diff --git "a/exp/log/log-train-2023-04-26-10-07-06-3" "b/exp/log/log-train-2023-04-26-10-07-06-3" new file mode 100644--- /dev/null +++ "b/exp/log/log-train-2023-04-26-10-07-06-3" @@ -0,0 +1,23370 @@ +2023-04-26 10:07:06,651 INFO [finetune.py:1046] (3/7) Training started +2023-04-26 10:07:06,651 INFO [finetune.py:1056] (3/7) Device: cuda:3 +2023-04-26 10:07:06,653 INFO [finetune.py:1065] (3/7) {'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/exp2'), '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_fr/ASR/pruned_transducer_stateless7_streaming/exp/english_pretrain/pretrained.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} +2023-04-26 10:07:06,653 INFO [finetune.py:1067] (3/7) About to create model +2023-04-26 10:07:07,030 INFO [zipformer.py:405] (3/7) 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. +2023-04-26 10:07:07,039 INFO [finetune.py:1071] (3/7) Number of model parameters: 70369391 +2023-04-26 10:07:07,040 INFO [finetune.py:626] (3/7) Loading checkpoint from /home/lishaojie/icefall/egs/commonvoice_fr/ASR/pruned_transducer_stateless7_streaming/exp/english_pretrain/pretrained.pt +2023-04-26 10:07:07,203 INFO [finetune.py:647] (3/7) Loading parameters starting with prefix encoder +2023-04-26 10:07:08,583 INFO [finetune.py:1093] (3/7) Using DDP +2023-04-26 10:07:09,243 INFO [commonvoice_fr.py:392] (3/7) About to get train cuts +2023-04-26 10:07:09,246 INFO [commonvoice_fr.py:218] (3/7) Enable MUSAN +2023-04-26 10:07:09,246 INFO [commonvoice_fr.py:219] (3/7) About to get Musan cuts +2023-04-26 10:07:10,773 INFO [commonvoice_fr.py:243] (3/7) Enable SpecAugment +2023-04-26 10:07:10,774 INFO [commonvoice_fr.py:244] (3/7) Time warp factor: 80 +2023-04-26 10:07:10,774 INFO [commonvoice_fr.py:254] (3/7) Num frame mask: 10 +2023-04-26 10:07:10,774 INFO [commonvoice_fr.py:267] (3/7) About to create train dataset +2023-04-26 10:07:10,774 INFO [commonvoice_fr.py:294] (3/7) Using DynamicBucketingSampler. +2023-04-26 10:07:13,473 INFO [commonvoice_fr.py:309] (3/7) About to create train dataloader +2023-04-26 10:07:13,473 INFO [commonvoice_fr.py:399] (3/7) About to get dev cuts +2023-04-26 10:07:13,474 INFO [commonvoice_fr.py:340] (3/7) About to create dev dataset +2023-04-26 10:07:13,884 INFO [commonvoice_fr.py:357] (3/7) About to create dev dataloader +2023-04-26 10:07:13,884 INFO [finetune.py:1289] (3/7) Sanity check -- see if any of the batches in epoch 1 would cause OOM. +2023-04-26 10:11:06,929 INFO [finetune.py:1317] (3/7) Maximum memory allocated so far is 5236MB +2023-04-26 10:11:07,627 INFO [finetune.py:1317] (3/7) Maximum memory allocated so far is 5768MB +2023-04-26 10:11:08,316 INFO [finetune.py:1317] (3/7) Maximum memory allocated so far is 5768MB +2023-04-26 10:11:08,990 INFO [finetune.py:1317] (3/7) Maximum memory allocated so far is 5768MB +2023-04-26 10:11:09,671 INFO [finetune.py:1317] (3/7) Maximum memory allocated so far is 5768MB +2023-04-26 10:11:10,373 INFO [finetune.py:1317] (3/7) Maximum memory allocated so far is 5768MB +2023-04-26 10:11:19,555 INFO [finetune.py:976] (3/7) Epoch 1, batch 0, loss[loss=7.447, simple_loss=6.75, pruned_loss=6.956, over 4703.00 frames. ], tot_loss[loss=7.447, simple_loss=6.75, pruned_loss=6.956, over 4703.00 frames. ], batch size: 23, lr: 2.00e-03, grad_scale: 2.0 +2023-04-26 10:11:19,555 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 10:11:40,203 INFO [finetune.py:1010] (3/7) Epoch 1, validation: loss=7.31, simple_loss=6.623, pruned_loss=6.857, over 2265189.00 frames. +2023-04-26 10:11:40,203 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 5768MB +2023-04-26 10:11:48,678 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:12:02,913 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0545, 0.7897, 0.4268, 0.4561, 0.5853, 0.9202, 0.3976, 0.5567], + device='cuda:3'), covar=tensor([1.7069, 1.4817, 2.2746, 2.1967, 1.8775, 1.9429, 2.8932, 1.2596], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0507, 0.0407, 0.0397, 0.0440, 0.0445, 0.0494, 0.0430], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:12:11,187 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:12:13,269 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8265, 4.0415, 3.4905, 4.1365, 3.1296, 3.6112, 2.2199, 3.6558], + device='cuda:3'), covar=tensor([0.2078, 0.1421, 0.2059, 0.1225, 0.6598, 0.2316, 0.3075, 0.1632], + device='cuda:3'), in_proj_covar=tensor([0.0265, 0.0237, 0.0290, 0.0333, 0.0330, 0.0273, 0.0293, 0.0287], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:12:31,704 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3910, 1.0595, 1.4582, 1.3164, 0.8988, 0.9085, 1.1088, 0.9373], + device='cuda:3'), covar=tensor([0.0111, 0.0265, 0.0167, 0.0131, 0.0212, 0.0147, 0.0170, 0.0342], + device='cuda:3'), in_proj_covar=tensor([0.0083, 0.0093, 0.0087, 0.0092, 0.0108, 0.0106, 0.0105, 0.0092], + device='cuda:3'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:3') +2023-04-26 10:12:43,282 INFO [finetune.py:976] (3/7) Epoch 1, batch 50, loss[loss=2.568, simple_loss=2.44, pruned_loss=1.292, over 4927.00 frames. ], tot_loss[loss=4.46, simple_loss=4.037, pruned_loss=4.085, over 213004.45 frames. ], batch size: 33, lr: 2.20e-03, grad_scale: 0.00390625 +2023-04-26 10:12:46,163 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.68 vs. limit=2.0 +2023-04-26 10:12:58,692 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.50 vs. limit=2.0 +2023-04-26 10:13:16,383 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=3.98 vs. limit=2.0 +2023-04-26 10:13:19,821 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:13:19,870 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=15.83 vs. limit=5.0 +2023-04-26 10:13:41,117 WARNING [finetune.py:966] (3/7) Grad scale is small: 6.103515625e-05 +2023-04-26 10:13:41,118 INFO [finetune.py:976] (3/7) Epoch 1, batch 100, loss[loss=2.337, simple_loss=2.204, pruned_loss=1.298, over 4887.00 frames. ], tot_loss[loss=3.481, simple_loss=3.22, pruned_loss=2.56, over 378676.73 frames. ], batch size: 32, lr: 2.40e-03, grad_scale: 0.0001220703125 +2023-04-26 10:13:50,957 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=28.28 vs. limit=5.0 +2023-04-26 10:14:03,492 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 4.683e+02 1.507e+03 7.833e+03 2.572e+04 3.214e+07, threshold=1.567e+04, percent-clipped=0.0 +2023-04-26 10:14:06,180 WARNING [optim.py:389] (3/7) Scaling gradients by 0.014711554162204266, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:06,253 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.88, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.001e+12, grad_sumsq = 2.310e+12, orig_rms_sq=4.331e-01 +2023-04-26 10:14:23,407 WARNING [optim.py:389] (3/7) Scaling gradients by 0.00018281130178365856, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:23,480 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.29, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=2.155e+15, grad_sumsq = 4.978e+15, orig_rms_sq=4.329e-01 +2023-04-26 10:14:23,844 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.72 vs. limit=2.0 +2023-04-26 10:14:24,113 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:14:27,645 INFO [finetune.py:976] (3/7) Epoch 1, batch 150, loss[loss=1.578, simple_loss=1.428, pruned_loss=1.216, over 4869.00 frames. ], tot_loss[loss=2.953, simple_loss=2.74, pruned_loss=2.054, over 505616.32 frames. ], batch size: 31, lr: 2.60e-03, grad_scale: 3.0517578125e-05 +2023-04-26 10:14:28,149 WARNING [optim.py:389] (3/7) Scaling gradients by 0.00022292081848718226, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:28,222 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.45, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=2.213e+15, grad_sumsq = 5.111e+15, orig_rms_sq=4.330e-01 +2023-04-26 10:14:40,641 WARNING [optim.py:389] (3/7) Scaling gradients by 0.05655747279524803, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:40,714 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.84, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=6.482e+10, grad_sumsq = 1.497e+11, orig_rms_sq=4.330e-01 +2023-04-26 10:14:56,529 WARNING [finetune.py:966] (3/7) Grad scale is small: 3.0517578125e-05 +2023-04-26 10:14:56,529 INFO [finetune.py:976] (3/7) Epoch 1, batch 200, loss[loss=1.449, simple_loss=1.252, pruned_loss=1.369, over 4815.00 frames. ], tot_loss[loss=2.438, simple_loss=2.24, pruned_loss=1.779, over 607044.57 frames. ], batch size: 51, lr: 2.80e-03, grad_scale: 6.103515625e-05 +2023-04-26 10:15:04,438 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.87 vs. limit=2.0 +2023-04-26 10:15:07,791 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 6.387e+01 5.300e+02 1.840e+03 7.547e+03 8.570e+07, threshold=3.680e+03, percent-clipped=20.0 +2023-04-26 10:15:12,965 WARNING [optim.py:389] (3/7) Scaling gradients by 0.011872046627104282, model_norm_threshold=3679.54541015625 +2023-04-26 10:15:13,039 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.57, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=5.451e+10, grad_sumsq = 1.259e+11, orig_rms_sq=4.329e-01 +2023-04-26 10:15:16,158 WARNING [optim.py:389] (3/7) Scaling gradients by 0.08515117317438126, model_norm_threshold=3679.54541015625 +2023-04-26 10:15:16,231 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.79, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.483e+09, grad_sumsq = 3.425e+09, orig_rms_sq=4.329e-01 +2023-04-26 10:15:16,775 WARNING [optim.py:389] (3/7) Scaling gradients by 0.04552413150668144, model_norm_threshold=3679.54541015625 +2023-04-26 10:15:16,848 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.84, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=5.493e+09, grad_sumsq = 1.269e+10, orig_rms_sq=4.329e-01 +2023-04-26 10:15:25,654 INFO [finetune.py:976] (3/7) Epoch 1, batch 250, loss[loss=1.541, simple_loss=1.31, pruned_loss=1.473, over 4839.00 frames. ], tot_loss[loss=2.101, simple_loss=1.904, pruned_loss=1.623, over 683643.92 frames. ], batch size: 47, lr: 3.00e-03, grad_scale: 6.103515625e-05 +2023-04-26 10:15:41,276 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=17.39 vs. limit=5.0 +2023-04-26 10:15:45,671 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8556, 1.9057, 1.8590, 1.3399, 1.5204, 1.5061, 1.6469, 1.3996], + device='cuda:3'), covar=tensor([0.0069, 0.0087, 0.0070, 0.0107, 0.0117, 0.0096, 0.0048, 0.0089], + device='cuda:3'), in_proj_covar=tensor([0.0224, 0.0255, 0.0225, 0.0248, 0.0263, 0.0220, 0.0213, 0.0237], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003, 0.0004], + device='cuda:3') +2023-04-26 10:15:49,829 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=26.93 vs. limit=5.0 +2023-04-26 10:15:50,800 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=296.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:15:52,835 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=300.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:15:53,285 WARNING [finetune.py:966] (3/7) Grad scale is small: 6.103515625e-05 +2023-04-26 10:15:53,285 INFO [finetune.py:976] (3/7) Epoch 1, batch 300, loss[loss=1.587, simple_loss=1.325, pruned_loss=1.539, over 4815.00 frames. ], tot_loss[loss=1.875, simple_loss=1.676, pruned_loss=1.519, over 742631.54 frames. ], batch size: 41, lr: 3.20e-03, grad_scale: 0.0001220703125 +2023-04-26 10:16:02,530 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 2.002e+01 7.343e+01 2.302e+02 1.070e+03 3.099e+05, threshold=4.604e+02, percent-clipped=16.0 +2023-04-26 10:16:21,500 INFO [finetune.py:976] (3/7) Epoch 1, batch 350, loss[loss=1.291, simple_loss=1.067, pruned_loss=1.231, over 4784.00 frames. ], tot_loss[loss=1.723, simple_loss=1.518, pruned_loss=1.449, over 789302.13 frames. ], batch size: 51, lr: 3.40e-03, grad_scale: 0.0001220703125 +2023-04-26 10:16:24,672 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=357.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:16:30,448 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=24.20 vs. limit=5.0 +2023-04-26 10:16:42,453 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=387.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:16:55,918 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=24.25 vs. limit=5.0 +2023-04-26 10:16:56,316 WARNING [finetune.py:966] (3/7) Grad scale is small: 0.0001220703125 +2023-04-26 10:16:56,316 INFO [finetune.py:976] (3/7) Epoch 1, batch 400, loss[loss=1.274, simple_loss=1.015, pruned_loss=1.285, over 4905.00 frames. ], tot_loss[loss=1.61, simple_loss=1.396, pruned_loss=1.398, over 825904.20 frames. ], batch size: 37, lr: 3.60e-03, grad_scale: 0.000244140625 +2023-04-26 10:17:16,619 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.830e+01 2.642e+01 7.119e+01 5.229e+02 3.680e+03, threshold=1.424e+02, percent-clipped=26.0 +2023-04-26 10:17:27,882 WARNING [optim.py:389] (3/7) Scaling gradients by 0.02133115753531456, model_norm_threshold=142.37583923339844 +2023-04-26 10:17:27,956 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.81, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=3.619e+07, grad_sumsq = 8.362e+07, orig_rms_sq=4.328e-01 +2023-04-26 10:17:41,635 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=439.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:17:52,442 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=448.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:17:53,957 INFO [finetune.py:976] (3/7) Epoch 1, batch 450, loss[loss=1.156, simple_loss=0.9051, pruned_loss=1.164, over 4903.00 frames. ], tot_loss[loss=1.507, simple_loss=1.285, pruned_loss=1.348, over 854315.78 frames. ], batch size: 36, lr: 3.80e-03, grad_scale: 0.000244140625 +2023-04-26 10:18:14,583 WARNING [optim.py:389] (3/7) Scaling gradients by 0.06225070729851723, model_norm_threshold=142.37583923339844 +2023-04-26 10:18:14,657 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.68, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=3.535e+06, grad_sumsq = 8.167e+06, orig_rms_sq=4.328e-01 +2023-04-26 10:18:22,849 WARNING [finetune.py:966] (3/7) Grad scale is small: 0.000244140625 +2023-04-26 10:18:22,849 INFO [finetune.py:976] (3/7) Epoch 1, batch 500, loss[loss=1.165, simple_loss=0.8952, pruned_loss=1.173, over 4939.00 frames. ], tot_loss[loss=1.415, simple_loss=1.185, pruned_loss=1.296, over 874732.85 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 0.00048828125 +2023-04-26 10:18:31,734 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.709e+01 2.273e+01 3.115e+01 1.071e+02 6.675e+03, threshold=6.230e+01, percent-clipped=18.0 +2023-04-26 10:18:34,949 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5065, 1.2987, 1.7008, 1.5870, 0.9274, 2.2176, 2.5604, 2.0744], + device='cuda:3'), covar=tensor([0.3590, 0.3934, 0.4670, 0.2675, 0.6055, 0.3766, 0.1887, 0.2106], + device='cuda:3'), in_proj_covar=tensor([0.0524, 0.0597, 0.0674, 0.0650, 0.0578, 0.0635, 0.0668, 0.0657], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:18:47,432 WARNING [optim.py:389] (3/7) Scaling gradients by 0.017591100186109543, model_norm_threshold=62.30100631713867 +2023-04-26 10:18:47,507 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.35, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=4.348e+06, grad_sumsq = 1.005e+07, orig_rms_sq=4.327e-01 +2023-04-26 10:18:56,306 WARNING [optim.py:389] (3/7) Scaling gradients by 0.005508477333933115, model_norm_threshold=62.30100631713867 +2023-04-26 10:18:56,378 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.80, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.024e+08, grad_sumsq = 2.367e+08, orig_rms_sq=4.327e-01 +2023-04-26 10:18:58,482 INFO [finetune.py:976] (3/7) Epoch 1, batch 550, loss[loss=1.09, simple_loss=0.8109, pruned_loss=1.122, over 4751.00 frames. ], tot_loss[loss=1.334, simple_loss=1.096, pruned_loss=1.248, over 891101.62 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 0.00048828125 +2023-04-26 10:19:04,297 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=562.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:19:13,086 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=568.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:19:35,171 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=590.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:19:45,890 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=600.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:19:46,342 WARNING [finetune.py:966] (3/7) Grad scale is small: 0.00048828125 +2023-04-26 10:19:46,343 INFO [finetune.py:976] (3/7) Epoch 1, batch 600, loss[loss=1.248, simple_loss=0.9085, pruned_loss=1.285, over 4798.00 frames. ], tot_loss[loss=1.276, simple_loss=1.028, pruned_loss=1.215, over 906430.90 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 0.0009765625 +2023-04-26 10:20:06,709 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.706e+01 2.227e+01 2.629e+01 6.305e+01 1.131e+04, threshold=5.258e+01, percent-clipped=26.0 +2023-04-26 10:20:09,427 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=623.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:20:18,147 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=629.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:20:29,919 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=648.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:20:31,452 INFO [finetune.py:976] (3/7) Epoch 1, batch 650, loss[loss=1.184, simple_loss=0.8598, pruned_loss=1.181, over 4854.00 frames. ], tot_loss[loss=1.239, simple_loss=0.9796, pruned_loss=1.192, over 918151.14 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 0.0009765625 +2023-04-26 10:20:31,539 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=651.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:20:32,028 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=652.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:20:41,809 WARNING [optim.py:389] (3/7) Scaling gradients by 0.07653743773698807, model_norm_threshold=52.5806770324707 +2023-04-26 10:20:41,881 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.93, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=4.392e+05, grad_sumsq = 1.015e+06, orig_rms_sq=4.327e-01 +2023-04-26 10:20:59,953 WARNING [finetune.py:966] (3/7) Grad scale is small: 0.0009765625 +2023-04-26 10:20:59,954 INFO [finetune.py:976] (3/7) Epoch 1, batch 700, loss[loss=1.148, simple_loss=0.8278, pruned_loss=1.121, over 4905.00 frames. ], tot_loss[loss=1.21, simple_loss=0.9388, pruned_loss=1.17, over 926312.73 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 0.001953125 +2023-04-26 10:21:09,242 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.790e+01 2.165e+01 2.490e+01 3.193e+01 6.870e+02, threshold=4.979e+01, percent-clipped=6.0 +2023-04-26 10:21:20,035 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=739.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:21:22,085 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=743.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:21:28,227 INFO [finetune.py:976] (3/7) Epoch 1, batch 750, loss[loss=1.12, simple_loss=0.8034, pruned_loss=1.068, over 4918.00 frames. ], tot_loss[loss=1.183, simple_loss=0.9037, pruned_loss=1.143, over 932224.79 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 0.001953125 +2023-04-26 10:21:48,071 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=787.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:21:53,230 WARNING [optim.py:389] (3/7) Scaling gradients by 0.039711207151412964, model_norm_threshold=49.79251480102539 +2023-04-26 10:21:53,323 INFO [optim.py:451] (3/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.81, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.271e+06, grad_sumsq = 2.939e+06, orig_rms_sq=4.325e-01 +2023-04-26 10:21:55,911 WARNING [finetune.py:966] (3/7) Grad scale is small: 0.001953125 +2023-04-26 10:21:55,911 INFO [finetune.py:976] (3/7) Epoch 1, batch 800, loss[loss=1.041, simple_loss=0.7346, pruned_loss=0.9857, over 4888.00 frames. ], tot_loss[loss=1.156, simple_loss=0.8711, pruned_loss=1.113, over 936632.91 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.00390625 +2023-04-26 10:22:11,121 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 2.049e+01 2.360e+01 2.808e+01 3.468e+01 1.254e+03, threshold=5.615e+01, percent-clipped=6.0 +2023-04-26 10:22:26,848 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=847.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:22:28,867 INFO [finetune.py:976] (3/7) Epoch 1, batch 850, loss[loss=1.049, simple_loss=0.7623, pruned_loss=0.9304, over 4771.00 frames. ], tot_loss[loss=1.123, simple_loss=0.8377, pruned_loss=1.071, over 937449.00 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 0.00390625 +2023-04-26 10:23:15,923 WARNING [finetune.py:966] (3/7) Grad scale is small: 0.00390625 +2023-04-26 10:23:15,923 INFO [finetune.py:976] (3/7) Epoch 1, batch 900, loss[loss=0.9303, simple_loss=0.6832, pruned_loss=0.7934, over 4246.00 frames. ], tot_loss[loss=1.089, simple_loss=0.8048, pruned_loss=1.027, over 940249.23 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 0.0078125 +2023-04-26 10:23:20,202 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=908.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:23:26,895 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 2.132e+01 2.534e+01 2.883e+01 3.456e+01 6.931e+01, threshold=5.766e+01, percent-clipped=4.0 +2023-04-26 10:23:26,977 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=918.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:23:28,272 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=11.07 vs. limit=5.0 +2023-04-26 10:23:30,107 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=924.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:23:32,794 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=16.31 vs. limit=5.0 +2023-04-26 10:23:37,199 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-04-26 10:23:42,107 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=946.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:23:44,693 INFO [finetune.py:976] (3/7) Epoch 1, batch 950, loss[loss=1.022, simple_loss=0.7217, pruned_loss=0.8937, over 4907.00 frames. ], tot_loss[loss=1.066, simple_loss=0.7814, pruned_loss=0.9903, over 944983.26 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 0.0078125 +2023-04-26 10:23:45,284 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=952.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:24:41,986 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1000.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:24:42,451 WARNING [finetune.py:966] (3/7) Grad scale is small: 0.0078125 +2023-04-26 10:24:42,451 INFO [finetune.py:976] (3/7) Epoch 1, batch 1000, loss[loss=0.9895, simple_loss=0.6962, pruned_loss=0.8494, over 4822.00 frames. ], tot_loss[loss=1.061, simple_loss=0.7722, pruned_loss=0.9686, over 947481.05 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 0.015625 +2023-04-26 10:25:00,170 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 2.486e+01 3.090e+01 3.594e+01 4.276e+01 7.170e+01, threshold=7.188e+01, percent-clipped=7.0 +2023-04-26 10:25:13,060 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.95 vs. limit=2.0 +2023-04-26 10:25:14,558 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1043.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:25:18,745 INFO [finetune.py:976] (3/7) Epoch 1, batch 1050, loss[loss=1.065, simple_loss=0.7529, pruned_loss=0.8898, over 4841.00 frames. ], tot_loss[loss=1.062, simple_loss=0.7689, pruned_loss=0.9522, over 947955.63 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 0.015625 +2023-04-26 10:25:42,489 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1091.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:25:48,291 INFO [finetune.py:976] (3/7) Epoch 1, batch 1100, loss[loss=1.045, simple_loss=0.7419, pruned_loss=0.8527, over 4884.00 frames. ], tot_loss[loss=1.056, simple_loss=0.761, pruned_loss=0.93, over 949134.25 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 0.03125 +2023-04-26 10:25:57,274 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 2.855e+01 3.563e+01 4.845e+01 6.131e+01 1.271e+02, threshold=9.690e+01, percent-clipped=11.0 +2023-04-26 10:26:17,718 INFO [finetune.py:976] (3/7) Epoch 1, batch 1150, loss[loss=1.083, simple_loss=0.7584, pruned_loss=0.8785, over 4817.00 frames. ], tot_loss[loss=1.049, simple_loss=0.7525, pruned_loss=0.9083, over 951348.64 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 0.03125 +2023-04-26 10:26:18,900 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1153.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:26:46,978 INFO [finetune.py:976] (3/7) Epoch 1, batch 1200, loss[loss=0.9503, simple_loss=0.6621, pruned_loss=0.7604, over 4884.00 frames. ], tot_loss[loss=1.037, simple_loss=0.7408, pruned_loss=0.8838, over 953296.56 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.0625 +2023-04-26 10:26:48,601 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1203.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:26:49,364 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=10.18 vs. limit=5.0 +2023-04-26 10:26:54,335 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1214.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:26:56,332 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 2.920e+01 3.405e+01 3.864e+01 4.670e+01 1.171e+02, threshold=7.728e+01, percent-clipped=2.0 +2023-04-26 10:26:56,428 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1218.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:26:59,564 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1224.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:27:19,095 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1246.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:27:21,547 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.20 vs. limit=5.0 +2023-04-26 10:27:21,698 INFO [finetune.py:976] (3/7) Epoch 1, batch 1250, loss[loss=0.9836, simple_loss=0.6843, pruned_loss=0.7742, over 4902.00 frames. ], tot_loss[loss=1.022, simple_loss=0.7268, pruned_loss=0.8575, over 953676.39 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 0.0625 +2023-04-26 10:27:40,180 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1266.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:27:49,524 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1272.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:28:01,502 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.52 vs. limit=5.0 +2023-04-26 10:28:15,537 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1294.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:28:24,915 INFO [finetune.py:976] (3/7) Epoch 1, batch 1300, loss[loss=1.029, simple_loss=0.7189, pruned_loss=0.7942, over 4857.00 frames. ], tot_loss[loss=1.007, simple_loss=0.7128, pruned_loss=0.8327, over 954479.09 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 0.125 +2023-04-26 10:28:40,245 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 3.024e+01 3.776e+01 4.391e+01 5.836e+01 1.199e+02, threshold=8.782e+01, percent-clipped=8.0 +2023-04-26 10:29:00,693 INFO [finetune.py:976] (3/7) Epoch 1, batch 1350, loss[loss=1.008, simple_loss=0.6905, pruned_loss=0.7781, over 4818.00 frames. ], tot_loss[loss=1.008, simple_loss=0.7099, pruned_loss=0.8204, over 952718.90 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 0.125 +2023-04-26 10:29:20,112 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.11 vs. limit=2.0 +2023-04-26 10:29:49,395 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 10:29:52,993 INFO [finetune.py:976] (3/7) Epoch 1, batch 1400, loss[loss=1.155, simple_loss=0.8036, pruned_loss=0.8667, over 4762.00 frames. ], tot_loss[loss=1.014, simple_loss=0.7101, pruned_loss=0.8139, over 954888.26 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 0.25 +2023-04-26 10:30:14,313 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 3.571e+01 4.693e+01 6.414e+01 7.905e+01 1.778e+02, threshold=1.283e+02, percent-clipped=17.0 +2023-04-26 10:30:34,171 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1434.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:30:55,950 INFO [finetune.py:976] (3/7) Epoch 1, batch 1450, loss[loss=1.044, simple_loss=0.7161, pruned_loss=0.7804, over 4918.00 frames. ], tot_loss[loss=1.02, simple_loss=0.7111, pruned_loss=0.8075, over 955124.67 frames. ], batch size: 42, lr: 4.00e-03, grad_scale: 0.25 +2023-04-26 10:31:30,404 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=72.42 vs. limit=5.0 +2023-04-26 10:31:42,612 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1495.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:31:53,098 INFO [finetune.py:976] (3/7) Epoch 1, batch 1500, loss[loss=1.044, simple_loss=0.723, pruned_loss=0.7633, over 4804.00 frames. ], tot_loss[loss=1.022, simple_loss=0.7102, pruned_loss=0.7975, over 955990.40 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 0.5 +2023-04-26 10:31:59,746 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1503.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:32:03,092 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1509.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:32:13,076 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 3.937e+01 5.249e+01 6.518e+01 8.001e+01 1.317e+02, threshold=1.304e+02, percent-clipped=3.0 +2023-04-26 10:32:34,189 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9081, 1.1920, 1.4869, 1.7882, 1.5026, 1.1977, 1.1456, 1.3705], + device='cuda:3'), covar=tensor([0.8969, 1.2334, 0.8451, 1.3631, 1.3586, 1.0105, 1.5253, 0.7501], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0319, 0.0252, 0.0415, 0.0276, 0.0264, 0.0313, 0.0260], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:32:48,632 INFO [finetune.py:976] (3/7) Epoch 1, batch 1550, loss[loss=0.9687, simple_loss=0.6724, pruned_loss=0.6981, over 4819.00 frames. ], tot_loss[loss=1.014, simple_loss=0.7039, pruned_loss=0.7788, over 956593.19 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 0.5 +2023-04-26 10:32:48,699 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:33:08,364 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1566.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:33:30,153 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=26.31 vs. limit=5.0 +2023-04-26 10:33:50,408 INFO [finetune.py:976] (3/7) Epoch 1, batch 1600, loss[loss=0.8841, simple_loss=0.6279, pruned_loss=0.6186, over 4898.00 frames. ], tot_loss[loss=0.9901, simple_loss=0.6895, pruned_loss=0.7486, over 957472.13 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:34:04,237 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1611.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:34:12,361 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-04-26 10:34:13,528 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 5.628e+01 9.315e+01 1.324e+02 1.713e+02 3.757e+02, threshold=2.648e+02, percent-clipped=49.0 +2023-04-26 10:34:24,072 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1627.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:34:48,572 INFO [finetune.py:976] (3/7) Epoch 1, batch 1650, loss[loss=0.8062, simple_loss=0.5847, pruned_loss=0.5491, over 4933.00 frames. ], tot_loss[loss=0.9601, simple_loss=0.6728, pruned_loss=0.7135, over 955942.19 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:35:08,303 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1672.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:35:32,133 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8498, 2.3028, 1.6120, 2.5318, 1.5146, 2.0102, 1.9853, 1.4605], + device='cuda:3'), covar=tensor([0.0812, 0.0592, 0.1148, 0.0430, 0.1463, 0.0770, 0.0801, 0.1302], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0307, 0.0224, 0.0284, 0.0301, 0.0260, 0.0262, 0.0273], + device='cuda:3'), out_proj_covar=tensor([1.1935e-04, 1.2593e-04, 9.1693e-05, 1.1496e-04, 1.2510e-04, 1.0488e-04, + 1.0901e-04, 1.1121e-04], device='cuda:3') +2023-04-26 10:35:36,524 INFO [finetune.py:976] (3/7) Epoch 1, batch 1700, loss[loss=0.8384, simple_loss=0.6236, pruned_loss=0.5545, over 4813.00 frames. ], tot_loss[loss=0.9285, simple_loss=0.6568, pruned_loss=0.6769, over 955506.91 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:35:59,302 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.556e+01 1.723e+02 2.061e+02 2.467e+02 4.417e+02, threshold=4.123e+02, percent-clipped=15.0 +2023-04-26 10:36:01,126 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=16.27 vs. limit=5.0 +2023-04-26 10:36:09,463 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8552, 2.6904, 2.1415, 3.2081, 2.7068, 2.8011, 1.1632, 2.6947], + device='cuda:3'), covar=tensor([0.1956, 0.1438, 0.2931, 0.1499, 0.2317, 0.1820, 0.5882, 0.2143], + device='cuda:3'), in_proj_covar=tensor([0.0264, 0.0239, 0.0293, 0.0332, 0.0330, 0.0273, 0.0290, 0.0287], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:36:24,059 INFO [finetune.py:976] (3/7) Epoch 1, batch 1750, loss[loss=0.736, simple_loss=0.5651, pruned_loss=0.471, over 4822.00 frames. ], tot_loss[loss=0.9032, simple_loss=0.6469, pruned_loss=0.6448, over 952998.11 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:36:58,581 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1790.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:37:11,178 INFO [finetune.py:976] (3/7) Epoch 1, batch 1800, loss[loss=0.6413, simple_loss=0.4962, pruned_loss=0.4048, over 4773.00 frames. ], tot_loss[loss=0.8747, simple_loss=0.6359, pruned_loss=0.611, over 952314.68 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:37:21,905 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1809.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:37:28,282 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.499e+02 2.323e+02 2.803e+02 3.399e+02 5.478e+02, threshold=5.607e+02, percent-clipped=9.0 +2023-04-26 10:37:40,486 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1838.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:37:41,660 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=6.28 vs. limit=5.0 +2023-04-26 10:37:47,698 INFO [finetune.py:976] (3/7) Epoch 1, batch 1850, loss[loss=0.6321, simple_loss=0.5164, pruned_loss=0.3801, over 4839.00 frames. ], tot_loss[loss=0.8399, simple_loss=0.6207, pruned_loss=0.574, over 953182.32 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:37:51,630 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1857.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:37:51,659 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1857.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:37:54,473 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.1915, 4.0451, 3.0522, 4.7678, 3.8800, 4.1457, 1.7082, 3.9794], + device='cuda:3'), covar=tensor([0.1550, 0.1073, 0.4516, 0.0996, 0.2539, 0.1527, 0.5788, 0.2241], + device='cuda:3'), in_proj_covar=tensor([0.0264, 0.0239, 0.0292, 0.0331, 0.0332, 0.0273, 0.0289, 0.0288], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:37:59,992 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1870.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:38:17,755 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1899.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:38:18,349 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.62 vs. limit=5.0 +2023-04-26 10:38:18,742 INFO [finetune.py:976] (3/7) Epoch 1, batch 1900, loss[loss=0.6462, simple_loss=0.5368, pruned_loss=0.3815, over 4743.00 frames. ], tot_loss[loss=0.8059, simple_loss=0.6065, pruned_loss=0.5386, over 953722.59 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:38:28,761 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.629e+02 2.504e+02 3.040e+02 3.671e+02 6.110e+02, threshold=6.080e+02, percent-clipped=2.0 +2023-04-26 10:38:28,866 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1918.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:38:32,527 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1922.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 10:38:39,021 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1931.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:38:49,094 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7717, 1.2379, 1.8917, 2.0123, 1.5496, 1.2018, 1.4508, 0.8776], + device='cuda:3'), covar=tensor([0.0519, 0.0484, 0.0391, 0.0231, 0.0531, 0.1104, 0.0528, 0.1078], + device='cuda:3'), in_proj_covar=tensor([0.0080, 0.0089, 0.0082, 0.0087, 0.0103, 0.0105, 0.0102, 0.0089], + device='cuda:3'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:3') +2023-04-26 10:38:50,104 INFO [finetune.py:976] (3/7) Epoch 1, batch 1950, loss[loss=0.6516, simple_loss=0.538, pruned_loss=0.3845, over 4797.00 frames. ], tot_loss[loss=0.7671, simple_loss=0.5872, pruned_loss=0.502, over 955214.24 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:38:57,508 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2721, 1.7452, 1.5584, 1.6648, 1.6043, 1.8759, 1.7300, 4.3153], + device='cuda:3'), covar=tensor([0.1332, 0.1545, 0.1733, 0.3110, 0.1636, 0.1202, 0.1531, 0.0186], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0044, 0.0044, 0.0050, 0.0045, 0.0042, 0.0044, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0016, 0.0014, 0.0013, 0.0014, 0.0018], + device='cuda:3') +2023-04-26 10:38:59,750 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1967.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:39:22,527 INFO [finetune.py:976] (3/7) Epoch 1, batch 2000, loss[loss=0.5439, simple_loss=0.4684, pruned_loss=0.3097, over 4794.00 frames. ], tot_loss[loss=0.7286, simple_loss=0.5667, pruned_loss=0.4675, over 956954.03 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 2.0 +2023-04-26 10:39:39,761 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.592e+02 2.794e+02 3.334e+02 3.918e+02 6.535e+02, threshold=6.668e+02, percent-clipped=3.0 +2023-04-26 10:39:57,485 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.25 vs. limit=5.0 +2023-04-26 10:40:01,846 INFO [finetune.py:976] (3/7) Epoch 1, batch 2050, loss[loss=0.4424, simple_loss=0.3955, pruned_loss=0.2447, over 4343.00 frames. ], tot_loss[loss=0.6886, simple_loss=0.5451, pruned_loss=0.4334, over 957675.53 frames. ], batch size: 19, lr: 4.00e-03, grad_scale: 2.0 +2023-04-26 10:40:10,127 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5936, 1.3300, 4.2861, 3.9990, 3.7952, 4.0048, 4.0165, 3.7394], + device='cuda:3'), covar=tensor([0.6436, 0.5862, 0.0914, 0.1519, 0.1135, 0.1848, 0.1482, 0.1396], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0309, 0.0444, 0.0456, 0.0371, 0.0425, 0.0340, 0.0395], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 10:40:27,118 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-04-26 10:40:37,622 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2090.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:40:50,242 INFO [finetune.py:976] (3/7) Epoch 1, batch 2100, loss[loss=0.517, simple_loss=0.4448, pruned_loss=0.2946, over 4759.00 frames. ], tot_loss[loss=0.6554, simple_loss=0.5281, pruned_loss=0.4048, over 955111.39 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 2.0 +2023-04-26 10:40:52,164 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4901, 3.2788, 2.1055, 2.4033, 1.8643, 1.8186, 1.9963, 1.6939], + device='cuda:3'), covar=tensor([0.4019, 0.3361, 0.7512, 0.6604, 0.6784, 0.7481, 0.5634, 0.6922], + device='cuda:3'), in_proj_covar=tensor([0.0217, 0.0242, 0.0220, 0.0237, 0.0256, 0.0217, 0.0211, 0.0229], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004], + device='cuda:3') +2023-04-26 10:41:11,503 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.614e+02 2.534e+02 2.914e+02 3.246e+02 6.149e+02, threshold=5.827e+02, percent-clipped=0.0 +2023-04-26 10:41:31,945 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2138.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:41:41,043 INFO [finetune.py:976] (3/7) Epoch 1, batch 2150, loss[loss=0.4639, simple_loss=0.4193, pruned_loss=0.2543, over 4750.00 frames. ], tot_loss[loss=0.634, simple_loss=0.5205, pruned_loss=0.3843, over 952425.67 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:42:25,865 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2194.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:42:30,519 INFO [finetune.py:976] (3/7) Epoch 1, batch 2200, loss[loss=0.5518, simple_loss=0.4994, pruned_loss=0.3021, over 4814.00 frames. ], tot_loss[loss=0.6131, simple_loss=0.5128, pruned_loss=0.3649, over 954304.23 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:42:37,619 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2213.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:42:40,428 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.685e+02 2.565e+02 3.033e+02 3.519e+02 5.393e+02, threshold=6.065e+02, percent-clipped=0.0 +2023-04-26 10:42:42,260 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6812, 3.5172, 0.9882, 1.9456, 1.9964, 2.5951, 2.2156, 1.0611], + device='cuda:3'), covar=tensor([0.1158, 0.0681, 0.1930, 0.1170, 0.0919, 0.0876, 0.1259, 0.1737], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0265, 0.0146, 0.0130, 0.0141, 0.0161, 0.0129, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 10:42:42,849 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2222.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:42:42,859 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6572, 1.4860, 1.6948, 2.0225, 2.2082, 1.6008, 1.2650, 1.7027], + device='cuda:3'), covar=tensor([0.1094, 0.1431, 0.0964, 0.0877, 0.0642, 0.0977, 0.1254, 0.0799], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0209, 0.0188, 0.0167, 0.0167, 0.0184, 0.0162, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 10:42:45,139 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2226.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:43:02,982 INFO [finetune.py:976] (3/7) Epoch 1, batch 2250, loss[loss=0.5522, simple_loss=0.5059, pruned_loss=0.2992, over 4845.00 frames. ], tot_loss[loss=0.5929, simple_loss=0.5045, pruned_loss=0.347, over 954811.96 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:43:11,249 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1342, 2.6874, 1.0269, 1.5520, 1.8947, 1.4483, 3.5047, 1.5924], + device='cuda:3'), covar=tensor([0.0778, 0.0866, 0.1242, 0.1101, 0.0615, 0.0994, 0.0133, 0.0721], + device='cuda:3'), in_proj_covar=tensor([0.0057, 0.0073, 0.0053, 0.0050, 0.0054, 0.0056, 0.0089, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0012, 0.0008], + device='cuda:3') +2023-04-26 10:43:12,377 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2267.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:43:14,137 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2270.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:43:34,422 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3664, 1.2780, 1.4164, 1.7328, 1.9114, 1.3212, 0.9244, 1.4546], + device='cuda:3'), covar=tensor([0.1047, 0.1724, 0.1211, 0.0724, 0.0548, 0.1024, 0.1274, 0.0852], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0208, 0.0187, 0.0167, 0.0167, 0.0183, 0.0162, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 10:43:57,383 INFO [finetune.py:976] (3/7) Epoch 1, batch 2300, loss[loss=0.5142, simple_loss=0.461, pruned_loss=0.2837, over 4120.00 frames. ], tot_loss[loss=0.5725, simple_loss=0.4952, pruned_loss=0.3298, over 954588.75 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:44:17,315 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2315.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:44:19,009 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.752e+02 2.423e+02 2.781e+02 3.392e+02 7.688e+02, threshold=5.562e+02, percent-clipped=1.0 +2023-04-26 10:44:43,409 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2340.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:44:51,819 INFO [finetune.py:976] (3/7) Epoch 1, batch 2350, loss[loss=0.5371, simple_loss=0.4887, pruned_loss=0.2928, over 4894.00 frames. ], tot_loss[loss=0.5474, simple_loss=0.4809, pruned_loss=0.3108, over 955471.15 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:45:40,784 INFO [finetune.py:976] (3/7) Epoch 1, batch 2400, loss[loss=0.4602, simple_loss=0.429, pruned_loss=0.2457, over 4909.00 frames. ], tot_loss[loss=0.5263, simple_loss=0.468, pruned_loss=0.2953, over 953687.52 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:45:40,904 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2401.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:45:51,693 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.659e+02 2.375e+02 2.859e+02 3.332e+02 6.118e+02, threshold=5.719e+02, percent-clipped=1.0 +2023-04-26 10:45:58,907 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=7.05 vs. limit=5.0 +2023-04-26 10:46:11,803 INFO [finetune.py:976] (3/7) Epoch 1, batch 2450, loss[loss=0.4511, simple_loss=0.4006, pruned_loss=0.2508, over 4829.00 frames. ], tot_loss[loss=0.5077, simple_loss=0.4567, pruned_loss=0.2817, over 955222.95 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:46:26,830 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7469, 1.2559, 4.3601, 4.0732, 3.8703, 4.0601, 4.0559, 3.8408], + device='cuda:3'), covar=tensor([0.6148, 0.5818, 0.0943, 0.1532, 0.1003, 0.1476, 0.1541, 0.1341], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0315, 0.0453, 0.0463, 0.0379, 0.0432, 0.0346, 0.0400], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 10:46:31,158 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 10:46:39,155 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2494.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:46:44,084 INFO [finetune.py:976] (3/7) Epoch 1, batch 2500, loss[loss=0.4752, simple_loss=0.4497, pruned_loss=0.2504, over 4862.00 frames. ], tot_loss[loss=0.4988, simple_loss=0.4531, pruned_loss=0.2741, over 954173.99 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:46:45,359 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9726, 1.1317, 3.2931, 3.0622, 2.9693, 3.1573, 3.2588, 2.9104], + device='cuda:3'), covar=tensor([0.7148, 0.5771, 0.1363, 0.2066, 0.1475, 0.1996, 0.1451, 0.1748], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0317, 0.0455, 0.0466, 0.0381, 0.0434, 0.0348, 0.0403], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 10:46:50,432 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0020, 1.1989, 0.8618, 1.3504, 1.2701, 0.9078, 1.1469, 0.8062], + device='cuda:3'), covar=tensor([0.1942, 0.1627, 0.1608, 0.1365, 0.3083, 0.1554, 0.1866, 0.2530], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0301, 0.0222, 0.0281, 0.0296, 0.0256, 0.0258, 0.0270], + device='cuda:3'), out_proj_covar=tensor([1.1806e-04, 1.2314e-04, 9.0916e-05, 1.1380e-04, 1.2315e-04, 1.0361e-04, + 1.0686e-04, 1.1014e-04], device='cuda:3') +2023-04-26 10:46:52,662 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2513.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:46:56,032 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.542e+02 2.485e+02 2.798e+02 3.305e+02 6.030e+02, threshold=5.597e+02, percent-clipped=1.0 +2023-04-26 10:47:00,959 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2526.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:47:03,423 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-04-26 10:47:08,128 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7491, 1.8341, 1.6627, 1.5643, 1.9391, 1.5840, 2.3926, 1.3943], + device='cuda:3'), covar=tensor([0.2936, 0.0967, 0.2124, 0.1642, 0.0980, 0.1559, 0.0550, 0.2453], + device='cuda:3'), in_proj_covar=tensor([0.0301, 0.0310, 0.0374, 0.0322, 0.0361, 0.0337, 0.0346, 0.0361], + device='cuda:3'), out_proj_covar=tensor([9.3273e-05, 9.5669e-05, 1.1584e-04, 1.0079e-04, 1.1080e-04, 1.0311e-04, + 1.0467e-04, 1.1200e-04], device='cuda:3') +2023-04-26 10:47:10,441 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2542.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:47:16,162 INFO [finetune.py:976] (3/7) Epoch 1, batch 2550, loss[loss=0.5055, simple_loss=0.4853, pruned_loss=0.2629, over 4819.00 frames. ], tot_loss[loss=0.4914, simple_loss=0.4519, pruned_loss=0.2669, over 954013.07 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:47:35,363 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2561.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:47:50,049 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2574.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:48:03,712 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-26 10:48:06,449 INFO [finetune.py:976] (3/7) Epoch 1, batch 2600, loss[loss=0.47, simple_loss=0.4569, pruned_loss=0.2415, over 4805.00 frames. ], tot_loss[loss=0.483, simple_loss=0.4485, pruned_loss=0.2598, over 953118.46 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:48:14,053 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4445, 3.5955, 1.0844, 1.9874, 1.9636, 2.4784, 2.2556, 1.0078], + device='cuda:3'), covar=tensor([0.1234, 0.0623, 0.1868, 0.1196, 0.1040, 0.0998, 0.1262, 0.1873], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0265, 0.0146, 0.0130, 0.0141, 0.0161, 0.0128, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 10:48:18,462 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.395e+02 2.465e+02 2.854e+02 3.439e+02 6.010e+02, threshold=5.707e+02, percent-clipped=1.0 +2023-04-26 10:48:38,252 INFO [finetune.py:976] (3/7) Epoch 1, batch 2650, loss[loss=0.4037, simple_loss=0.4095, pruned_loss=0.1989, over 4913.00 frames. ], tot_loss[loss=0.4742, simple_loss=0.4448, pruned_loss=0.2526, over 953717.65 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:48:42,089 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-04-26 10:49:01,819 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2685.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:49:14,009 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2696.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:49:14,118 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.93 vs. limit=5.0 +2023-04-26 10:49:16,820 INFO [finetune.py:976] (3/7) Epoch 1, batch 2700, loss[loss=0.4087, simple_loss=0.3943, pruned_loss=0.2115, over 4781.00 frames. ], tot_loss[loss=0.4638, simple_loss=0.4388, pruned_loss=0.2451, over 953442.70 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:49:27,768 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9250, 1.7456, 1.9644, 2.1869, 2.2377, 1.6513, 1.1661, 1.9376], + device='cuda:3'), covar=tensor([0.1286, 0.1396, 0.0967, 0.1124, 0.0755, 0.1380, 0.1545, 0.1023], + device='cuda:3'), in_proj_covar=tensor([0.0204, 0.0211, 0.0191, 0.0174, 0.0173, 0.0190, 0.0168, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 10:49:38,653 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-26 10:49:39,497 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.764e+02 2.585e+02 2.984e+02 3.485e+02 4.746e+02, threshold=5.968e+02, percent-clipped=0.0 +2023-04-26 10:50:04,125 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4557, 3.4636, 1.0196, 2.0080, 1.9766, 2.5455, 2.2935, 1.0842], + device='cuda:3'), covar=tensor([0.1139, 0.0742, 0.1846, 0.1080, 0.0921, 0.0900, 0.1160, 0.1832], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0266, 0.0147, 0.0130, 0.0141, 0.0162, 0.0128, 0.0131], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 10:50:04,161 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2746.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:50:13,026 INFO [finetune.py:976] (3/7) Epoch 1, batch 2750, loss[loss=0.4185, simple_loss=0.3982, pruned_loss=0.2195, over 4819.00 frames. ], tot_loss[loss=0.4497, simple_loss=0.4286, pruned_loss=0.2359, over 953402.73 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:51:16,383 INFO [finetune.py:976] (3/7) Epoch 1, batch 2800, loss[loss=0.3867, simple_loss=0.3826, pruned_loss=0.1954, over 4729.00 frames. ], tot_loss[loss=0.4375, simple_loss=0.4202, pruned_loss=0.2278, over 953674.90 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:51:26,537 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=8.50 vs. limit=5.0 +2023-04-26 10:51:38,601 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.584e+02 2.321e+02 2.765e+02 3.400e+02 8.366e+02, threshold=5.531e+02, percent-clipped=2.0 +2023-04-26 10:51:49,420 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2824.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:52:24,239 INFO [finetune.py:976] (3/7) Epoch 1, batch 2850, loss[loss=0.3873, simple_loss=0.381, pruned_loss=0.1967, over 4898.00 frames. ], tot_loss[loss=0.4293, simple_loss=0.4154, pruned_loss=0.2219, over 955454.64 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:52:26,215 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-26 10:53:07,971 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2885.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:53:29,244 INFO [finetune.py:976] (3/7) Epoch 1, batch 2900, loss[loss=0.4824, simple_loss=0.4672, pruned_loss=0.2488, over 4819.00 frames. ], tot_loss[loss=0.4302, simple_loss=0.4179, pruned_loss=0.2215, over 954938.80 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:53:46,013 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.613e+02 2.467e+02 2.930e+02 3.474e+02 6.951e+02, threshold=5.860e+02, percent-clipped=1.0 +2023-04-26 10:53:52,496 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.43 vs. limit=5.0 +2023-04-26 10:53:56,410 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6466, 2.1055, 1.5236, 2.1006, 1.4943, 1.5095, 2.0074, 1.3071], + device='cuda:3'), covar=tensor([0.2032, 0.1535, 0.1597, 0.1548, 0.3247, 0.1826, 0.1718, 0.3109], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0306, 0.0226, 0.0286, 0.0299, 0.0261, 0.0261, 0.0276], + device='cuda:3'), out_proj_covar=tensor([1.1958e-04, 1.2538e-04, 9.2741e-05, 1.1560e-04, 1.2418e-04, 1.0574e-04, + 1.0835e-04, 1.1247e-04], device='cuda:3') +2023-04-26 10:53:59,810 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2937.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:54:08,444 INFO [finetune.py:976] (3/7) Epoch 1, batch 2950, loss[loss=0.4613, simple_loss=0.4513, pruned_loss=0.2357, over 4923.00 frames. ], tot_loss[loss=0.4286, simple_loss=0.419, pruned_loss=0.2193, over 955803.43 frames. ], batch size: 42, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:54:17,538 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.16 vs. limit=5.0 +2023-04-26 10:54:22,089 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6507, 1.9382, 1.4973, 1.9615, 1.5127, 1.5218, 1.8139, 1.4074], + device='cuda:3'), covar=tensor([0.1649, 0.1158, 0.1285, 0.1082, 0.2270, 0.1394, 0.1533, 0.2210], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0305, 0.0225, 0.0285, 0.0298, 0.0260, 0.0261, 0.0275], + device='cuda:3'), out_proj_covar=tensor([1.1910e-04, 1.2505e-04, 9.2404e-05, 1.1528e-04, 1.2391e-04, 1.0540e-04, + 1.0803e-04, 1.1199e-04], device='cuda:3') +2023-04-26 10:54:31,045 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9483, 2.8488, 2.3787, 3.3096, 2.9337, 2.8881, 1.2575, 2.8032], + device='cuda:3'), covar=tensor([0.1727, 0.1197, 0.2705, 0.2193, 0.2531, 0.1810, 0.5002, 0.2340], + device='cuda:3'), in_proj_covar=tensor([0.0260, 0.0232, 0.0282, 0.0325, 0.0320, 0.0270, 0.0282, 0.0283], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:54:33,990 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=7.99 vs. limit=5.0 +2023-04-26 10:54:37,959 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2996.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:54:39,142 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2998.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:54:40,894 INFO [finetune.py:976] (3/7) Epoch 1, batch 3000, loss[loss=0.4142, simple_loss=0.4086, pruned_loss=0.2099, over 4731.00 frames. ], tot_loss[loss=0.4256, simple_loss=0.418, pruned_loss=0.2167, over 953089.82 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:54:40,895 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 10:54:44,276 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4999, 1.3987, 3.8995, 3.6317, 3.5157, 3.6849, 3.8392, 3.5086], + device='cuda:3'), covar=tensor([0.6546, 0.5602, 0.1157, 0.1826, 0.1398, 0.1571, 0.0992, 0.1555], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0320, 0.0460, 0.0469, 0.0386, 0.0440, 0.0352, 0.0410], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 10:54:48,742 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4868, 1.4367, 1.6431, 1.8367, 1.8564, 1.3365, 0.9853, 1.6634], + device='cuda:3'), covar=tensor([0.1177, 0.1385, 0.0946, 0.0889, 0.0772, 0.1410, 0.1379, 0.0794], + device='cuda:3'), in_proj_covar=tensor([0.0203, 0.0209, 0.0189, 0.0173, 0.0172, 0.0189, 0.0167, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 10:54:48,871 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0366, 0.7401, 0.8208, 0.9091, 0.7691, 0.7432, 0.5330, 0.5990], + device='cuda:3'), covar=tensor([ 7.8251, 9.6581, 4.2682, 14.5803, 11.2061, 8.2109, 10.6995, 11.0226], + device='cuda:3'), in_proj_covar=tensor([0.0267, 0.0286, 0.0226, 0.0354, 0.0247, 0.0236, 0.0281, 0.0233], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 10:54:51,385 INFO [finetune.py:1010] (3/7) Epoch 1, validation: loss=0.4217, simple_loss=0.4614, pruned_loss=0.191, over 2265189.00 frames. +2023-04-26 10:54:51,386 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 5768MB +2023-04-26 10:55:01,574 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.562e+02 2.517e+02 2.904e+02 3.818e+02 1.122e+03, threshold=5.808e+02, percent-clipped=2.0 +2023-04-26 10:55:02,291 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3019.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:55:16,258 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3041.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:55:18,055 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3044.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:55:23,634 INFO [finetune.py:976] (3/7) Epoch 1, batch 3050, loss[loss=0.3507, simple_loss=0.3608, pruned_loss=0.1703, over 4735.00 frames. ], tot_loss[loss=0.4199, simple_loss=0.4152, pruned_loss=0.2124, over 953441.62 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:55:42,240 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3080.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:55:42,864 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0721, 0.7356, 0.8151, 0.6508, 1.1522, 1.0040, 0.8625, 0.9164], + device='cuda:3'), covar=tensor([0.1130, 0.1712, 0.2247, 0.1613, 0.0833, 0.1249, 0.1428, 0.1685], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0358, 0.0356, 0.0318, 0.0370, 0.0392, 0.0340, 0.0378], + device='cuda:3'), out_proj_covar=tensor([7.5612e-05, 7.7117e-05, 7.6788e-05, 6.6344e-05, 7.8918e-05, 8.5881e-05, + 7.4155e-05, 8.1737e-05], device='cuda:3') +2023-04-26 10:55:45,842 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7506, 0.6306, 0.7490, 1.3060, 1.1054, 0.8257, 0.8394, 0.8646], + device='cuda:3'), covar=tensor([18.7425, 29.4004, 26.9399, 18.3214, 23.2865, 32.9897, 37.0616, 20.7454], + device='cuda:3'), in_proj_covar=tensor([0.0473, 0.0540, 0.0615, 0.0584, 0.0513, 0.0581, 0.0592, 0.0592], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 10:55:56,456 INFO [finetune.py:976] (3/7) Epoch 1, batch 3100, loss[loss=0.4867, simple_loss=0.4445, pruned_loss=0.2644, over 4929.00 frames. ], tot_loss[loss=0.4118, simple_loss=0.4099, pruned_loss=0.207, over 955841.36 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:56:12,966 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.494e+02 2.339e+02 2.765e+02 3.275e+02 7.045e+02, threshold=5.531e+02, percent-clipped=2.0 +2023-04-26 10:56:55,554 INFO [finetune.py:976] (3/7) Epoch 1, batch 3150, loss[loss=0.3543, simple_loss=0.3686, pruned_loss=0.17, over 4860.00 frames. ], tot_loss[loss=0.4037, simple_loss=0.4036, pruned_loss=0.202, over 956384.79 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:57:30,078 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4206, 2.3894, 1.9955, 2.3281, 2.5946, 2.0857, 3.3049, 1.9126], + device='cuda:3'), covar=tensor([0.4222, 0.1365, 0.3098, 0.2135, 0.1855, 0.2421, 0.0780, 0.3005], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0316, 0.0386, 0.0329, 0.0364, 0.0339, 0.0354, 0.0367], + device='cuda:3'), out_proj_covar=tensor([9.4690e-05, 9.7693e-05, 1.1951e-04, 1.0275e-04, 1.1156e-04, 1.0364e-04, + 1.0709e-04, 1.1401e-04], device='cuda:3') +2023-04-26 10:57:32,984 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3180.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:57:46,515 INFO [finetune.py:976] (3/7) Epoch 1, batch 3200, loss[loss=0.3509, simple_loss=0.3625, pruned_loss=0.1696, over 4778.00 frames. ], tot_loss[loss=0.3927, simple_loss=0.3949, pruned_loss=0.1953, over 953987.81 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:58:16,454 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.736e+02 2.453e+02 2.807e+02 3.222e+02 7.994e+02, threshold=5.615e+02, percent-clipped=2.0 +2023-04-26 10:58:42,948 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-26 10:58:52,915 INFO [finetune.py:976] (3/7) Epoch 1, batch 3250, loss[loss=0.3328, simple_loss=0.3487, pruned_loss=0.1585, over 4164.00 frames. ], tot_loss[loss=0.3887, simple_loss=0.3922, pruned_loss=0.1927, over 953642.64 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:59:47,338 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3288.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:59:50,301 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3293.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:00:00,688 INFO [finetune.py:976] (3/7) Epoch 1, batch 3300, loss[loss=0.3627, simple_loss=0.3878, pruned_loss=0.1688, over 4753.00 frames. ], tot_loss[loss=0.3915, simple_loss=0.3963, pruned_loss=0.1934, over 952508.85 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:00:19,087 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.582e+02 2.400e+02 2.649e+02 3.193e+02 6.501e+02, threshold=5.297e+02, percent-clipped=1.0 +2023-04-26 11:00:33,904 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3341.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:00:38,726 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3349.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 11:00:39,797 INFO [finetune.py:976] (3/7) Epoch 1, batch 3350, loss[loss=0.3432, simple_loss=0.3576, pruned_loss=0.1645, over 4746.00 frames. ], tot_loss[loss=0.3909, simple_loss=0.3971, pruned_loss=0.1924, over 950105.00 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:00:55,377 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-04-26 11:00:56,443 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 11:00:57,398 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3375.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:01:01,643 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7126, 1.6591, 1.5274, 1.4028, 1.8480, 1.3948, 2.2573, 1.3097], + device='cuda:3'), covar=tensor([0.3617, 0.1185, 0.3580, 0.2277, 0.1335, 0.1974, 0.0749, 0.3690], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0321, 0.0392, 0.0333, 0.0367, 0.0343, 0.0360, 0.0373], + device='cuda:3'), out_proj_covar=tensor([9.5614e-05, 9.9111e-05, 1.2132e-04, 1.0422e-04, 1.1241e-04, 1.0487e-04, + 1.0886e-04, 1.1560e-04], device='cuda:3') +2023-04-26 11:01:05,772 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3389.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:01:13,016 INFO [finetune.py:976] (3/7) Epoch 1, batch 3400, loss[loss=0.3748, simple_loss=0.3946, pruned_loss=0.1775, over 4821.00 frames. ], tot_loss[loss=0.3886, simple_loss=0.3968, pruned_loss=0.1902, over 951571.93 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:01:23,593 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.622e+02 2.298e+02 2.697e+02 3.198e+02 5.981e+02, threshold=5.394e+02, percent-clipped=4.0 +2023-04-26 11:01:57,865 INFO [finetune.py:976] (3/7) Epoch 1, batch 3450, loss[loss=0.3172, simple_loss=0.3516, pruned_loss=0.1414, over 4803.00 frames. ], tot_loss[loss=0.3864, simple_loss=0.3955, pruned_loss=0.1887, over 953015.42 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:02:20,931 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.3943, 1.3530, 1.3640, 1.1577, 1.5062, 1.1014, 1.7831, 1.2148], + device='cuda:3'), covar=tensor([0.3394, 0.1357, 0.3827, 0.2065, 0.1209, 0.1848, 0.1002, 0.3359], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0322, 0.0393, 0.0334, 0.0368, 0.0344, 0.0362, 0.0374], + device='cuda:3'), out_proj_covar=tensor([9.6033e-05, 9.9374e-05, 1.2176e-04, 1.0453e-04, 1.1277e-04, 1.0526e-04, + 1.0939e-04, 1.1594e-04], device='cuda:3') +2023-04-26 11:02:35,986 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3480.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:02:49,059 INFO [finetune.py:976] (3/7) Epoch 1, batch 3500, loss[loss=0.4387, simple_loss=0.421, pruned_loss=0.2282, over 4912.00 frames. ], tot_loss[loss=0.3806, simple_loss=0.3906, pruned_loss=0.1853, over 953512.73 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:02:54,179 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.03 vs. limit=5.0 +2023-04-26 11:02:55,922 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0217, 1.2805, 3.2953, 3.0685, 2.9521, 3.2127, 3.2246, 2.9007], + device='cuda:3'), covar=tensor([0.7049, 0.5514, 0.1506, 0.2104, 0.1494, 0.1721, 0.1506, 0.1793], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0320, 0.0459, 0.0469, 0.0387, 0.0440, 0.0351, 0.0409], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:03:00,467 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.505e+02 2.494e+02 2.871e+02 4.319e+02 1.287e+03, threshold=5.742e+02, percent-clipped=13.0 +2023-04-26 11:03:00,569 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3170, 1.3526, 3.8519, 3.5665, 3.3955, 3.6720, 3.7453, 3.3636], + device='cuda:3'), covar=tensor([0.7217, 0.6071, 0.1231, 0.2073, 0.1382, 0.1547, 0.1350, 0.1727], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0320, 0.0458, 0.0469, 0.0387, 0.0440, 0.0351, 0.0410], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:03:07,525 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3528.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:03:24,389 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3546.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:03:27,269 INFO [finetune.py:976] (3/7) Epoch 1, batch 3550, loss[loss=0.3395, simple_loss=0.3573, pruned_loss=0.1608, over 4695.00 frames. ], tot_loss[loss=0.3775, simple_loss=0.3868, pruned_loss=0.1841, over 952324.19 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:03:39,887 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-04-26 11:04:11,261 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3593.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:04:16,044 INFO [finetune.py:976] (3/7) Epoch 1, batch 3600, loss[loss=0.3419, simple_loss=0.3687, pruned_loss=0.1576, over 4822.00 frames. ], tot_loss[loss=0.3703, simple_loss=0.3813, pruned_loss=0.1796, over 953095.92 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:04:19,839 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3607.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 11:04:26,345 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.597e+02 2.647e+02 3.095e+02 3.748e+02 7.550e+02, threshold=6.190e+02, percent-clipped=4.0 +2023-04-26 11:04:37,145 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.74 vs. limit=5.0 +2023-04-26 11:04:42,910 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3641.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:04:44,771 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3644.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 11:04:48,958 INFO [finetune.py:976] (3/7) Epoch 1, batch 3650, loss[loss=0.4127, simple_loss=0.426, pruned_loss=0.1997, over 4819.00 frames. ], tot_loss[loss=0.3714, simple_loss=0.3839, pruned_loss=0.1795, over 953777.17 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:05:04,827 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3675.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:05:05,436 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3676.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:05:23,024 INFO [finetune.py:976] (3/7) Epoch 1, batch 3700, loss[loss=0.3907, simple_loss=0.4099, pruned_loss=0.1857, over 4933.00 frames. ], tot_loss[loss=0.3725, simple_loss=0.3864, pruned_loss=0.1793, over 953426.40 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:05:29,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5188, 2.2256, 1.3069, 1.2374, 1.0105, 1.0986, 1.2489, 0.9018], + device='cuda:3'), covar=tensor([0.3153, 0.2983, 0.4041, 0.5039, 0.4984, 0.3839, 0.3285, 0.4421], + device='cuda:3'), in_proj_covar=tensor([0.0201, 0.0224, 0.0205, 0.0220, 0.0241, 0.0201, 0.0197, 0.0214], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:05:44,466 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.676e+02 2.590e+02 3.040e+02 3.768e+02 5.314e+02, threshold=6.080e+02, percent-clipped=0.0 +2023-04-26 11:05:53,053 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3723.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:06:12,551 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3737.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 11:06:21,955 INFO [finetune.py:976] (3/7) Epoch 1, batch 3750, loss[loss=0.3524, simple_loss=0.377, pruned_loss=0.164, over 4820.00 frames. ], tot_loss[loss=0.3706, simple_loss=0.3866, pruned_loss=0.1773, over 953622.07 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:06:32,400 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9946, 2.3870, 1.0019, 1.3387, 1.6726, 1.2050, 3.0664, 1.6778], + device='cuda:3'), covar=tensor([0.0725, 0.0622, 0.0887, 0.1133, 0.0543, 0.1039, 0.0195, 0.0617], + device='cuda:3'), in_proj_covar=tensor([0.0057, 0.0073, 0.0053, 0.0050, 0.0055, 0.0056, 0.0089, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0012, 0.0008], + device='cuda:3') +2023-04-26 11:06:41,925 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9772, 0.9095, 1.1955, 1.1118, 0.9670, 0.8115, 0.9359, 0.5927], + device='cuda:3'), covar=tensor([0.0912, 0.0888, 0.0667, 0.0744, 0.0934, 0.1573, 0.0748, 0.1459], + device='cuda:3'), in_proj_covar=tensor([0.0076, 0.0081, 0.0077, 0.0079, 0.0095, 0.0098, 0.0095, 0.0083], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:3') +2023-04-26 11:07:07,658 INFO [finetune.py:976] (3/7) Epoch 1, batch 3800, loss[loss=0.37, simple_loss=0.4014, pruned_loss=0.1693, over 4900.00 frames. ], tot_loss[loss=0.3719, simple_loss=0.3879, pruned_loss=0.1779, over 953784.76 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:07:29,606 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.740e+02 2.601e+02 3.098e+02 3.867e+02 7.221e+02, threshold=6.196e+02, percent-clipped=5.0 +2023-04-26 11:07:43,475 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9236, 2.2862, 0.9440, 1.2642, 1.4366, 1.2136, 2.5757, 1.4645], + device='cuda:3'), covar=tensor([0.0715, 0.0500, 0.0723, 0.1198, 0.0521, 0.0992, 0.0229, 0.0656], + device='cuda:3'), in_proj_covar=tensor([0.0058, 0.0073, 0.0053, 0.0050, 0.0055, 0.0056, 0.0089, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0012, 0.0008], + device='cuda:3') +2023-04-26 11:08:14,577 INFO [finetune.py:976] (3/7) Epoch 1, batch 3850, loss[loss=0.3109, simple_loss=0.3463, pruned_loss=0.1377, over 4896.00 frames. ], tot_loss[loss=0.3653, simple_loss=0.383, pruned_loss=0.1738, over 954630.63 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:08:26,067 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9422, 2.2087, 0.9686, 1.2506, 1.4395, 1.1779, 2.5449, 1.4448], + device='cuda:3'), covar=tensor([0.0718, 0.0518, 0.0682, 0.1114, 0.0508, 0.1010, 0.0245, 0.0674], + device='cuda:3'), in_proj_covar=tensor([0.0057, 0.0073, 0.0053, 0.0050, 0.0055, 0.0056, 0.0089, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0012, 0.0008], + device='cuda:3') +2023-04-26 11:08:27,592 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 11:09:20,593 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6943, 1.3245, 4.1693, 3.8831, 3.6812, 3.8271, 3.7691, 3.6921], + device='cuda:3'), covar=tensor([0.5888, 0.5577, 0.0915, 0.1587, 0.1026, 0.1298, 0.2028, 0.1393], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0319, 0.0455, 0.0465, 0.0382, 0.0438, 0.0346, 0.0407], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:09:22,875 INFO [finetune.py:976] (3/7) Epoch 1, batch 3900, loss[loss=0.3098, simple_loss=0.348, pruned_loss=0.1358, over 4936.00 frames. ], tot_loss[loss=0.3611, simple_loss=0.3787, pruned_loss=0.1718, over 953439.23 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:09:29,502 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3902.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 11:09:44,792 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.635e+02 2.527e+02 3.007e+02 3.749e+02 8.787e+02, threshold=6.015e+02, percent-clipped=2.0 +2023-04-26 11:09:49,804 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3926.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:09:55,405 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0841, 0.3307, 0.6834, 0.7458, 0.8293, 0.8895, 0.5312, 0.6238], + device='cuda:3'), covar=tensor([ 6.6599, 24.3899, 14.5193, 12.4006, 12.4736, 16.7086, 19.4979, 14.3195], + device='cuda:3'), in_proj_covar=tensor([0.0257, 0.0368, 0.0299, 0.0296, 0.0326, 0.0337, 0.0358, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:10:02,243 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3944.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:10:07,403 INFO [finetune.py:976] (3/7) Epoch 1, batch 3950, loss[loss=0.2729, simple_loss=0.3022, pruned_loss=0.1218, over 4827.00 frames. ], tot_loss[loss=0.3528, simple_loss=0.3717, pruned_loss=0.167, over 953825.68 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:10:16,492 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.4819, 0.5242, 0.6268, 0.6987, 0.5944, 0.5346, 0.5310, 0.5853], + device='cuda:3'), covar=tensor([13.9526, 20.2536, 18.8180, 15.3976, 20.1395, 24.5942, 25.3539, 14.7225], + device='cuda:3'), in_proj_covar=tensor([0.0427, 0.0488, 0.0556, 0.0530, 0.0457, 0.0516, 0.0523, 0.0530], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:10:22,654 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.24 vs. limit=5.0 +2023-04-26 11:10:30,554 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3987.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:10:34,041 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3992.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:10:42,149 INFO [finetune.py:976] (3/7) Epoch 1, batch 4000, loss[loss=0.3039, simple_loss=0.3527, pruned_loss=0.1275, over 4816.00 frames. ], tot_loss[loss=0.3503, simple_loss=0.3698, pruned_loss=0.1654, over 955038.97 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:10:50,595 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-26 11:10:50,808 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-26 11:10:54,137 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.677e+02 2.386e+02 2.816e+02 3.337e+02 7.046e+02, threshold=5.633e+02, percent-clipped=3.0 +2023-04-26 11:11:02,814 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4032.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 11:11:15,866 INFO [finetune.py:976] (3/7) Epoch 1, batch 4050, loss[loss=0.3315, simple_loss=0.3787, pruned_loss=0.1422, over 4902.00 frames. ], tot_loss[loss=0.3489, simple_loss=0.3692, pruned_loss=0.1643, over 952809.47 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:11:49,877 INFO [finetune.py:976] (3/7) Epoch 1, batch 4100, loss[loss=0.3323, simple_loss=0.3627, pruned_loss=0.1509, over 4858.00 frames. ], tot_loss[loss=0.3492, simple_loss=0.3708, pruned_loss=0.1638, over 953120.69 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:11:55,878 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.8393, 4.7721, 3.3782, 5.4017, 4.7546, 4.7353, 2.6484, 4.6277], + device='cuda:3'), covar=tensor([0.1058, 0.0739, 0.2515, 0.0778, 0.2834, 0.1488, 0.4173, 0.1776], + device='cuda:3'), in_proj_covar=tensor([0.0261, 0.0232, 0.0283, 0.0328, 0.0322, 0.0271, 0.0285, 0.0286], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:12:08,253 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.683e+02 2.322e+02 2.767e+02 3.338e+02 6.077e+02, threshold=5.534e+02, percent-clipped=1.0 +2023-04-26 11:12:34,762 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4141.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:12:41,353 INFO [finetune.py:976] (3/7) Epoch 1, batch 4150, loss[loss=0.358, simple_loss=0.3838, pruned_loss=0.1661, over 4893.00 frames. ], tot_loss[loss=0.3487, simple_loss=0.3716, pruned_loss=0.1629, over 954160.93 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:13:15,084 INFO [finetune.py:976] (3/7) Epoch 1, batch 4200, loss[loss=0.2934, simple_loss=0.3238, pruned_loss=0.1315, over 4704.00 frames. ], tot_loss[loss=0.3461, simple_loss=0.3704, pruned_loss=0.1609, over 952834.63 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:13:16,204 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4202.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:13:16,223 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4202.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:13:28,643 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.716e+02 2.286e+02 2.811e+02 3.257e+02 1.063e+03, threshold=5.622e+02, percent-clipped=1.0 +2023-04-26 11:14:05,467 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4250.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:14:05,988 INFO [finetune.py:976] (3/7) Epoch 1, batch 4250, loss[loss=0.3317, simple_loss=0.3497, pruned_loss=0.1569, over 4825.00 frames. ], tot_loss[loss=0.3409, simple_loss=0.3657, pruned_loss=0.1581, over 951946.43 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:14:49,631 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 11:14:51,198 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4282.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:14:59,317 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4288.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:15:02,405 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3432, 3.3324, 0.9632, 1.7850, 1.8899, 2.3335, 2.1077, 1.0376], + device='cuda:3'), covar=tensor([0.1397, 0.0741, 0.2030, 0.1333, 0.1019, 0.1033, 0.1328, 0.1710], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0270, 0.0149, 0.0131, 0.0143, 0.0166, 0.0129, 0.0133], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:15:07,278 INFO [finetune.py:976] (3/7) Epoch 1, batch 4300, loss[loss=0.3219, simple_loss=0.3444, pruned_loss=0.1497, over 4719.00 frames. ], tot_loss[loss=0.3361, simple_loss=0.3614, pruned_loss=0.1553, over 951198.95 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:15:12,243 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 11:15:16,905 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6096, 2.2020, 1.4580, 1.3604, 1.1795, 1.2465, 1.4219, 1.0764], + device='cuda:3'), covar=tensor([0.2226, 0.2124, 0.3013, 0.3464, 0.3913, 0.2890, 0.2272, 0.3233], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0220, 0.0200, 0.0216, 0.0236, 0.0198, 0.0193, 0.0210], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:15:18,981 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-26 11:15:19,787 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.630e+02 2.200e+02 2.676e+02 3.122e+02 6.239e+02, threshold=5.353e+02, percent-clipped=1.0 +2023-04-26 11:15:29,484 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4332.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 11:15:39,942 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4349.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:15:41,059 INFO [finetune.py:976] (3/7) Epoch 1, batch 4350, loss[loss=0.2942, simple_loss=0.3254, pruned_loss=0.1315, over 4820.00 frames. ], tot_loss[loss=0.3316, simple_loss=0.357, pruned_loss=0.1531, over 952155.97 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:16:02,409 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4380.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:16:15,320 INFO [finetune.py:976] (3/7) Epoch 1, batch 4400, loss[loss=0.3702, simple_loss=0.3993, pruned_loss=0.1706, over 4818.00 frames. ], tot_loss[loss=0.332, simple_loss=0.3576, pruned_loss=0.1532, over 953243.74 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:16:26,704 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.735e+02 2.527e+02 2.839e+02 3.555e+02 1.567e+03, threshold=5.678e+02, percent-clipped=5.0 +2023-04-26 11:16:39,833 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7693, 2.1140, 1.7475, 2.0590, 1.6689, 1.7490, 2.0187, 1.5856], + device='cuda:3'), covar=tensor([0.1867, 0.1284, 0.1191, 0.1193, 0.2325, 0.1277, 0.1518, 0.2347], + device='cuda:3'), in_proj_covar=tensor([0.0304, 0.0321, 0.0237, 0.0300, 0.0306, 0.0275, 0.0272, 0.0291], + device='cuda:3'), out_proj_covar=tensor([1.2448e-04, 1.3155e-04, 9.7316e-05, 1.2139e-04, 1.2676e-04, 1.1107e-04, + 1.1299e-04, 1.1873e-04], device='cuda:3') +2023-04-26 11:16:48,906 INFO [finetune.py:976] (3/7) Epoch 1, batch 4450, loss[loss=0.3356, simple_loss=0.3717, pruned_loss=0.1498, over 4831.00 frames. ], tot_loss[loss=0.3343, simple_loss=0.3609, pruned_loss=0.1539, over 953595.06 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:16:53,300 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4918, 1.1352, 1.1446, 0.9943, 1.5953, 1.4313, 1.1252, 1.2330], + device='cuda:3'), covar=tensor([0.1297, 0.1988, 0.2618, 0.2533, 0.1133, 0.1771, 0.2119, 0.2344], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0345, 0.0348, 0.0313, 0.0357, 0.0381, 0.0327, 0.0365], + device='cuda:3'), out_proj_covar=tensor([7.2854e-05, 7.4322e-05, 7.5325e-05, 6.5594e-05, 7.6030e-05, 8.3306e-05, + 7.1274e-05, 7.8906e-05], device='cuda:3') +2023-04-26 11:17:19,615 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4497.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:17:22,029 INFO [finetune.py:976] (3/7) Epoch 1, batch 4500, loss[loss=0.2617, simple_loss=0.302, pruned_loss=0.1107, over 4735.00 frames. ], tot_loss[loss=0.3352, simple_loss=0.3627, pruned_loss=0.1539, over 953828.72 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:17:32,938 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.661e+02 2.259e+02 2.800e+02 3.318e+02 7.116e+02, threshold=5.601e+02, percent-clipped=2.0 +2023-04-26 11:18:15,788 INFO [finetune.py:976] (3/7) Epoch 1, batch 4550, loss[loss=0.3035, simple_loss=0.3662, pruned_loss=0.1204, over 4811.00 frames. ], tot_loss[loss=0.3348, simple_loss=0.363, pruned_loss=0.1532, over 953591.04 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:18:36,988 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-26 11:19:01,291 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4582.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:19:13,032 INFO [finetune.py:976] (3/7) Epoch 1, batch 4600, loss[loss=0.2764, simple_loss=0.3013, pruned_loss=0.1257, over 4738.00 frames. ], tot_loss[loss=0.3307, simple_loss=0.3597, pruned_loss=0.1509, over 953460.56 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:19:23,472 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.373e+02 2.361e+02 2.771e+02 3.480e+02 8.913e+02, threshold=5.542e+02, percent-clipped=3.0 +2023-04-26 11:19:33,262 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4630.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:19:42,843 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4644.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:19:47,130 INFO [finetune.py:976] (3/7) Epoch 1, batch 4650, loss[loss=0.2493, simple_loss=0.2928, pruned_loss=0.1029, over 4807.00 frames. ], tot_loss[loss=0.3239, simple_loss=0.3534, pruned_loss=0.1472, over 953104.45 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:20:20,046 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8257, 1.3436, 4.4761, 4.1210, 3.9086, 4.2056, 4.2048, 3.9254], + device='cuda:3'), covar=tensor([0.6738, 0.6321, 0.1045, 0.1900, 0.1165, 0.1351, 0.1206, 0.1445], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0321, 0.0457, 0.0465, 0.0385, 0.0441, 0.0347, 0.0411], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:20:35,057 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1996, 1.7048, 1.4142, 1.8509, 1.7267, 1.9934, 1.5631, 3.5937], + device='cuda:3'), covar=tensor([0.0699, 0.0710, 0.0815, 0.1274, 0.0641, 0.0561, 0.0703, 0.0163], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0041, 0.0041, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 11:20:42,427 INFO [finetune.py:976] (3/7) Epoch 1, batch 4700, loss[loss=0.2141, simple_loss=0.2749, pruned_loss=0.0766, over 4824.00 frames. ], tot_loss[loss=0.3183, simple_loss=0.3487, pruned_loss=0.1439, over 955588.13 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:21:03,818 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.981e+02 2.474e+02 3.057e+02 6.452e+02, threshold=4.948e+02, percent-clipped=2.0 +2023-04-26 11:21:39,264 INFO [finetune.py:976] (3/7) Epoch 1, batch 4750, loss[loss=0.2712, simple_loss=0.305, pruned_loss=0.1187, over 4749.00 frames. ], tot_loss[loss=0.3165, simple_loss=0.347, pruned_loss=0.143, over 957956.53 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:21:47,130 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.11 vs. limit=2.0 +2023-04-26 11:21:59,555 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 11:22:11,016 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4797.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:22:13,545 INFO [finetune.py:976] (3/7) Epoch 1, batch 4800, loss[loss=0.3342, simple_loss=0.3694, pruned_loss=0.1495, over 4933.00 frames. ], tot_loss[loss=0.3202, simple_loss=0.3503, pruned_loss=0.1451, over 955424.11 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:22:15,132 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-26 11:22:24,011 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.522e+02 2.397e+02 2.799e+02 3.310e+02 5.685e+02, threshold=5.598e+02, percent-clipped=2.0 +2023-04-26 11:22:43,636 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4845.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:22:47,739 INFO [finetune.py:976] (3/7) Epoch 1, batch 4850, loss[loss=0.2578, simple_loss=0.2788, pruned_loss=0.1184, over 4419.00 frames. ], tot_loss[loss=0.3232, simple_loss=0.3543, pruned_loss=0.146, over 956708.65 frames. ], batch size: 19, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:22:54,700 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5626, 1.5748, 0.7468, 1.2684, 1.8064, 1.4663, 1.3153, 1.4563], + device='cuda:3'), covar=tensor([0.0629, 0.0496, 0.0546, 0.0664, 0.0378, 0.0644, 0.0616, 0.0743], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0043, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 11:23:09,357 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4838, 1.1420, 1.1048, 1.0858, 1.7177, 1.5108, 1.1664, 1.2095], + device='cuda:3'), covar=tensor([0.1348, 0.1822, 0.2460, 0.2080, 0.0868, 0.1401, 0.1947, 0.1932], + device='cuda:3'), in_proj_covar=tensor([0.0333, 0.0346, 0.0350, 0.0315, 0.0357, 0.0381, 0.0328, 0.0364], + device='cuda:3'), out_proj_covar=tensor([7.2924e-05, 7.4460e-05, 7.5723e-05, 6.5956e-05, 7.6194e-05, 8.3549e-05, + 7.1612e-05, 7.8826e-05], device='cuda:3') +2023-04-26 11:23:33,605 INFO [finetune.py:976] (3/7) Epoch 1, batch 4900, loss[loss=0.3247, simple_loss=0.3491, pruned_loss=0.1502, over 4816.00 frames. ], tot_loss[loss=0.3231, simple_loss=0.3546, pruned_loss=0.1458, over 954252.56 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:23:46,701 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-04-26 11:23:49,557 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.488e+02 2.175e+02 2.572e+02 3.151e+02 5.494e+02, threshold=5.143e+02, percent-clipped=0.0 +2023-04-26 11:24:22,584 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 11:24:29,536 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4944.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:24:41,249 INFO [finetune.py:976] (3/7) Epoch 1, batch 4950, loss[loss=0.3106, simple_loss=0.3342, pruned_loss=0.1435, over 4763.00 frames. ], tot_loss[loss=0.3228, simple_loss=0.3546, pruned_loss=0.1455, over 954353.52 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:25:18,996 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4992.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:25:26,585 INFO [finetune.py:976] (3/7) Epoch 1, batch 5000, loss[loss=0.2875, simple_loss=0.3175, pruned_loss=0.1287, over 4142.00 frames. ], tot_loss[loss=0.3188, simple_loss=0.3514, pruned_loss=0.1431, over 953236.65 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:25:37,464 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.472e+02 2.232e+02 2.673e+02 3.232e+02 7.029e+02, threshold=5.346e+02, percent-clipped=5.0 +2023-04-26 11:26:10,647 INFO [finetune.py:976] (3/7) Epoch 1, batch 5050, loss[loss=0.2827, simple_loss=0.3239, pruned_loss=0.1208, over 4937.00 frames. ], tot_loss[loss=0.3135, simple_loss=0.3464, pruned_loss=0.1403, over 954010.79 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:27:11,308 INFO [finetune.py:976] (3/7) Epoch 1, batch 5100, loss[loss=0.3124, simple_loss=0.3344, pruned_loss=0.1452, over 4032.00 frames. ], tot_loss[loss=0.3084, simple_loss=0.341, pruned_loss=0.1379, over 952215.73 frames. ], batch size: 17, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:27:40,805 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.449e+02 2.109e+02 2.481e+02 2.814e+02 6.642e+02, threshold=4.963e+02, percent-clipped=1.0 +2023-04-26 11:27:52,994 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 11:28:15,612 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7293, 0.3237, 0.5128, 1.0388, 0.9150, 0.6769, 0.6126, 0.6379], + device='cuda:3'), covar=tensor([12.3810, 16.3767, 15.0670, 14.7813, 13.5937, 16.8980, 16.6848, 9.9040], + device='cuda:3'), in_proj_covar=tensor([0.0425, 0.0488, 0.0561, 0.0535, 0.0451, 0.0511, 0.0516, 0.0524], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:28:17,911 INFO [finetune.py:976] (3/7) Epoch 1, batch 5150, loss[loss=0.2232, simple_loss=0.282, pruned_loss=0.08216, over 4762.00 frames. ], tot_loss[loss=0.3075, simple_loss=0.3401, pruned_loss=0.1374, over 953844.84 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:29:07,361 INFO [finetune.py:976] (3/7) Epoch 1, batch 5200, loss[loss=0.3154, simple_loss=0.3568, pruned_loss=0.1371, over 4828.00 frames. ], tot_loss[loss=0.3121, simple_loss=0.3457, pruned_loss=0.1393, over 955011.31 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:29:21,206 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.499e+02 2.511e+02 2.846e+02 3.705e+02 8.558e+02, threshold=5.692e+02, percent-clipped=9.0 +2023-04-26 11:29:33,738 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2620, 1.3082, 1.2670, 1.2779, 1.4115, 1.4864, 1.3489, 1.3357], + device='cuda:3'), covar=tensor([ 4.6038, 12.7228, 8.6796, 6.7447, 6.6629, 9.9831, 11.2198, 8.8522], + device='cuda:3'), in_proj_covar=tensor([0.0259, 0.0362, 0.0293, 0.0291, 0.0319, 0.0336, 0.0351, 0.0320], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:29:35,546 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3097, 1.6782, 1.5552, 1.9185, 1.7346, 2.0855, 1.6248, 3.2890], + device='cuda:3'), covar=tensor([0.0681, 0.0643, 0.0682, 0.1042, 0.0584, 0.0645, 0.0654, 0.0182], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0041, 0.0041, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 11:29:41,593 INFO [finetune.py:976] (3/7) Epoch 1, batch 5250, loss[loss=0.3065, simple_loss=0.3401, pruned_loss=0.1365, over 4889.00 frames. ], tot_loss[loss=0.3133, simple_loss=0.3476, pruned_loss=0.1395, over 955154.27 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:30:00,415 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8214, 1.5108, 2.2387, 2.3587, 1.6410, 1.2710, 1.7522, 1.0015], + device='cuda:3'), covar=tensor([0.1302, 0.1912, 0.0688, 0.0760, 0.1560, 0.1718, 0.1152, 0.1825], + device='cuda:3'), in_proj_covar=tensor([0.0073, 0.0080, 0.0077, 0.0077, 0.0091, 0.0096, 0.0093, 0.0082], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:3') +2023-04-26 11:30:00,436 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0155, 1.0766, 1.4425, 2.1513, 1.4974, 1.1400, 0.9517, 1.5147], + device='cuda:3'), covar=tensor([0.6978, 0.9555, 0.4319, 1.1806, 1.0793, 0.7179, 1.6670, 0.9885], + device='cuda:3'), in_proj_covar=tensor([0.0255, 0.0274, 0.0215, 0.0334, 0.0232, 0.0227, 0.0268, 0.0220], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:30:04,713 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5070, 1.2586, 1.1270, 1.0985, 1.7753, 1.4797, 1.1703, 1.1625], + device='cuda:3'), covar=tensor([0.1575, 0.1675, 0.2431, 0.2036, 0.0901, 0.1850, 0.2156, 0.2037], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0349, 0.0351, 0.0318, 0.0360, 0.0383, 0.0332, 0.0366], + device='cuda:3'), out_proj_covar=tensor([7.3207e-05, 7.5155e-05, 7.5787e-05, 6.6557e-05, 7.6818e-05, 8.4027e-05, + 7.2360e-05, 7.9213e-05], device='cuda:3') +2023-04-26 11:30:06,532 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4896, 1.1988, 0.6500, 1.1794, 1.4404, 1.3734, 1.2708, 1.2980], + device='cuda:3'), covar=tensor([0.0641, 0.0542, 0.0559, 0.0685, 0.0405, 0.0647, 0.0615, 0.0761], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 11:30:12,705 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4240, 1.5685, 1.7348, 1.7645, 1.8685, 1.4333, 1.0595, 1.5720], + device='cuda:3'), covar=tensor([0.1223, 0.1185, 0.0840, 0.0854, 0.0798, 0.1222, 0.1368, 0.0829], + device='cuda:3'), in_proj_covar=tensor([0.0207, 0.0208, 0.0189, 0.0180, 0.0176, 0.0195, 0.0173, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:30:15,016 INFO [finetune.py:976] (3/7) Epoch 1, batch 5300, loss[loss=0.2555, simple_loss=0.3119, pruned_loss=0.09951, over 4866.00 frames. ], tot_loss[loss=0.3133, simple_loss=0.3483, pruned_loss=0.1392, over 955054.29 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:30:27,274 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.471e+02 2.212e+02 2.624e+02 3.137e+02 5.522e+02, threshold=5.248e+02, percent-clipped=0.0 +2023-04-26 11:30:49,038 INFO [finetune.py:976] (3/7) Epoch 1, batch 5350, loss[loss=0.288, simple_loss=0.3328, pruned_loss=0.1216, over 4753.00 frames. ], tot_loss[loss=0.3103, simple_loss=0.3463, pruned_loss=0.1372, over 953647.65 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:31:04,574 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7975, 2.7800, 2.4265, 2.7327, 3.2112, 2.5913, 3.6677, 2.3037], + device='cuda:3'), covar=tensor([0.4472, 0.1602, 0.3656, 0.2527, 0.1315, 0.2510, 0.1023, 0.3699], + device='cuda:3'), in_proj_covar=tensor([0.0330, 0.0341, 0.0422, 0.0356, 0.0390, 0.0363, 0.0388, 0.0397], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:31:14,624 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8198, 2.1649, 1.0431, 1.5760, 2.3902, 1.7446, 1.5150, 1.7880], + device='cuda:3'), covar=tensor([0.0562, 0.0445, 0.0466, 0.0625, 0.0320, 0.0580, 0.0587, 0.0724], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 11:31:35,565 INFO [finetune.py:976] (3/7) Epoch 1, batch 5400, loss[loss=0.3079, simple_loss=0.3399, pruned_loss=0.1379, over 4883.00 frames. ], tot_loss[loss=0.3063, simple_loss=0.3418, pruned_loss=0.1354, over 953358.91 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:31:52,845 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.437e+02 2.058e+02 2.416e+02 3.015e+02 5.928e+02, threshold=4.831e+02, percent-clipped=1.0 +2023-04-26 11:32:15,539 INFO [finetune.py:976] (3/7) Epoch 1, batch 5450, loss[loss=0.2635, simple_loss=0.317, pruned_loss=0.105, over 4906.00 frames. ], tot_loss[loss=0.2999, simple_loss=0.3363, pruned_loss=0.1318, over 954390.84 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:32:24,899 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-26 11:32:49,240 INFO [finetune.py:976] (3/7) Epoch 1, batch 5500, loss[loss=0.3298, simple_loss=0.3526, pruned_loss=0.1535, over 4765.00 frames. ], tot_loss[loss=0.2957, simple_loss=0.3319, pruned_loss=0.1297, over 952626.10 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:32:58,628 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6188, 2.2880, 1.4520, 1.4774, 1.1579, 1.2283, 1.5455, 1.0656], + device='cuda:3'), covar=tensor([0.2563, 0.2475, 0.3170, 0.3736, 0.4288, 0.3195, 0.2382, 0.3492], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0222, 0.0199, 0.0217, 0.0236, 0.0198, 0.0193, 0.0209], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:32:59,728 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.389e+02 2.158e+02 2.515e+02 3.075e+02 5.534e+02, threshold=5.030e+02, percent-clipped=1.0 +2023-04-26 11:33:46,366 INFO [finetune.py:976] (3/7) Epoch 1, batch 5550, loss[loss=0.2762, simple_loss=0.3159, pruned_loss=0.1182, over 4774.00 frames. ], tot_loss[loss=0.2986, simple_loss=0.3347, pruned_loss=0.1312, over 953164.03 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:34:09,578 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4734, 2.2517, 1.3126, 1.4227, 1.0135, 1.1002, 1.4160, 0.9615], + device='cuda:3'), covar=tensor([0.2685, 0.2525, 0.3278, 0.3905, 0.4487, 0.3288, 0.2526, 0.3579], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0221, 0.0199, 0.0217, 0.0236, 0.0198, 0.0193, 0.0209], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:34:53,183 INFO [finetune.py:976] (3/7) Epoch 1, batch 5600, loss[loss=0.3142, simple_loss=0.3534, pruned_loss=0.1375, over 4767.00 frames. ], tot_loss[loss=0.3018, simple_loss=0.3393, pruned_loss=0.1321, over 954829.62 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:35:09,275 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0326, 1.4180, 1.2646, 1.6948, 1.4832, 1.8920, 1.3474, 3.2950], + device='cuda:3'), covar=tensor([0.0753, 0.0736, 0.0850, 0.1273, 0.0706, 0.0594, 0.0777, 0.0175], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0040, 0.0041, 0.0046, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 11:35:13,911 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.486e+02 2.261e+02 2.635e+02 3.434e+02 7.043e+02, threshold=5.269e+02, percent-clipped=4.0 +2023-04-26 11:35:39,241 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-04-26 11:35:44,916 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5650.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:35:45,421 INFO [finetune.py:976] (3/7) Epoch 1, batch 5650, loss[loss=0.3478, simple_loss=0.3841, pruned_loss=0.1558, over 4792.00 frames. ], tot_loss[loss=0.304, simple_loss=0.342, pruned_loss=0.133, over 953921.81 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:36:05,790 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1110, 1.5421, 1.8086, 2.0915, 1.7851, 1.4939, 1.3443, 1.6434], + device='cuda:3'), covar=tensor([1.0282, 0.8196, 0.3750, 1.1485, 0.8113, 0.6571, 1.0895, 0.8866], + device='cuda:3'), in_proj_covar=tensor([0.0257, 0.0276, 0.0217, 0.0337, 0.0234, 0.0229, 0.0270, 0.0222], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:36:12,390 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9702, 1.2585, 1.3162, 1.2600, 1.3135, 0.9988, 0.6629, 1.2021], + device='cuda:3'), covar=tensor([0.1324, 0.1457, 0.1090, 0.0971, 0.0894, 0.1215, 0.1485, 0.0850], + device='cuda:3'), in_proj_covar=tensor([0.0210, 0.0210, 0.0190, 0.0182, 0.0179, 0.0197, 0.0175, 0.0192], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:36:15,238 INFO [finetune.py:976] (3/7) Epoch 1, batch 5700, loss[loss=0.2604, simple_loss=0.2905, pruned_loss=0.1152, over 4319.00 frames. ], tot_loss[loss=0.3007, simple_loss=0.3373, pruned_loss=0.1321, over 935921.18 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:36:21,274 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5711.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:36:31,943 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.325e+02 1.970e+02 2.507e+02 3.043e+02 6.051e+02, threshold=5.014e+02, percent-clipped=1.0 +2023-04-26 11:37:01,404 INFO [finetune.py:976] (3/7) Epoch 2, batch 0, loss[loss=0.2529, simple_loss=0.3015, pruned_loss=0.1021, over 4753.00 frames. ], tot_loss[loss=0.2529, simple_loss=0.3015, pruned_loss=0.1021, over 4753.00 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:37:01,404 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 11:37:23,940 INFO [finetune.py:1010] (3/7) Epoch 2, validation: loss=0.2101, simple_loss=0.2777, pruned_loss=0.0712, over 2265189.00 frames. +2023-04-26 11:37:23,940 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6040MB +2023-04-26 11:37:46,574 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5755.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:37:55,024 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-26 11:38:02,080 INFO [finetune.py:976] (3/7) Epoch 2, batch 50, loss[loss=0.3135, simple_loss=0.3434, pruned_loss=0.1418, over 4763.00 frames. ], tot_loss[loss=0.2962, simple_loss=0.3348, pruned_loss=0.1288, over 216813.80 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:38:06,794 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9214, 1.3290, 1.6304, 1.9039, 1.5591, 1.3180, 1.1031, 1.4982], + device='cuda:3'), covar=tensor([0.9830, 0.8798, 0.4170, 1.1328, 0.8623, 0.6550, 1.1913, 0.9038], + device='cuda:3'), in_proj_covar=tensor([0.0258, 0.0277, 0.0218, 0.0339, 0.0235, 0.0230, 0.0271, 0.0222], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:38:17,937 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9127, 0.7808, 1.0261, 1.1753, 1.0835, 0.9401, 0.9259, 1.0046], + device='cuda:3'), covar=tensor([ 8.4129, 12.6576, 12.1650, 11.9841, 10.2637, 14.3897, 13.6937, 9.2659], + device='cuda:3'), in_proj_covar=tensor([0.0432, 0.0497, 0.0572, 0.0551, 0.0460, 0.0520, 0.0525, 0.0535], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:38:25,328 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6558, 3.5156, 1.1715, 2.2284, 2.0693, 2.5611, 2.5014, 1.3693], + device='cuda:3'), covar=tensor([0.1175, 0.0823, 0.1828, 0.1160, 0.1020, 0.0993, 0.1038, 0.2072], + device='cuda:3'), in_proj_covar=tensor([0.0124, 0.0272, 0.0151, 0.0132, 0.0143, 0.0166, 0.0130, 0.0132], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:38:27,683 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5816.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:38:29,285 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.431e+02 2.066e+02 2.410e+02 2.890e+02 4.929e+02, threshold=4.819e+02, percent-clipped=0.0 +2023-04-26 11:38:32,822 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0265, 1.2126, 5.1754, 4.8110, 4.5858, 4.8548, 4.5798, 4.5489], + device='cuda:3'), covar=tensor([0.6440, 0.6427, 0.0922, 0.1818, 0.1058, 0.1192, 0.1119, 0.1567], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0313, 0.0448, 0.0456, 0.0375, 0.0432, 0.0341, 0.0400], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:38:35,791 INFO [finetune.py:976] (3/7) Epoch 2, batch 100, loss[loss=0.2946, simple_loss=0.3355, pruned_loss=0.1268, over 4896.00 frames. ], tot_loss[loss=0.2969, simple_loss=0.3342, pruned_loss=0.1298, over 380540.01 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:38:55,167 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5858.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:39:09,189 INFO [finetune.py:976] (3/7) Epoch 2, batch 150, loss[loss=0.3039, simple_loss=0.3336, pruned_loss=0.1371, over 4935.00 frames. ], tot_loss[loss=0.2876, simple_loss=0.3259, pruned_loss=0.1246, over 507345.02 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:39:17,673 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8488, 1.4247, 4.6188, 4.3097, 4.0626, 4.2707, 4.2288, 4.0135], + device='cuda:3'), covar=tensor([0.6317, 0.6038, 0.1157, 0.1998, 0.1068, 0.1710, 0.1231, 0.1700], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0314, 0.0449, 0.0458, 0.0377, 0.0434, 0.0342, 0.0401], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:39:34,934 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.317e+02 2.049e+02 2.408e+02 2.939e+02 5.453e+02, threshold=4.816e+02, percent-clipped=1.0 +2023-04-26 11:39:35,705 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5919.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:39:39,696 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0607, 1.2235, 5.1452, 4.7453, 4.6002, 4.8235, 4.5878, 4.5761], + device='cuda:3'), covar=tensor([0.5992, 0.6470, 0.0901, 0.1867, 0.0943, 0.1511, 0.1281, 0.1248], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0314, 0.0450, 0.0459, 0.0378, 0.0436, 0.0343, 0.0402], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:39:48,163 INFO [finetune.py:976] (3/7) Epoch 2, batch 200, loss[loss=0.2392, simple_loss=0.3006, pruned_loss=0.08889, over 4840.00 frames. ], tot_loss[loss=0.2841, simple_loss=0.3224, pruned_loss=0.1229, over 609618.80 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:40:20,382 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0539, 2.4321, 1.1252, 1.4309, 1.8778, 1.3632, 2.8477, 1.5231], + device='cuda:3'), covar=tensor([0.0656, 0.0714, 0.0889, 0.1037, 0.0488, 0.0899, 0.0205, 0.0647], + device='cuda:3'), in_proj_covar=tensor([0.0058, 0.0074, 0.0055, 0.0051, 0.0057, 0.0057, 0.0089, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0012, 0.0008], + device='cuda:3') +2023-04-26 11:40:52,013 INFO [finetune.py:976] (3/7) Epoch 2, batch 250, loss[loss=0.3218, simple_loss=0.3633, pruned_loss=0.1402, over 4920.00 frames. ], tot_loss[loss=0.2876, simple_loss=0.3266, pruned_loss=0.1243, over 685767.59 frames. ], batch size: 42, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:41:01,556 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6170, 1.2635, 1.0988, 1.2978, 1.8224, 1.5790, 1.3844, 1.1041], + device='cuda:3'), covar=tensor([0.1522, 0.1877, 0.2280, 0.2327, 0.0853, 0.1747, 0.2132, 0.2106], + device='cuda:3'), in_proj_covar=tensor([0.0329, 0.0343, 0.0347, 0.0315, 0.0353, 0.0377, 0.0326, 0.0359], + device='cuda:3'), out_proj_covar=tensor([7.2026e-05, 7.4040e-05, 7.4851e-05, 6.6041e-05, 7.5382e-05, 8.2653e-05, + 7.0969e-05, 7.7633e-05], device='cuda:3') +2023-04-26 11:41:05,636 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4516, 2.9652, 0.9278, 1.6285, 2.1183, 1.5968, 4.3567, 1.9958], + device='cuda:3'), covar=tensor([0.0695, 0.0842, 0.1102, 0.1336, 0.0671, 0.1021, 0.0173, 0.0680], + device='cuda:3'), in_proj_covar=tensor([0.0058, 0.0074, 0.0055, 0.0051, 0.0057, 0.0057, 0.0089, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0012, 0.0008], + device='cuda:3') +2023-04-26 11:41:06,881 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5990.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:41:09,832 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.54 vs. limit=2.0 +2023-04-26 11:41:18,234 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6006.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:41:25,498 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.356e+02 2.171e+02 2.695e+02 3.137e+02 1.432e+03, threshold=5.390e+02, percent-clipped=4.0 +2023-04-26 11:41:30,163 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-26 11:41:33,100 INFO [finetune.py:976] (3/7) Epoch 2, batch 300, loss[loss=0.4162, simple_loss=0.4205, pruned_loss=0.2059, over 4919.00 frames. ], tot_loss[loss=0.2946, simple_loss=0.3334, pruned_loss=0.1278, over 744844.62 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:41:49,306 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6043.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:41:54,638 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6051.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:42:10,440 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5512, 2.1640, 1.4480, 1.3412, 1.1969, 1.2265, 1.4183, 1.0970], + device='cuda:3'), covar=tensor([0.2723, 0.2117, 0.3032, 0.3485, 0.4089, 0.3395, 0.2396, 0.3299], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0222, 0.0200, 0.0218, 0.0236, 0.0198, 0.0194, 0.0210], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:42:21,221 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0617, 2.0864, 2.4135, 2.5364, 2.4780, 1.8030, 1.5186, 2.0852], + device='cuda:3'), covar=tensor([0.1040, 0.0993, 0.0558, 0.0690, 0.0625, 0.1130, 0.1354, 0.0719], + device='cuda:3'), in_proj_covar=tensor([0.0209, 0.0210, 0.0189, 0.0182, 0.0178, 0.0196, 0.0175, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:42:22,355 INFO [finetune.py:976] (3/7) Epoch 2, batch 350, loss[loss=0.2465, simple_loss=0.3058, pruned_loss=0.09361, over 4831.00 frames. ], tot_loss[loss=0.2969, simple_loss=0.3363, pruned_loss=0.1287, over 790153.85 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:42:47,486 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6104.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:42:57,934 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6111.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:43:02,078 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.403e+02 2.085e+02 2.458e+02 2.985e+02 5.837e+02, threshold=4.917e+02, percent-clipped=2.0 +2023-04-26 11:43:02,862 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7390, 1.0233, 1.1037, 1.3539, 1.8082, 1.6096, 1.3744, 1.0758], + device='cuda:3'), covar=tensor([0.1313, 0.2222, 0.2286, 0.1803, 0.0982, 0.1568, 0.2007, 0.2159], + device='cuda:3'), in_proj_covar=tensor([0.0328, 0.0345, 0.0348, 0.0316, 0.0354, 0.0380, 0.0327, 0.0360], + device='cuda:3'), out_proj_covar=tensor([7.1940e-05, 7.4299e-05, 7.5046e-05, 6.6210e-05, 7.5486e-05, 8.3164e-05, + 7.1208e-05, 7.7816e-05], device='cuda:3') +2023-04-26 11:43:12,556 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7186, 1.3630, 1.2325, 1.2929, 1.9529, 1.6532, 1.3667, 1.2547], + device='cuda:3'), covar=tensor([0.1441, 0.1885, 0.2292, 0.2055, 0.0768, 0.1705, 0.2001, 0.2110], + device='cuda:3'), in_proj_covar=tensor([0.0329, 0.0345, 0.0348, 0.0316, 0.0354, 0.0380, 0.0327, 0.0360], + device='cuda:3'), out_proj_covar=tensor([7.1982e-05, 7.4310e-05, 7.5090e-05, 6.6269e-05, 7.5522e-05, 8.3172e-05, + 7.1251e-05, 7.7858e-05], device='cuda:3') +2023-04-26 11:43:14,254 INFO [finetune.py:976] (3/7) Epoch 2, batch 400, loss[loss=0.2601, simple_loss=0.3203, pruned_loss=0.09997, over 4808.00 frames. ], tot_loss[loss=0.2959, simple_loss=0.3363, pruned_loss=0.1277, over 826791.15 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:43:20,170 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6135.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:43:43,639 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6170.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:43:48,410 INFO [finetune.py:976] (3/7) Epoch 2, batch 450, loss[loss=0.2754, simple_loss=0.3155, pruned_loss=0.1176, over 4848.00 frames. ], tot_loss[loss=0.2935, simple_loss=0.3347, pruned_loss=0.1262, over 856440.08 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:44:00,945 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6196.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:06,766 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6203.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:11,649 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6211.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:13,447 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6214.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:15,816 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.476e+02 2.104e+02 2.551e+02 3.094e+02 4.755e+02, threshold=5.103e+02, percent-clipped=0.0 +2023-04-26 11:44:21,877 INFO [finetune.py:976] (3/7) Epoch 2, batch 500, loss[loss=0.2143, simple_loss=0.2719, pruned_loss=0.07832, over 4780.00 frames. ], tot_loss[loss=0.2874, simple_loss=0.3293, pruned_loss=0.1227, over 877956.46 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:44:23,802 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6231.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:46,833 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6264.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:48,037 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6266.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:51,699 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6272.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:55,243 INFO [finetune.py:976] (3/7) Epoch 2, batch 550, loss[loss=0.3081, simple_loss=0.3306, pruned_loss=0.1428, over 4708.00 frames. ], tot_loss[loss=0.2832, simple_loss=0.325, pruned_loss=0.1207, over 895451.06 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:45:14,994 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6306.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 11:45:28,233 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.445e+02 2.269e+02 2.692e+02 3.237e+02 6.108e+02, threshold=5.385e+02, percent-clipped=2.0 +2023-04-26 11:45:30,250 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7744, 1.1086, 1.4748, 1.8431, 1.4098, 1.1123, 0.9336, 1.3073], + device='cuda:3'), covar=tensor([0.6442, 0.8459, 0.3818, 1.0641, 0.9003, 0.6727, 1.2550, 0.8835], + device='cuda:3'), in_proj_covar=tensor([0.0262, 0.0279, 0.0221, 0.0343, 0.0237, 0.0233, 0.0274, 0.0223], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:45:37,837 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6327.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:45:38,337 INFO [finetune.py:976] (3/7) Epoch 2, batch 600, loss[loss=0.2813, simple_loss=0.3277, pruned_loss=0.1175, over 4833.00 frames. ], tot_loss[loss=0.2829, simple_loss=0.3248, pruned_loss=0.1205, over 910887.30 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:45:39,381 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 11:46:00,716 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6346.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:46:12,545 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6354.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:46:39,283 INFO [finetune.py:976] (3/7) Epoch 2, batch 650, loss[loss=0.3253, simple_loss=0.3644, pruned_loss=0.1432, over 4902.00 frames. ], tot_loss[loss=0.2862, simple_loss=0.328, pruned_loss=0.1222, over 921934.39 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:46:41,219 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4952, 1.2456, 1.1085, 1.1776, 1.7470, 1.4958, 1.2382, 1.0642], + device='cuda:3'), covar=tensor([0.1657, 0.1820, 0.2113, 0.2160, 0.0778, 0.1623, 0.2197, 0.2329], + device='cuda:3'), in_proj_covar=tensor([0.0328, 0.0344, 0.0347, 0.0317, 0.0354, 0.0379, 0.0328, 0.0361], + device='cuda:3'), out_proj_covar=tensor([7.1921e-05, 7.4071e-05, 7.5023e-05, 6.6385e-05, 7.5560e-05, 8.3110e-05, + 7.1353e-05, 7.8106e-05], device='cuda:3') +2023-04-26 11:46:45,448 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4371, 1.6369, 0.7103, 1.1462, 1.6676, 1.3603, 1.2681, 1.3128], + device='cuda:3'), covar=tensor([0.0635, 0.0478, 0.0559, 0.0678, 0.0402, 0.0639, 0.0615, 0.0760], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 11:46:53,614 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6399.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:02,619 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6411.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:07,257 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.650e+02 2.377e+02 2.867e+02 3.488e+02 9.630e+02, threshold=5.734e+02, percent-clipped=2.0 +2023-04-26 11:47:13,366 INFO [finetune.py:976] (3/7) Epoch 2, batch 700, loss[loss=0.3505, simple_loss=0.384, pruned_loss=0.1585, over 4815.00 frames. ], tot_loss[loss=0.2883, simple_loss=0.3308, pruned_loss=0.1229, over 931092.06 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:47:46,401 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6459.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:48,730 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6462.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:50,521 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4407, 0.8988, 0.4968, 1.1269, 1.1713, 1.3522, 1.2170, 1.2426], + device='cuda:3'), covar=tensor([0.0635, 0.0545, 0.0547, 0.0666, 0.0374, 0.0642, 0.0603, 0.0760], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 11:47:50,548 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3820, 1.6053, 1.1204, 0.8975, 1.0993, 1.0924, 1.1183, 1.0191], + device='cuda:3'), covar=tensor([0.2444, 0.2175, 0.3219, 0.3180, 0.4085, 0.3125, 0.2416, 0.3152], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0224, 0.0200, 0.0218, 0.0237, 0.0199, 0.0194, 0.0210], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:47:51,479 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.05 vs. limit=5.0 +2023-04-26 11:47:51,714 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:58,781 INFO [finetune.py:976] (3/7) Epoch 2, batch 750, loss[loss=0.3099, simple_loss=0.3503, pruned_loss=0.1347, over 4920.00 frames. ], tot_loss[loss=0.2902, simple_loss=0.3328, pruned_loss=0.1238, over 937013.73 frames. ], batch size: 42, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:48:18,131 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6491.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:48:50,940 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6514.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:48:53,700 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.458e+02 2.088e+02 2.383e+02 2.881e+02 6.100e+02, threshold=4.765e+02, percent-clipped=1.0 +2023-04-26 11:49:03,675 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6523.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 11:49:03,999 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-26 11:49:05,430 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6526.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:06,607 INFO [finetune.py:976] (3/7) Epoch 2, batch 800, loss[loss=0.3167, simple_loss=0.346, pruned_loss=0.1437, over 4856.00 frames. ], tot_loss[loss=0.2891, simple_loss=0.3318, pruned_loss=0.1232, over 940831.51 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:49:06,748 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6528.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:15,367 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9922, 1.0398, 1.1377, 1.2013, 1.2744, 1.3957, 1.2332, 1.2279], + device='cuda:3'), covar=tensor([3.5333, 7.7600, 5.2512, 4.4047, 4.9407, 8.0934, 6.8217, 5.7816], + device='cuda:3'), in_proj_covar=tensor([0.0269, 0.0371, 0.0298, 0.0299, 0.0327, 0.0351, 0.0359, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 11:49:27,196 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6559.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:29,549 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6562.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:30,223 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7370, 2.2613, 1.8500, 2.2531, 1.7678, 1.7540, 2.1501, 1.4744], + device='cuda:3'), covar=tensor([0.2821, 0.2050, 0.1575, 0.1833, 0.3285, 0.2055, 0.2251, 0.3612], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0335, 0.0245, 0.0310, 0.0313, 0.0284, 0.0280, 0.0301], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:49:33,148 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6567.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:40,229 INFO [finetune.py:976] (3/7) Epoch 2, batch 850, loss[loss=0.2801, simple_loss=0.3247, pruned_loss=0.1178, over 4798.00 frames. ], tot_loss[loss=0.2859, simple_loss=0.3288, pruned_loss=0.1215, over 944208.67 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:50:18,614 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.341e+02 2.044e+02 2.405e+02 3.038e+02 6.612e+02, threshold=4.810e+02, percent-clipped=4.0 +2023-04-26 11:50:21,182 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6622.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:50:25,268 INFO [finetune.py:976] (3/7) Epoch 2, batch 900, loss[loss=0.2596, simple_loss=0.3049, pruned_loss=0.1072, over 4869.00 frames. ], tot_loss[loss=0.281, simple_loss=0.3232, pruned_loss=0.1194, over 944254.51 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:50:25,996 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6629.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:50:27,251 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9444, 0.6064, 0.8417, 1.2576, 1.2172, 0.9164, 0.9018, 0.8393], + device='cuda:3'), covar=tensor([ 6.1875, 9.8376, 10.2461, 10.2861, 7.3857, 10.5180, 10.5455, 6.9879], + device='cuda:3'), in_proj_covar=tensor([0.0439, 0.0504, 0.0586, 0.0568, 0.0469, 0.0526, 0.0533, 0.0542], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:50:32,367 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 11:50:36,414 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6646.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:50:44,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0270, 2.0262, 1.7353, 1.6894, 2.0323, 1.7401, 2.5101, 1.4260], + device='cuda:3'), covar=tensor([0.4248, 0.1521, 0.4538, 0.2972, 0.2011, 0.2330, 0.1304, 0.4245], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0347, 0.0430, 0.0363, 0.0400, 0.0370, 0.0395, 0.0404], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:50:55,493 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7193, 2.4019, 1.5186, 1.5888, 1.2609, 1.2563, 1.5596, 1.1419], + device='cuda:3'), covar=tensor([0.2504, 0.2220, 0.2960, 0.3397, 0.3819, 0.3109, 0.2409, 0.3260], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0223, 0.0199, 0.0218, 0.0236, 0.0198, 0.0193, 0.0210], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:50:58,409 INFO [finetune.py:976] (3/7) Epoch 2, batch 950, loss[loss=0.3034, simple_loss=0.3314, pruned_loss=0.1377, over 4805.00 frames. ], tot_loss[loss=0.28, simple_loss=0.3219, pruned_loss=0.1191, over 947508.71 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:51:06,458 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6690.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:51:08,886 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6694.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:51:11,951 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6699.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:51:41,856 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.328e+02 2.213e+02 2.518e+02 3.022e+02 8.936e+02, threshold=5.037e+02, percent-clipped=4.0 +2023-04-26 11:51:43,849 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0070, 2.7996, 2.0515, 2.4100, 1.9931, 2.1085, 2.5219, 1.7522], + device='cuda:3'), covar=tensor([0.2805, 0.1801, 0.1634, 0.2123, 0.3163, 0.1895, 0.2600, 0.3501], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0334, 0.0245, 0.0310, 0.0312, 0.0283, 0.0280, 0.0301], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:51:48,517 INFO [finetune.py:976] (3/7) Epoch 2, batch 1000, loss[loss=0.2958, simple_loss=0.3423, pruned_loss=0.1247, over 4819.00 frames. ], tot_loss[loss=0.2844, simple_loss=0.3259, pruned_loss=0.1215, over 950295.52 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:52:11,814 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6747.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:52:55,094 INFO [finetune.py:976] (3/7) Epoch 2, batch 1050, loss[loss=0.2776, simple_loss=0.3314, pruned_loss=0.1119, over 4798.00 frames. ], tot_loss[loss=0.2888, simple_loss=0.3311, pruned_loss=0.1233, over 952543.23 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:53:03,744 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6791.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:53:33,223 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.642e+02 2.078e+02 2.516e+02 2.901e+02 4.864e+02, threshold=5.033e+02, percent-clipped=0.0 +2023-04-26 11:53:33,309 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6818.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 11:53:42,764 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6823.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:53:44,630 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6826.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:53:51,484 INFO [finetune.py:976] (3/7) Epoch 2, batch 1100, loss[loss=0.2945, simple_loss=0.342, pruned_loss=0.1235, over 4823.00 frames. ], tot_loss[loss=0.2886, simple_loss=0.3317, pruned_loss=0.1227, over 954974.34 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:54:04,442 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6839.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:05,782 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7453, 2.3137, 1.5019, 1.4562, 1.2453, 1.2726, 1.5693, 1.1649], + device='cuda:3'), covar=tensor([0.2430, 0.2461, 0.3035, 0.3481, 0.4092, 0.3105, 0.2341, 0.3374], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0224, 0.0199, 0.0220, 0.0237, 0.0199, 0.0194, 0.0211], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:54:09,761 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-04-26 11:54:16,610 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:17,903 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9463, 0.8411, 1.0960, 1.2413, 1.1747, 1.0048, 1.0736, 1.0699], + device='cuda:3'), covar=tensor([ 5.9473, 8.8426, 9.6981, 9.8748, 6.6010, 10.5929, 10.3081, 7.2953], + device='cuda:3'), in_proj_covar=tensor([0.0439, 0.0504, 0.0586, 0.0570, 0.0469, 0.0525, 0.0532, 0.0541], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 11:54:22,454 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6867.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:26,488 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4739, 2.0640, 1.3341, 1.1852, 1.1529, 1.1243, 1.3919, 1.0758], + device='cuda:3'), covar=tensor([0.2661, 0.2368, 0.3108, 0.3760, 0.4340, 0.3107, 0.2405, 0.3437], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0224, 0.0199, 0.0219, 0.0237, 0.0198, 0.0194, 0.0211], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 11:54:27,636 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6874.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:30,495 INFO [finetune.py:976] (3/7) Epoch 2, batch 1150, loss[loss=0.314, simple_loss=0.3557, pruned_loss=0.1361, over 4884.00 frames. ], tot_loss[loss=0.2898, simple_loss=0.3328, pruned_loss=0.1234, over 955340.76 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:54:30,597 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6878.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:54:49,252 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6907.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:50,557 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6909.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:54,162 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6915.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:56,911 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.447e+02 2.231e+02 2.530e+02 3.126e+02 7.008e+02, threshold=5.060e+02, percent-clipped=3.0 +2023-04-26 11:54:59,928 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6922.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:55:04,519 INFO [finetune.py:976] (3/7) Epoch 2, batch 1200, loss[loss=0.2956, simple_loss=0.3282, pruned_loss=0.1314, over 4878.00 frames. ], tot_loss[loss=0.2863, simple_loss=0.3295, pruned_loss=0.1215, over 955766.22 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:55:11,766 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6939.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:55:29,368 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6515, 1.2631, 4.3946, 4.0731, 3.8687, 4.1214, 4.0838, 3.9111], + device='cuda:3'), covar=tensor([0.6598, 0.5960, 0.0988, 0.1751, 0.1042, 0.1329, 0.1517, 0.1449], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0316, 0.0453, 0.0458, 0.0379, 0.0438, 0.0346, 0.0401], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 11:55:32,156 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:55:32,226 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:55:38,018 INFO [finetune.py:976] (3/7) Epoch 2, batch 1250, loss[loss=0.2559, simple_loss=0.3012, pruned_loss=0.1053, over 4932.00 frames. ], tot_loss[loss=0.2804, simple_loss=0.324, pruned_loss=0.1184, over 956275.13 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:55:43,403 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6985.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:56:04,201 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.519e+02 2.106e+02 2.592e+02 3.021e+02 5.805e+02, threshold=5.184e+02, percent-clipped=1.0 +2023-04-26 11:56:11,779 INFO [finetune.py:976] (3/7) Epoch 2, batch 1300, loss[loss=0.2609, simple_loss=0.2784, pruned_loss=0.1217, over 4185.00 frames. ], tot_loss[loss=0.2751, simple_loss=0.3187, pruned_loss=0.1157, over 953845.90 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:56:25,550 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7048.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:56:55,870 INFO [finetune.py:976] (3/7) Epoch 2, batch 1350, loss[loss=0.2722, simple_loss=0.3128, pruned_loss=0.1158, over 4755.00 frames. ], tot_loss[loss=0.2743, simple_loss=0.3178, pruned_loss=0.1154, over 953623.63 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:57:34,477 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7109.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:57:39,841 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.388e+02 2.079e+02 2.494e+02 3.022e+02 7.754e+02, threshold=4.988e+02, percent-clipped=2.0 +2023-04-26 11:57:39,963 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7118.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:57:48,867 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7123.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:57:56,137 INFO [finetune.py:976] (3/7) Epoch 2, batch 1400, loss[loss=0.2321, simple_loss=0.2996, pruned_loss=0.08228, over 4905.00 frames. ], tot_loss[loss=0.2768, simple_loss=0.3213, pruned_loss=0.1161, over 955133.47 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:58:36,313 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7163.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:58:36,338 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6631, 1.3691, 1.9707, 1.8495, 1.4952, 1.2124, 1.5283, 0.9917], + device='cuda:3'), covar=tensor([0.0800, 0.1161, 0.0571, 0.0852, 0.1055, 0.1490, 0.0892, 0.1342], + device='cuda:3'), in_proj_covar=tensor([0.0072, 0.0080, 0.0076, 0.0075, 0.0088, 0.0096, 0.0090, 0.0081], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 11:58:38,105 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7166.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:58:47,201 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7171.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:58:51,415 INFO [finetune.py:976] (3/7) Epoch 2, batch 1450, loss[loss=0.2935, simple_loss=0.3404, pruned_loss=0.1232, over 4850.00 frames. ], tot_loss[loss=0.2774, simple_loss=0.3223, pruned_loss=0.1162, over 954591.41 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:58:56,892 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-26 11:59:19,363 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.461e+02 2.205e+02 2.539e+02 2.992e+02 8.732e+02, threshold=5.079e+02, percent-clipped=2.0 +2023-04-26 11:59:22,653 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-26 11:59:23,178 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7224.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 11:59:24,477 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.63 vs. limit=2.0 +2023-04-26 11:59:25,498 INFO [finetune.py:976] (3/7) Epoch 2, batch 1500, loss[loss=0.3103, simple_loss=0.3499, pruned_loss=0.1353, over 4812.00 frames. ], tot_loss[loss=0.2804, simple_loss=0.3252, pruned_loss=0.1178, over 956149.67 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:59:30,170 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7234.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:59:51,523 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:59:57,731 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7275.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:59:59,532 INFO [finetune.py:976] (3/7) Epoch 2, batch 1550, loss[loss=0.3203, simple_loss=0.3463, pruned_loss=0.1471, over 4248.00 frames. ], tot_loss[loss=0.2797, simple_loss=0.3252, pruned_loss=0.1171, over 956137.81 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:00:04,485 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7285.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:00:18,871 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6937, 2.2819, 1.6779, 1.4123, 1.2626, 1.3080, 1.7129, 1.2860], + device='cuda:3'), covar=tensor([0.2588, 0.2449, 0.2819, 0.3577, 0.4054, 0.2973, 0.2263, 0.3101], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0224, 0.0199, 0.0220, 0.0237, 0.0198, 0.0194, 0.0211], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 12:00:27,655 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.588e+02 2.228e+02 2.487e+02 3.016e+02 5.047e+02, threshold=4.974e+02, percent-clipped=0.0 +2023-04-26 12:00:33,829 INFO [finetune.py:976] (3/7) Epoch 2, batch 1600, loss[loss=0.2395, simple_loss=0.2933, pruned_loss=0.09288, over 4753.00 frames. ], tot_loss[loss=0.2763, simple_loss=0.3215, pruned_loss=0.1156, over 957477.93 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:00:36,971 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7333.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:00:38,898 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7336.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:00:53,668 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 12:01:07,695 INFO [finetune.py:976] (3/7) Epoch 2, batch 1650, loss[loss=0.198, simple_loss=0.2584, pruned_loss=0.06877, over 4815.00 frames. ], tot_loss[loss=0.272, simple_loss=0.3171, pruned_loss=0.1134, over 957511.19 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:01:25,936 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7404.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:01:28,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0998, 1.2914, 1.1966, 1.5816, 1.4190, 1.6282, 1.2906, 2.5219], + device='cuda:3'), covar=tensor([0.0676, 0.0864, 0.0850, 0.1298, 0.0722, 0.0563, 0.0816, 0.0264], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 12:01:28,264 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7407.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:01:34,836 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.570e+02 2.142e+02 2.529e+02 2.973e+02 5.158e+02, threshold=5.058e+02, percent-clipped=1.0 +2023-04-26 12:01:41,461 INFO [finetune.py:976] (3/7) Epoch 2, batch 1700, loss[loss=0.2383, simple_loss=0.2952, pruned_loss=0.09072, over 4809.00 frames. ], tot_loss[loss=0.2697, simple_loss=0.3148, pruned_loss=0.1123, over 958250.82 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:02:09,801 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7468.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:02:21,931 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 12:02:22,154 INFO [finetune.py:976] (3/7) Epoch 2, batch 1750, loss[loss=0.3254, simple_loss=0.3709, pruned_loss=0.1399, over 4835.00 frames. ], tot_loss[loss=0.274, simple_loss=0.3189, pruned_loss=0.1145, over 958489.54 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:02:45,817 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4245, 3.7630, 0.7922, 1.9027, 1.9087, 2.4371, 2.3013, 0.8868], + device='cuda:3'), covar=tensor([0.1464, 0.0908, 0.2209, 0.1443, 0.1210, 0.1251, 0.1477, 0.2571], + device='cuda:3'), in_proj_covar=tensor([0.0125, 0.0275, 0.0153, 0.0134, 0.0145, 0.0167, 0.0131, 0.0134], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 12:03:11,389 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.401e+02 2.126e+02 2.582e+02 2.994e+02 4.895e+02, threshold=5.165e+02, percent-clipped=0.0 +2023-04-26 12:03:12,089 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7519.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:03:12,196 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 12:03:12,745 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0007, 2.4214, 1.9803, 2.3257, 1.6915, 2.0063, 2.0957, 1.6583], + device='cuda:3'), covar=tensor([0.2346, 0.1328, 0.1279, 0.1602, 0.3176, 0.1480, 0.2259, 0.3218], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0333, 0.0244, 0.0309, 0.0312, 0.0284, 0.0278, 0.0301], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:03:23,649 INFO [finetune.py:976] (3/7) Epoch 2, batch 1800, loss[loss=0.2094, simple_loss=0.2553, pruned_loss=0.08174, over 3578.00 frames. ], tot_loss[loss=0.2763, simple_loss=0.3219, pruned_loss=0.1153, over 957018.54 frames. ], batch size: 15, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:03:33,051 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7534.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:04:17,654 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:04:31,080 INFO [finetune.py:976] (3/7) Epoch 2, batch 1850, loss[loss=0.3278, simple_loss=0.3517, pruned_loss=0.152, over 4824.00 frames. ], tot_loss[loss=0.2784, simple_loss=0.3238, pruned_loss=0.1165, over 956785.07 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:04:34,062 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7582.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:05:07,098 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7613.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:05:10,536 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.515e+02 2.096e+02 2.728e+02 3.385e+02 6.441e+02, threshold=5.455e+02, percent-clipped=3.0 +2023-04-26 12:05:14,350 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7624.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:05:16,721 INFO [finetune.py:976] (3/7) Epoch 2, batch 1900, loss[loss=0.2556, simple_loss=0.3174, pruned_loss=0.09689, over 4918.00 frames. ], tot_loss[loss=0.2791, simple_loss=0.325, pruned_loss=0.1166, over 955757.39 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:05:18,632 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7631.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:05:47,120 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0463, 1.2592, 4.8949, 4.4916, 4.2520, 4.4808, 4.4098, 4.3544], + device='cuda:3'), covar=tensor([0.6668, 0.6307, 0.1083, 0.2180, 0.1152, 0.1374, 0.1509, 0.1872], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0316, 0.0451, 0.0459, 0.0380, 0.0435, 0.0345, 0.0401], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:05:49,631 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-04-26 12:05:50,040 INFO [finetune.py:976] (3/7) Epoch 2, batch 1950, loss[loss=0.2622, simple_loss=0.3092, pruned_loss=0.1076, over 4834.00 frames. ], tot_loss[loss=0.2752, simple_loss=0.3217, pruned_loss=0.1144, over 954815.78 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:05:54,910 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7685.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:06:06,996 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7704.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:06:08,158 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3407, 1.2813, 1.5951, 1.5350, 1.4565, 1.2765, 1.3918, 1.3799], + device='cuda:3'), covar=tensor([4.4395, 6.3222, 7.7343, 7.5738, 4.8017, 8.5893, 7.9630, 5.7942], + device='cuda:3'), in_proj_covar=tensor([0.0445, 0.0509, 0.0595, 0.0580, 0.0477, 0.0530, 0.0537, 0.0548], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:06:14,090 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.43 vs. limit=5.0 +2023-04-26 12:06:17,408 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.610e+02 2.113e+02 2.438e+02 2.790e+02 7.164e+02, threshold=4.875e+02, percent-clipped=1.0 +2023-04-26 12:06:23,986 INFO [finetune.py:976] (3/7) Epoch 2, batch 2000, loss[loss=0.2894, simple_loss=0.3047, pruned_loss=0.137, over 4152.00 frames. ], tot_loss[loss=0.2723, simple_loss=0.3184, pruned_loss=0.1131, over 951993.05 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:06:33,718 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7743.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:06:39,111 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7752.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:06:40,953 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5134, 1.7079, 1.6693, 1.7158, 1.6804, 1.7875, 1.7913, 1.7143], + device='cuda:3'), covar=tensor([1.1504, 2.8499, 2.4319, 1.9176, 2.2782, 3.5502, 2.7577, 2.5984], + device='cuda:3'), in_proj_covar=tensor([0.0275, 0.0377, 0.0302, 0.0302, 0.0331, 0.0362, 0.0363, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:06:47,288 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:06:49,688 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.55 vs. limit=2.0 +2023-04-26 12:06:50,449 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 12:06:57,782 INFO [finetune.py:976] (3/7) Epoch 2, batch 2050, loss[loss=0.2024, simple_loss=0.2521, pruned_loss=0.07639, over 4807.00 frames. ], tot_loss[loss=0.2693, simple_loss=0.3148, pruned_loss=0.1119, over 952348.38 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:07:13,098 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7802.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:07:14,334 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7804.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:07:24,632 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.397e+02 2.038e+02 2.548e+02 3.001e+02 7.131e+02, threshold=5.096e+02, percent-clipped=2.0 +2023-04-26 12:07:25,341 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7819.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:07:29,683 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.10 vs. limit=5.0 +2023-04-26 12:07:31,154 INFO [finetune.py:976] (3/7) Epoch 2, batch 2100, loss[loss=0.312, simple_loss=0.3388, pruned_loss=0.1426, over 4785.00 frames. ], tot_loss[loss=0.2685, simple_loss=0.3134, pruned_loss=0.1118, over 949777.71 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:07:40,890 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6095, 1.3325, 1.1970, 1.3997, 1.8276, 1.5661, 1.3093, 1.1463], + device='cuda:3'), covar=tensor([0.1955, 0.2143, 0.2358, 0.2172, 0.1073, 0.2223, 0.2628, 0.2238], + device='cuda:3'), in_proj_covar=tensor([0.0327, 0.0342, 0.0347, 0.0317, 0.0354, 0.0374, 0.0326, 0.0358], + device='cuda:3'), out_proj_covar=tensor([7.1592e-05, 7.3557e-05, 7.4867e-05, 6.6526e-05, 7.5566e-05, 8.1904e-05, + 7.1018e-05, 7.7268e-05], device='cuda:3') +2023-04-26 12:07:54,038 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7863.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:07:56,931 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7867.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:08:15,064 INFO [finetune.py:976] (3/7) Epoch 2, batch 2150, loss[loss=0.2456, simple_loss=0.3163, pruned_loss=0.08748, over 4755.00 frames. ], tot_loss[loss=0.2728, simple_loss=0.318, pruned_loss=0.1138, over 949822.81 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:08:52,497 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2229, 1.4175, 1.8521, 2.5728, 1.7822, 1.3649, 1.2032, 1.7064], + device='cuda:3'), covar=tensor([0.6754, 0.7954, 0.3619, 0.7413, 0.8763, 0.6063, 1.1274, 0.8752], + device='cuda:3'), in_proj_covar=tensor([0.0270, 0.0283, 0.0226, 0.0351, 0.0241, 0.0238, 0.0277, 0.0224], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:09:02,673 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 2.251e+02 2.703e+02 3.381e+02 5.777e+02, threshold=5.406e+02, percent-clipped=2.0 +2023-04-26 12:09:20,981 INFO [finetune.py:976] (3/7) Epoch 2, batch 2200, loss[loss=0.2619, simple_loss=0.3228, pruned_loss=0.1005, over 4891.00 frames. ], tot_loss[loss=0.2738, simple_loss=0.3196, pruned_loss=0.114, over 952058.98 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:09:22,952 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7931.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:10:21,348 INFO [finetune.py:976] (3/7) Epoch 2, batch 2250, loss[loss=0.2471, simple_loss=0.3117, pruned_loss=0.09128, over 4923.00 frames. ], tot_loss[loss=0.2755, simple_loss=0.3213, pruned_loss=0.1148, over 950756.79 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:10:22,018 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7979.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:10:23,126 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7980.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:10:30,865 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-04-26 12:10:47,702 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.449e+02 2.028e+02 2.420e+02 2.875e+02 5.113e+02, threshold=4.840e+02, percent-clipped=0.0 +2023-04-26 12:10:55,299 INFO [finetune.py:976] (3/7) Epoch 2, batch 2300, loss[loss=0.306, simple_loss=0.3446, pruned_loss=0.1337, over 4716.00 frames. ], tot_loss[loss=0.2748, simple_loss=0.3215, pruned_loss=0.114, over 952537.91 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:11:18,750 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8063.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:11:28,366 INFO [finetune.py:976] (3/7) Epoch 2, batch 2350, loss[loss=0.2598, simple_loss=0.3136, pruned_loss=0.103, over 4893.00 frames. ], tot_loss[loss=0.2723, simple_loss=0.319, pruned_loss=0.1128, over 953633.60 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:11:42,661 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8099.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:11:48,699 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4303, 1.1989, 1.6256, 1.5264, 1.2977, 1.0487, 1.2209, 0.8381], + device='cuda:3'), covar=tensor([0.0772, 0.0991, 0.0540, 0.0990, 0.1013, 0.1300, 0.0743, 0.1054], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0079, 0.0075, 0.0074, 0.0087, 0.0095, 0.0089, 0.0080], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 12:11:50,390 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8111.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:11:54,656 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 2.063e+02 2.510e+02 2.746e+02 4.758e+02, threshold=5.020e+02, percent-clipped=0.0 +2023-04-26 12:12:00,036 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-04-26 12:12:01,655 INFO [finetune.py:976] (3/7) Epoch 2, batch 2400, loss[loss=0.2655, simple_loss=0.313, pruned_loss=0.109, over 4750.00 frames. ], tot_loss[loss=0.2693, simple_loss=0.3154, pruned_loss=0.1116, over 955190.22 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:12:21,843 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8158.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:12:34,354 INFO [finetune.py:976] (3/7) Epoch 2, batch 2450, loss[loss=0.2615, simple_loss=0.3163, pruned_loss=0.1034, over 4858.00 frames. ], tot_loss[loss=0.2646, simple_loss=0.3107, pruned_loss=0.1093, over 954840.27 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:12:49,769 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9454, 2.8129, 2.6279, 2.7765, 3.0280, 2.5549, 3.6545, 2.3260], + device='cuda:3'), covar=tensor([0.4072, 0.1626, 0.3282, 0.2796, 0.1965, 0.2443, 0.1406, 0.3463], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0351, 0.0434, 0.0367, 0.0402, 0.0373, 0.0398, 0.0408], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:13:01,849 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.405e+02 2.093e+02 2.421e+02 2.814e+02 5.190e+02, threshold=4.843e+02, percent-clipped=1.0 +2023-04-26 12:13:05,596 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2986, 3.2805, 2.5746, 3.8119, 3.3569, 3.2947, 1.3884, 3.3093], + device='cuda:3'), covar=tensor([0.1918, 0.1275, 0.3557, 0.2322, 0.3236, 0.2144, 0.5817, 0.2352], + device='cuda:3'), in_proj_covar=tensor([0.0260, 0.0233, 0.0278, 0.0327, 0.0322, 0.0273, 0.0287, 0.0287], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:13:07,998 INFO [finetune.py:976] (3/7) Epoch 2, batch 2500, loss[loss=0.3051, simple_loss=0.3363, pruned_loss=0.137, over 4868.00 frames. ], tot_loss[loss=0.2664, simple_loss=0.3119, pruned_loss=0.1104, over 954894.28 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:13:27,566 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.91 vs. limit=5.0 +2023-04-26 12:13:49,845 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:13:59,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2365, 2.9194, 0.8641, 1.5336, 1.5905, 2.0488, 1.8452, 0.8828], + device='cuda:3'), covar=tensor([0.1622, 0.1607, 0.2304, 0.1637, 0.1313, 0.1340, 0.1787, 0.2229], + device='cuda:3'), in_proj_covar=tensor([0.0124, 0.0274, 0.0152, 0.0133, 0.0145, 0.0167, 0.0130, 0.0135], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 12:14:03,521 INFO [finetune.py:976] (3/7) Epoch 2, batch 2550, loss[loss=0.2769, simple_loss=0.3309, pruned_loss=0.1114, over 4799.00 frames. ], tot_loss[loss=0.2694, simple_loss=0.3157, pruned_loss=0.1116, over 954344.07 frames. ], batch size: 41, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:14:04,877 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8280.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:14:37,932 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9009, 1.1473, 4.9643, 4.5805, 4.3845, 4.6648, 4.4618, 4.3989], + device='cuda:3'), covar=tensor([0.6715, 0.6455, 0.1049, 0.1791, 0.0981, 0.1278, 0.1515, 0.1678], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0318, 0.0453, 0.0458, 0.0381, 0.0436, 0.0345, 0.0402], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:15:02,134 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.269e+02 2.032e+02 2.420e+02 2.910e+02 5.283e+02, threshold=4.841e+02, percent-clipped=1.0 +2023-04-26 12:15:13,556 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8326.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:15:20,190 INFO [finetune.py:976] (3/7) Epoch 2, batch 2600, loss[loss=0.2988, simple_loss=0.3379, pruned_loss=0.1299, over 4799.00 frames. ], tot_loss[loss=0.2704, simple_loss=0.3176, pruned_loss=0.1116, over 955457.64 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:15:20,256 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8328.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:16:05,389 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7428, 1.3261, 1.2196, 1.4464, 1.9272, 1.6317, 1.3055, 1.2589], + device='cuda:3'), covar=tensor([0.1543, 0.1791, 0.2221, 0.1602, 0.0928, 0.1863, 0.2439, 0.2054], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0338, 0.0345, 0.0314, 0.0351, 0.0369, 0.0322, 0.0355], + device='cuda:3'), out_proj_covar=tensor([7.0608e-05, 7.2717e-05, 7.4500e-05, 6.5795e-05, 7.4792e-05, 8.0897e-05, + 7.0132e-05, 7.6589e-05], device='cuda:3') +2023-04-26 12:16:23,392 INFO [finetune.py:976] (3/7) Epoch 2, batch 2650, loss[loss=0.3156, simple_loss=0.3505, pruned_loss=0.1403, over 4815.00 frames. ], tot_loss[loss=0.2719, simple_loss=0.3192, pruned_loss=0.1123, over 954771.02 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:16:49,117 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8399.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:17:13,452 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.592e+02 2.071e+02 2.411e+02 3.038e+02 9.260e+02, threshold=4.823e+02, percent-clipped=4.0 +2023-04-26 12:17:15,386 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5459, 1.3919, 0.5854, 1.2140, 1.4076, 1.4289, 1.3305, 1.3799], + device='cuda:3'), covar=tensor([0.0589, 0.0488, 0.0525, 0.0654, 0.0355, 0.0601, 0.0575, 0.0724], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0048, 0.0044, 0.0039, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 12:17:19,412 INFO [finetune.py:976] (3/7) Epoch 2, batch 2700, loss[loss=0.3244, simple_loss=0.3527, pruned_loss=0.1481, over 4858.00 frames. ], tot_loss[loss=0.2713, simple_loss=0.3189, pruned_loss=0.1118, over 955284.88 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:17:32,537 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8447.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:17:40,762 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:17:41,979 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8933, 1.6391, 2.3384, 2.4003, 1.6911, 1.3552, 1.9561, 1.0268], + device='cuda:3'), covar=tensor([0.1397, 0.1270, 0.0688, 0.1080, 0.1626, 0.1735, 0.1042, 0.1760], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0079, 0.0075, 0.0074, 0.0087, 0.0095, 0.0089, 0.0080], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 12:17:53,364 INFO [finetune.py:976] (3/7) Epoch 2, batch 2750, loss[loss=0.2783, simple_loss=0.3147, pruned_loss=0.1209, over 4781.00 frames. ], tot_loss[loss=0.2674, simple_loss=0.3146, pruned_loss=0.1101, over 954561.39 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:18:12,773 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:18:21,018 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.402e+02 1.950e+02 2.341e+02 2.831e+02 5.180e+02, threshold=4.681e+02, percent-clipped=1.0 +2023-04-26 12:18:26,545 INFO [finetune.py:976] (3/7) Epoch 2, batch 2800, loss[loss=0.2276, simple_loss=0.279, pruned_loss=0.08806, over 4838.00 frames. ], tot_loss[loss=0.264, simple_loss=0.3108, pruned_loss=0.1085, over 956685.80 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:18:48,000 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2945, 1.0082, 1.3593, 1.6136, 1.4394, 1.2648, 1.3348, 1.3132], + device='cuda:3'), covar=tensor([4.7921, 6.9727, 7.7965, 7.8554, 5.2415, 8.2391, 7.9309, 5.7888], + device='cuda:3'), in_proj_covar=tensor([0.0446, 0.0508, 0.0596, 0.0586, 0.0478, 0.0527, 0.0536, 0.0548], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:19:00,280 INFO [finetune.py:976] (3/7) Epoch 2, batch 2850, loss[loss=0.2317, simple_loss=0.28, pruned_loss=0.0917, over 4849.00 frames. ], tot_loss[loss=0.2601, simple_loss=0.3074, pruned_loss=0.1064, over 957348.36 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:19:45,301 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.328e+02 2.166e+02 2.478e+02 2.838e+02 4.479e+02, threshold=4.956e+02, percent-clipped=0.0 +2023-04-26 12:19:52,097 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8621.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:19:56,364 INFO [finetune.py:976] (3/7) Epoch 2, batch 2900, loss[loss=0.259, simple_loss=0.3078, pruned_loss=0.1051, over 4753.00 frames. ], tot_loss[loss=0.2627, simple_loss=0.3105, pruned_loss=0.1074, over 956333.79 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:20:17,636 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5620, 0.7912, 1.0043, 1.1264, 1.1838, 1.3265, 1.0883, 1.0835], + device='cuda:3'), covar=tensor([2.0287, 3.8386, 3.0016, 2.7596, 2.8891, 4.8044, 3.7445, 3.3766], + device='cuda:3'), in_proj_covar=tensor([0.0278, 0.0379, 0.0303, 0.0306, 0.0333, 0.0367, 0.0366, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:20:26,910 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6474, 1.8543, 1.0157, 1.3732, 2.1276, 1.5507, 1.4461, 1.6060], + device='cuda:3'), covar=tensor([0.0597, 0.0468, 0.0439, 0.0651, 0.0291, 0.0638, 0.0588, 0.0716], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:3'), out_proj_covar=tensor([0.0048, 0.0044, 0.0039, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 12:20:55,916 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-26 12:21:07,306 INFO [finetune.py:976] (3/7) Epoch 2, batch 2950, loss[loss=0.2424, simple_loss=0.3018, pruned_loss=0.09146, over 4868.00 frames. ], tot_loss[loss=0.267, simple_loss=0.3147, pruned_loss=0.1096, over 955934.35 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:22:03,513 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.340e+02 2.015e+02 2.424e+02 2.946e+02 5.788e+02, threshold=4.848e+02, percent-clipped=1.0 +2023-04-26 12:22:15,447 INFO [finetune.py:976] (3/7) Epoch 2, batch 3000, loss[loss=0.274, simple_loss=0.3239, pruned_loss=0.112, over 4806.00 frames. ], tot_loss[loss=0.2675, simple_loss=0.3153, pruned_loss=0.1098, over 955785.74 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:22:15,447 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 12:22:18,469 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4468, 1.5140, 3.8715, 3.5484, 3.4654, 3.6609, 3.8095, 3.4419], + device='cuda:3'), covar=tensor([0.7273, 0.5388, 0.1341, 0.2431, 0.1429, 0.1713, 0.0873, 0.1766], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0317, 0.0451, 0.0459, 0.0382, 0.0436, 0.0345, 0.0403], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:22:19,212 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3773, 3.0133, 1.0275, 1.7250, 1.9020, 2.2855, 2.0822, 0.9889], + device='cuda:3'), covar=tensor([0.1316, 0.1181, 0.1805, 0.1423, 0.0988, 0.0994, 0.1352, 0.1747], + device='cuda:3'), in_proj_covar=tensor([0.0125, 0.0275, 0.0153, 0.0134, 0.0145, 0.0168, 0.0131, 0.0136], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 12:22:32,547 INFO [finetune.py:1010] (3/7) Epoch 2, validation: loss=0.1863, simple_loss=0.2571, pruned_loss=0.0578, over 2265189.00 frames. +2023-04-26 12:22:32,547 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6048MB +2023-04-26 12:22:38,468 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9753, 2.4825, 2.0697, 2.5176, 2.0065, 2.1182, 2.1724, 1.6781], + device='cuda:3'), covar=tensor([0.1905, 0.1241, 0.1139, 0.1275, 0.2676, 0.1328, 0.1739, 0.2482], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0338, 0.0247, 0.0313, 0.0318, 0.0288, 0.0281, 0.0304], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:23:09,088 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8760.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:23:10,868 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4297, 3.3720, 2.6344, 3.9338, 3.3723, 3.4178, 1.4630, 3.2881], + device='cuda:3'), covar=tensor([0.1869, 0.1233, 0.2931, 0.1972, 0.2675, 0.2024, 0.5760, 0.2305], + device='cuda:3'), in_proj_covar=tensor([0.0259, 0.0232, 0.0275, 0.0325, 0.0321, 0.0270, 0.0284, 0.0284], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:23:21,916 INFO [finetune.py:976] (3/7) Epoch 2, batch 3050, loss[loss=0.2833, simple_loss=0.3302, pruned_loss=0.1182, over 4787.00 frames. ], tot_loss[loss=0.2671, simple_loss=0.3158, pruned_loss=0.1092, over 953376.74 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:23:49,151 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.298e+02 2.073e+02 2.370e+02 3.091e+02 5.908e+02, threshold=4.740e+02, percent-clipped=3.0 +2023-04-26 12:23:51,001 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8821.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:23:56,177 INFO [finetune.py:976] (3/7) Epoch 2, batch 3100, loss[loss=0.2414, simple_loss=0.2971, pruned_loss=0.09282, over 4850.00 frames. ], tot_loss[loss=0.2637, simple_loss=0.3124, pruned_loss=0.1075, over 953165.09 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:23:56,961 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 12:24:10,032 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7623, 2.2957, 1.8054, 2.2423, 1.6633, 1.8200, 2.0629, 1.4602], + device='cuda:3'), covar=tensor([0.2322, 0.1618, 0.1226, 0.1526, 0.3251, 0.1688, 0.1846, 0.2846], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0336, 0.0247, 0.0312, 0.0316, 0.0288, 0.0279, 0.0303], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:24:29,315 INFO [finetune.py:976] (3/7) Epoch 2, batch 3150, loss[loss=0.2107, simple_loss=0.2661, pruned_loss=0.07762, over 4827.00 frames. ], tot_loss[loss=0.2611, simple_loss=0.3087, pruned_loss=0.1068, over 954761.52 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:24:47,629 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 12:24:56,258 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.408e+02 2.116e+02 2.518e+02 3.010e+02 6.088e+02, threshold=5.037e+02, percent-clipped=1.0 +2023-04-26 12:24:57,598 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8921.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:01,753 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8927.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:02,226 INFO [finetune.py:976] (3/7) Epoch 2, batch 3200, loss[loss=0.2492, simple_loss=0.2904, pruned_loss=0.104, over 4904.00 frames. ], tot_loss[loss=0.2573, simple_loss=0.3049, pruned_loss=0.1049, over 956089.59 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:25:30,260 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8969.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:41,813 INFO [finetune.py:976] (3/7) Epoch 2, batch 3250, loss[loss=0.3419, simple_loss=0.3797, pruned_loss=0.1521, over 4906.00 frames. ], tot_loss[loss=0.2596, simple_loss=0.3074, pruned_loss=0.1059, over 954520.53 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:25:53,829 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8986.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:55,583 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8988.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:26:16,799 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.04 vs. limit=5.0 +2023-04-26 12:26:27,410 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9013.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:26:37,526 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.210e+02 2.159e+02 2.478e+02 3.003e+02 4.851e+02, threshold=4.955e+02, percent-clipped=0.0 +2023-04-26 12:26:48,551 INFO [finetune.py:976] (3/7) Epoch 2, batch 3300, loss[loss=0.292, simple_loss=0.3385, pruned_loss=0.1228, over 4829.00 frames. ], tot_loss[loss=0.2661, simple_loss=0.3135, pruned_loss=0.1094, over 953796.30 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:27:08,796 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9047.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:27:26,250 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9074.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:27:28,528 INFO [finetune.py:976] (3/7) Epoch 2, batch 3350, loss[loss=0.3214, simple_loss=0.3522, pruned_loss=0.1453, over 4832.00 frames. ], tot_loss[loss=0.2686, simple_loss=0.3165, pruned_loss=0.1103, over 954310.49 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:28:00,088 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9116.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:28:07,186 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.418e+02 2.168e+02 2.587e+02 3.094e+02 5.996e+02, threshold=5.175e+02, percent-clipped=1.0 +2023-04-26 12:28:17,531 INFO [finetune.py:976] (3/7) Epoch 2, batch 3400, loss[loss=0.2613, simple_loss=0.3204, pruned_loss=0.1011, over 4877.00 frames. ], tot_loss[loss=0.2695, simple_loss=0.3177, pruned_loss=0.1106, over 954431.28 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:28:54,819 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4329, 1.1434, 0.3679, 1.1654, 0.9729, 1.3244, 1.2252, 1.2215], + device='cuda:3'), covar=tensor([0.0667, 0.0491, 0.0587, 0.0663, 0.0416, 0.0631, 0.0616, 0.0765], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 12:28:59,899 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.21 vs. limit=5.0 +2023-04-26 12:29:01,798 INFO [finetune.py:976] (3/7) Epoch 2, batch 3450, loss[loss=0.2837, simple_loss=0.3164, pruned_loss=0.1255, over 4765.00 frames. ], tot_loss[loss=0.2673, simple_loss=0.3159, pruned_loss=0.1094, over 954662.22 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:29:29,511 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.432e+02 2.133e+02 2.511e+02 2.934e+02 6.196e+02, threshold=5.021e+02, percent-clipped=2.0 +2023-04-26 12:29:35,078 INFO [finetune.py:976] (3/7) Epoch 2, batch 3500, loss[loss=0.2511, simple_loss=0.3021, pruned_loss=0.1, over 4802.00 frames. ], tot_loss[loss=0.2632, simple_loss=0.3117, pruned_loss=0.1074, over 954371.97 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:29:47,519 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7480, 2.4771, 1.6067, 1.6043, 1.3028, 1.3100, 1.6230, 1.2462], + device='cuda:3'), covar=tensor([0.2617, 0.2524, 0.2974, 0.3529, 0.4071, 0.3360, 0.2325, 0.3491], + device='cuda:3'), in_proj_covar=tensor([0.0200, 0.0227, 0.0197, 0.0219, 0.0236, 0.0198, 0.0192, 0.0211], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 12:29:51,065 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8983, 3.8282, 2.8621, 4.4520, 3.9436, 3.9026, 1.8632, 3.7260], + device='cuda:3'), covar=tensor([0.1679, 0.1121, 0.2866, 0.1493, 0.2802, 0.1758, 0.5702, 0.2102], + device='cuda:3'), in_proj_covar=tensor([0.0256, 0.0231, 0.0274, 0.0325, 0.0320, 0.0268, 0.0284, 0.0283], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:30:08,958 INFO [finetune.py:976] (3/7) Epoch 2, batch 3550, loss[loss=0.2153, simple_loss=0.2664, pruned_loss=0.08208, over 4870.00 frames. ], tot_loss[loss=0.2583, simple_loss=0.3071, pruned_loss=0.1048, over 954929.44 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:30:12,102 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9283.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:30:37,458 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 2.002e+02 2.435e+02 2.997e+02 6.904e+02, threshold=4.871e+02, percent-clipped=3.0 +2023-04-26 12:30:48,982 INFO [finetune.py:976] (3/7) Epoch 2, batch 3600, loss[loss=0.2113, simple_loss=0.2622, pruned_loss=0.08015, over 4781.00 frames. ], tot_loss[loss=0.2561, simple_loss=0.3042, pruned_loss=0.104, over 956754.75 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:30:49,111 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4881, 2.9644, 1.2888, 1.6363, 2.1233, 1.6280, 3.5812, 1.9127], + device='cuda:3'), covar=tensor([0.0566, 0.0986, 0.0914, 0.1034, 0.0524, 0.0851, 0.0191, 0.0603], + device='cuda:3'), in_proj_covar=tensor([0.0057, 0.0074, 0.0055, 0.0051, 0.0056, 0.0056, 0.0087, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 12:30:50,314 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3363, 3.2095, 0.9812, 1.6569, 1.6992, 2.3116, 1.8993, 1.0175], + device='cuda:3'), covar=tensor([0.1553, 0.0967, 0.2241, 0.1500, 0.1320, 0.1116, 0.1646, 0.2125], + device='cuda:3'), in_proj_covar=tensor([0.0127, 0.0278, 0.0155, 0.0135, 0.0147, 0.0169, 0.0132, 0.0136], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 12:31:03,419 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9342.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:31:44,878 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9369.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:31:55,586 INFO [finetune.py:976] (3/7) Epoch 2, batch 3650, loss[loss=0.3252, simple_loss=0.3609, pruned_loss=0.1448, over 4760.00 frames. ], tot_loss[loss=0.2602, simple_loss=0.3082, pruned_loss=0.1061, over 956011.75 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:32:25,410 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9414.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:32:27,091 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9416.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:32:28,804 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.507e+02 2.166e+02 2.483e+02 2.970e+02 6.745e+02, threshold=4.965e+02, percent-clipped=1.0 +2023-04-26 12:32:34,828 INFO [finetune.py:976] (3/7) Epoch 2, batch 3700, loss[loss=0.2377, simple_loss=0.2955, pruned_loss=0.08995, over 4821.00 frames. ], tot_loss[loss=0.2619, simple_loss=0.3112, pruned_loss=0.1063, over 955935.53 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:32:40,468 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2805, 1.1878, 1.4588, 1.5287, 1.4733, 1.2495, 1.3597, 1.3662], + device='cuda:3'), covar=tensor([3.8677, 5.5948, 6.5979, 6.5181, 4.0249, 7.3273, 7.3000, 5.0420], + device='cuda:3'), in_proj_covar=tensor([0.0449, 0.0510, 0.0602, 0.0594, 0.0482, 0.0529, 0.0538, 0.0549], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:32:47,554 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-26 12:32:58,815 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9464.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:33:07,017 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9475.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:33:08,732 INFO [finetune.py:976] (3/7) Epoch 2, batch 3750, loss[loss=0.3171, simple_loss=0.3682, pruned_loss=0.133, over 4930.00 frames. ], tot_loss[loss=0.2649, simple_loss=0.3137, pruned_loss=0.1081, over 955598.12 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:33:47,547 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.577e+02 2.140e+02 2.529e+02 3.098e+02 5.505e+02, threshold=5.058e+02, percent-clipped=1.0 +2023-04-26 12:33:59,190 INFO [finetune.py:976] (3/7) Epoch 2, batch 3800, loss[loss=0.2439, simple_loss=0.302, pruned_loss=0.09289, over 4879.00 frames. ], tot_loss[loss=0.2627, simple_loss=0.3119, pruned_loss=0.1068, over 953331.49 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:34:20,226 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7005, 2.2110, 1.7170, 2.1545, 1.6632, 1.7503, 1.8925, 1.4392], + device='cuda:3'), covar=tensor([0.2194, 0.1436, 0.1279, 0.1479, 0.3256, 0.1610, 0.1883, 0.2835], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0336, 0.0246, 0.0312, 0.0315, 0.0288, 0.0278, 0.0302], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:34:55,619 INFO [finetune.py:976] (3/7) Epoch 2, batch 3850, loss[loss=0.2301, simple_loss=0.288, pruned_loss=0.08612, over 4867.00 frames. ], tot_loss[loss=0.2594, simple_loss=0.3091, pruned_loss=0.1048, over 954210.91 frames. ], batch size: 34, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:34:58,835 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9583.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:34:59,484 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9584.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:35:22,631 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 2.119e+02 2.508e+02 2.913e+02 5.368e+02, threshold=5.017e+02, percent-clipped=1.0 +2023-04-26 12:35:29,156 INFO [finetune.py:976] (3/7) Epoch 2, batch 3900, loss[loss=0.2411, simple_loss=0.289, pruned_loss=0.09666, over 4751.00 frames. ], tot_loss[loss=0.2568, simple_loss=0.3059, pruned_loss=0.1038, over 953979.08 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:35:29,852 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3076, 1.7812, 1.5267, 2.0485, 1.8480, 2.0834, 1.6303, 3.8144], + device='cuda:3'), covar=tensor([0.0699, 0.0758, 0.0852, 0.1225, 0.0652, 0.0544, 0.0741, 0.0154], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0047, 0.0042, 0.0041, 0.0041, 0.0066], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 12:35:38,161 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9631.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:35:50,795 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9642.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:35:52,632 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9645.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:36:02,300 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6856, 1.7356, 1.0437, 1.3795, 1.8908, 1.5954, 1.5054, 1.5348], + device='cuda:3'), covar=tensor([0.0599, 0.0440, 0.0433, 0.0632, 0.0329, 0.0582, 0.0596, 0.0717], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 12:36:12,996 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9658.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:36:26,762 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9669.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:36:38,413 INFO [finetune.py:976] (3/7) Epoch 2, batch 3950, loss[loss=0.2771, simple_loss=0.3089, pruned_loss=0.1227, over 4896.00 frames. ], tot_loss[loss=0.2523, simple_loss=0.3011, pruned_loss=0.1018, over 953483.48 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:36:55,197 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9690.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:37:17,361 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1529, 1.4392, 1.2940, 1.7102, 1.4861, 1.7049, 1.3283, 3.0958], + device='cuda:3'), covar=tensor([0.0679, 0.0820, 0.0815, 0.1251, 0.0663, 0.0559, 0.0776, 0.0200], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0047, 0.0042, 0.0041, 0.0041, 0.0066], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 12:37:29,748 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9717.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:37:36,036 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.382e+02 1.910e+02 2.412e+02 2.849e+02 4.927e+02, threshold=4.824e+02, percent-clipped=0.0 +2023-04-26 12:37:36,683 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9719.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:37:47,957 INFO [finetune.py:976] (3/7) Epoch 2, batch 4000, loss[loss=0.2634, simple_loss=0.3061, pruned_loss=0.1103, over 4709.00 frames. ], tot_loss[loss=0.2506, simple_loss=0.2994, pruned_loss=0.1009, over 954102.60 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:37:58,326 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9743.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:38:11,104 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1990, 0.8577, 1.1912, 1.4752, 1.3408, 1.1415, 1.1822, 1.1231], + device='cuda:3'), covar=tensor([3.4406, 4.5653, 5.2547, 5.8838, 3.6966, 5.8198, 5.8678, 4.2487], + device='cuda:3'), in_proj_covar=tensor([0.0447, 0.0507, 0.0598, 0.0594, 0.0480, 0.0527, 0.0534, 0.0547], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:38:15,587 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9770.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:38:20,830 INFO [finetune.py:976] (3/7) Epoch 2, batch 4050, loss[loss=0.272, simple_loss=0.3289, pruned_loss=0.1075, over 4740.00 frames. ], tot_loss[loss=0.2562, simple_loss=0.3049, pruned_loss=0.1037, over 952867.82 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:38:38,396 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9804.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:38:47,823 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 2.123e+02 2.473e+02 2.970e+02 7.758e+02, threshold=4.946e+02, percent-clipped=3.0 +2023-04-26 12:38:54,757 INFO [finetune.py:976] (3/7) Epoch 2, batch 4100, loss[loss=0.2695, simple_loss=0.3205, pruned_loss=0.1092, over 4901.00 frames. ], tot_loss[loss=0.2587, simple_loss=0.3082, pruned_loss=0.1046, over 954596.08 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:38:57,279 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9499, 1.9433, 2.1719, 2.2730, 2.4591, 1.8447, 1.2954, 1.9169], + device='cuda:3'), covar=tensor([0.1270, 0.1201, 0.0752, 0.0887, 0.0662, 0.1361, 0.1486, 0.0873], + device='cuda:3'), in_proj_covar=tensor([0.0213, 0.0212, 0.0191, 0.0185, 0.0181, 0.0201, 0.0178, 0.0196], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:39:33,881 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 12:39:34,296 INFO [finetune.py:976] (3/7) Epoch 2, batch 4150, loss[loss=0.2421, simple_loss=0.3162, pruned_loss=0.08402, over 4894.00 frames. ], tot_loss[loss=0.2624, simple_loss=0.3116, pruned_loss=0.1066, over 954346.96 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:39:46,981 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1746, 1.3527, 1.8302, 2.5136, 1.8012, 1.3956, 1.1854, 1.6956], + device='cuda:3'), covar=tensor([0.5811, 0.7222, 0.3340, 0.6372, 0.7503, 0.5277, 1.0036, 0.7253], + device='cuda:3'), in_proj_covar=tensor([0.0267, 0.0276, 0.0222, 0.0346, 0.0235, 0.0234, 0.0269, 0.0217], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:39:58,249 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1108, 0.7017, 0.9217, 1.4031, 1.2292, 1.0406, 1.0365, 1.0352], + device='cuda:3'), covar=tensor([3.6560, 4.8650, 5.5529, 6.1537, 3.9918, 6.1724, 5.8987, 4.4892], + device='cuda:3'), in_proj_covar=tensor([0.0449, 0.0508, 0.0600, 0.0596, 0.0481, 0.0527, 0.0535, 0.0548], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:40:18,084 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.356e+02 2.059e+02 2.438e+02 3.061e+02 5.791e+02, threshold=4.877e+02, percent-clipped=3.0 +2023-04-26 12:40:29,028 INFO [finetune.py:976] (3/7) Epoch 2, batch 4200, loss[loss=0.2746, simple_loss=0.3166, pruned_loss=0.1163, over 4717.00 frames. ], tot_loss[loss=0.2596, simple_loss=0.31, pruned_loss=0.1046, over 953219.25 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:40:43,344 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:41:36,131 INFO [finetune.py:976] (3/7) Epoch 2, batch 4250, loss[loss=0.2228, simple_loss=0.2802, pruned_loss=0.08269, over 4842.00 frames. ], tot_loss[loss=0.2563, simple_loss=0.3064, pruned_loss=0.1031, over 953580.64 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:41:36,262 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6559, 1.7441, 1.6435, 1.2693, 1.8412, 1.4952, 2.2951, 1.3927], + device='cuda:3'), covar=tensor([0.3991, 0.1523, 0.4572, 0.3257, 0.1698, 0.2101, 0.1323, 0.4440], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0351, 0.0437, 0.0370, 0.0404, 0.0375, 0.0402, 0.0414], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:42:30,339 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10014.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:42:33,325 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.966e+02 2.336e+02 2.838e+02 4.911e+02, threshold=4.671e+02, percent-clipped=1.0 +2023-04-26 12:42:43,920 INFO [finetune.py:976] (3/7) Epoch 2, batch 4300, loss[loss=0.2416, simple_loss=0.2863, pruned_loss=0.09846, over 4822.00 frames. ], tot_loss[loss=0.2543, simple_loss=0.3037, pruned_loss=0.1024, over 950842.69 frames. ], batch size: 30, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:43:02,744 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10039.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:43:18,131 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10062.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:43:23,022 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10070.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:43:27,739 INFO [finetune.py:976] (3/7) Epoch 2, batch 4350, loss[loss=0.2347, simple_loss=0.2925, pruned_loss=0.08844, over 4792.00 frames. ], tot_loss[loss=0.2525, simple_loss=0.3007, pruned_loss=0.1021, over 949904.56 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:43:31,503 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6589, 1.7206, 1.6783, 1.2361, 1.8270, 1.3973, 2.3326, 1.4548], + device='cuda:3'), covar=tensor([0.4203, 0.1531, 0.4699, 0.3205, 0.1721, 0.2318, 0.1464, 0.4220], + device='cuda:3'), in_proj_covar=tensor([0.0348, 0.0352, 0.0437, 0.0370, 0.0404, 0.0374, 0.0402, 0.0414], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:43:41,996 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10099.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:43:42,666 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10100.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:43:54,962 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10118.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:43:55,494 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.384e+02 2.038e+02 2.484e+02 2.915e+02 5.390e+02, threshold=4.968e+02, percent-clipped=2.0 +2023-04-26 12:43:58,077 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10123.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:43:59,292 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5200, 1.6421, 1.5878, 2.2317, 2.4183, 2.1438, 1.8990, 1.7164], + device='cuda:3'), covar=tensor([0.1733, 0.2419, 0.2353, 0.2395, 0.1403, 0.2296, 0.2734, 0.2323], + device='cuda:3'), in_proj_covar=tensor([0.0324, 0.0339, 0.0348, 0.0316, 0.0353, 0.0369, 0.0321, 0.0354], + device='cuda:3'), out_proj_covar=tensor([7.0753e-05, 7.2846e-05, 7.5288e-05, 6.6345e-05, 7.5219e-05, 8.0660e-05, + 6.9919e-05, 7.6458e-05], device='cuda:3') +2023-04-26 12:43:59,532 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 12:44:00,982 INFO [finetune.py:976] (3/7) Epoch 2, batch 4400, loss[loss=0.2661, simple_loss=0.3185, pruned_loss=0.1068, over 4886.00 frames. ], tot_loss[loss=0.2546, simple_loss=0.3032, pruned_loss=0.103, over 952802.93 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:44:30,040 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4765, 2.2652, 2.5105, 2.7216, 2.6093, 2.3649, 1.8920, 2.4484], + device='cuda:3'), covar=tensor([0.1031, 0.1030, 0.0610, 0.0671, 0.0689, 0.1021, 0.1143, 0.0655], + device='cuda:3'), in_proj_covar=tensor([0.0211, 0.0211, 0.0190, 0.0184, 0.0181, 0.0201, 0.0178, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:44:34,793 INFO [finetune.py:976] (3/7) Epoch 2, batch 4450, loss[loss=0.318, simple_loss=0.3498, pruned_loss=0.1431, over 4871.00 frames. ], tot_loss[loss=0.2574, simple_loss=0.3065, pruned_loss=0.1042, over 951833.73 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:45:03,115 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.522e+02 2.117e+02 2.495e+02 3.123e+02 6.793e+02, threshold=4.991e+02, percent-clipped=1.0 +2023-04-26 12:45:08,636 INFO [finetune.py:976] (3/7) Epoch 2, batch 4500, loss[loss=0.2569, simple_loss=0.3154, pruned_loss=0.09921, over 4906.00 frames. ], tot_loss[loss=0.2587, simple_loss=0.3084, pruned_loss=0.1046, over 952498.70 frames. ], batch size: 46, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:45:08,735 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8932, 2.3375, 1.0362, 1.1471, 1.7216, 1.0917, 3.0047, 1.4202], + device='cuda:3'), covar=tensor([0.0759, 0.0691, 0.0851, 0.1326, 0.0524, 0.1115, 0.0250, 0.0747], + device='cuda:3'), in_proj_covar=tensor([0.0056, 0.0073, 0.0054, 0.0050, 0.0055, 0.0056, 0.0086, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 12:45:15,988 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10240.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:45:26,436 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-04-26 12:45:37,121 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 12:45:42,229 INFO [finetune.py:976] (3/7) Epoch 2, batch 4550, loss[loss=0.226, simple_loss=0.2734, pruned_loss=0.08927, over 4726.00 frames. ], tot_loss[loss=0.2593, simple_loss=0.3094, pruned_loss=0.1046, over 953212.01 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:45:48,411 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10288.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:46:21,450 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10314.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:46:28,743 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5678, 1.4767, 0.7918, 1.2480, 1.6606, 1.4707, 1.3207, 1.4407], + device='cuda:3'), covar=tensor([0.0613, 0.0478, 0.0512, 0.0663, 0.0368, 0.0615, 0.0645, 0.0718], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0051], + device='cuda:3') +2023-04-26 12:46:29,829 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.925e+02 2.197e+02 2.697e+02 5.130e+02, threshold=4.395e+02, percent-clipped=1.0 +2023-04-26 12:46:41,914 INFO [finetune.py:976] (3/7) Epoch 2, batch 4600, loss[loss=0.2571, simple_loss=0.3071, pruned_loss=0.1036, over 4766.00 frames. ], tot_loss[loss=0.2572, simple_loss=0.3081, pruned_loss=0.1032, over 955396.19 frames. ], batch size: 28, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:47:03,393 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5905, 1.3328, 4.1586, 3.8780, 3.6896, 3.9201, 3.8556, 3.6534], + device='cuda:3'), covar=tensor([0.6678, 0.5852, 0.1071, 0.1754, 0.1090, 0.1442, 0.2047, 0.1409], + device='cuda:3'), in_proj_covar=tensor([0.0333, 0.0313, 0.0446, 0.0451, 0.0378, 0.0433, 0.0340, 0.0398], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:47:09,972 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10362.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:47:29,150 INFO [finetune.py:976] (3/7) Epoch 2, batch 4650, loss[loss=0.2529, simple_loss=0.3018, pruned_loss=0.102, over 4911.00 frames. ], tot_loss[loss=0.2545, simple_loss=0.3051, pruned_loss=0.102, over 955942.71 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:47:40,074 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3764, 1.2607, 3.7881, 3.5371, 3.3867, 3.5797, 3.6013, 3.3455], + device='cuda:3'), covar=tensor([0.6389, 0.5828, 0.1139, 0.1901, 0.1092, 0.1697, 0.1820, 0.1593], + device='cuda:3'), in_proj_covar=tensor([0.0331, 0.0311, 0.0445, 0.0449, 0.0377, 0.0431, 0.0339, 0.0397], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:47:50,960 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10395.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:47:58,953 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10399.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:48:12,185 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 12:48:12,473 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10409.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:48:23,999 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10418.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:48:25,137 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 2.119e+02 2.411e+02 2.762e+02 5.438e+02, threshold=4.823e+02, percent-clipped=2.0 +2023-04-26 12:48:35,916 INFO [finetune.py:976] (3/7) Epoch 2, batch 4700, loss[loss=0.1658, simple_loss=0.2296, pruned_loss=0.05097, over 4764.00 frames. ], tot_loss[loss=0.2519, simple_loss=0.3022, pruned_loss=0.1008, over 957087.76 frames. ], batch size: 28, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:48:42,901 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2126, 1.4140, 1.4041, 1.4907, 1.4690, 1.6411, 1.4870, 1.4862], + device='cuda:3'), covar=tensor([1.7619, 3.3102, 2.6283, 2.2502, 2.4657, 4.2508, 3.3434, 2.7349], + device='cuda:3'), in_proj_covar=tensor([0.0289, 0.0390, 0.0310, 0.0314, 0.0342, 0.0385, 0.0375, 0.0339], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:48:53,255 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10447.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:48:57,648 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10454.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:49:09,536 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10470.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:49:09,618 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-26 12:49:15,193 INFO [finetune.py:976] (3/7) Epoch 2, batch 4750, loss[loss=0.2851, simple_loss=0.3217, pruned_loss=0.1242, over 4860.00 frames. ], tot_loss[loss=0.2467, simple_loss=0.2972, pruned_loss=0.09804, over 958329.89 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:49:23,778 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8859, 1.3608, 4.9788, 4.6071, 4.3644, 4.7280, 4.4372, 4.3599], + device='cuda:3'), covar=tensor([0.6735, 0.6190, 0.1059, 0.1924, 0.1062, 0.1480, 0.1613, 0.1532], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0311, 0.0443, 0.0447, 0.0377, 0.0431, 0.0337, 0.0397], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:49:32,952 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5840, 1.7002, 1.4645, 1.6605, 1.5960, 1.8729, 1.6344, 1.6002], + device='cuda:3'), covar=tensor([1.7499, 3.6399, 2.9386, 2.4871, 2.7079, 4.1819, 3.7139, 3.1854], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0388, 0.0309, 0.0313, 0.0341, 0.0383, 0.0373, 0.0338], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:49:37,012 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9542, 1.2696, 1.6920, 2.2606, 1.6368, 1.2968, 1.1145, 1.4901], + device='cuda:3'), covar=tensor([0.5472, 0.6701, 0.3200, 0.5494, 0.6656, 0.4968, 0.9121, 0.6944], + device='cuda:3'), in_proj_covar=tensor([0.0270, 0.0277, 0.0223, 0.0347, 0.0235, 0.0235, 0.0269, 0.0216], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:49:40,440 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10515.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 12:49:43,424 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.363e+02 1.957e+02 2.292e+02 2.912e+02 6.222e+02, threshold=4.583e+02, percent-clipped=3.0 +2023-04-26 12:49:49,337 INFO [finetune.py:976] (3/7) Epoch 2, batch 4800, loss[loss=0.2959, simple_loss=0.3327, pruned_loss=0.1296, over 4904.00 frames. ], tot_loss[loss=0.2491, simple_loss=0.2994, pruned_loss=0.09941, over 957278.81 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:50:23,191 INFO [finetune.py:976] (3/7) Epoch 2, batch 4850, loss[loss=0.2637, simple_loss=0.2964, pruned_loss=0.1155, over 4349.00 frames. ], tot_loss[loss=0.2521, simple_loss=0.303, pruned_loss=0.1006, over 956924.03 frames. ], batch size: 19, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:50:31,016 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4689, 2.4414, 2.7209, 2.9915, 2.9120, 2.3647, 1.7968, 2.5441], + device='cuda:3'), covar=tensor([0.1108, 0.1038, 0.0583, 0.0700, 0.0592, 0.1146, 0.1189, 0.0716], + device='cuda:3'), in_proj_covar=tensor([0.0214, 0.0213, 0.0191, 0.0186, 0.0183, 0.0203, 0.0179, 0.0197], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:50:44,370 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7061, 1.3850, 4.1761, 3.9341, 3.7216, 3.8921, 3.8162, 3.7168], + device='cuda:3'), covar=tensor([0.6669, 0.5296, 0.1022, 0.1470, 0.0942, 0.1433, 0.2385, 0.1462], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0313, 0.0446, 0.0449, 0.0377, 0.0433, 0.0339, 0.0400], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:50:50,776 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 2.100e+02 2.315e+02 2.802e+02 4.090e+02, threshold=4.629e+02, percent-clipped=1.0 +2023-04-26 12:50:56,078 INFO [finetune.py:976] (3/7) Epoch 2, batch 4900, loss[loss=0.234, simple_loss=0.277, pruned_loss=0.09554, over 4747.00 frames. ], tot_loss[loss=0.2544, simple_loss=0.3054, pruned_loss=0.1017, over 956949.09 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:51:46,561 INFO [finetune.py:976] (3/7) Epoch 2, batch 4950, loss[loss=0.2224, simple_loss=0.2898, pruned_loss=0.07753, over 4917.00 frames. ], tot_loss[loss=0.2549, simple_loss=0.3061, pruned_loss=0.1019, over 954891.82 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:52:02,825 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3144, 0.9140, 1.2610, 1.6166, 1.4884, 1.2575, 1.2684, 1.2459], + device='cuda:3'), covar=tensor([3.8172, 5.1519, 6.2798, 7.4279, 4.0343, 6.2394, 6.5844, 4.6669], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0506, 0.0599, 0.0597, 0.0481, 0.0523, 0.0534, 0.0546], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:52:10,676 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10695.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:52:42,699 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10718.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:52:43,792 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.438e+02 1.913e+02 2.320e+02 2.882e+02 5.075e+02, threshold=4.640e+02, percent-clipped=3.0 +2023-04-26 12:52:54,985 INFO [finetune.py:976] (3/7) Epoch 2, batch 5000, loss[loss=0.2245, simple_loss=0.2757, pruned_loss=0.08669, over 4904.00 frames. ], tot_loss[loss=0.2522, simple_loss=0.3032, pruned_loss=0.1006, over 953692.17 frames. ], batch size: 46, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:53:16,589 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10743.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:53:16,720 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-26 12:53:17,249 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10744.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:53:17,277 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5177, 1.2618, 1.6548, 1.5718, 1.3747, 1.2186, 1.3821, 0.9565], + device='cuda:3'), covar=tensor([0.0757, 0.1156, 0.0900, 0.0905, 0.1063, 0.1590, 0.0747, 0.1235], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0079, 0.0075, 0.0072, 0.0085, 0.0095, 0.0088, 0.0080], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 12:53:18,507 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7735, 1.4308, 1.3257, 1.3955, 2.0126, 1.6869, 1.3066, 1.3405], + device='cuda:3'), covar=tensor([0.1482, 0.1785, 0.2604, 0.1867, 0.0801, 0.1757, 0.2333, 0.1857], + device='cuda:3'), in_proj_covar=tensor([0.0325, 0.0341, 0.0349, 0.0319, 0.0354, 0.0370, 0.0323, 0.0355], + device='cuda:3'), out_proj_covar=tensor([7.0995e-05, 7.3253e-05, 7.5532e-05, 6.6971e-05, 7.5515e-05, 8.0910e-05, + 7.0276e-05, 7.6839e-05], device='cuda:3') +2023-04-26 12:53:29,933 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:53:30,549 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10766.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:53:42,275 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10774.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:53:45,587 INFO [finetune.py:976] (3/7) Epoch 2, batch 5050, loss[loss=0.2015, simple_loss=0.2518, pruned_loss=0.07559, over 4855.00 frames. ], tot_loss[loss=0.2495, simple_loss=0.2998, pruned_loss=0.09958, over 954721.55 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:54:03,777 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7309, 3.7412, 2.7429, 4.2367, 3.7596, 3.7310, 1.7518, 3.5634], + device='cuda:3'), covar=tensor([0.1848, 0.1183, 0.2860, 0.1669, 0.2636, 0.1885, 0.5533, 0.2333], + device='cuda:3'), in_proj_covar=tensor([0.0255, 0.0228, 0.0271, 0.0320, 0.0316, 0.0266, 0.0279, 0.0283], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:54:25,687 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10805.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:54:28,748 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10810.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:54:40,510 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.434e+02 1.942e+02 2.267e+02 2.719e+02 4.388e+02, threshold=4.533e+02, percent-clipped=0.0 +2023-04-26 12:54:47,727 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6370, 1.2651, 4.4715, 4.1320, 3.9122, 4.2279, 4.2012, 3.9649], + device='cuda:3'), covar=tensor([0.6969, 0.6267, 0.0996, 0.1705, 0.1132, 0.1571, 0.1076, 0.1370], + device='cuda:3'), in_proj_covar=tensor([0.0333, 0.0313, 0.0445, 0.0449, 0.0377, 0.0431, 0.0339, 0.0399], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 12:54:51,195 INFO [finetune.py:976] (3/7) Epoch 2, batch 5100, loss[loss=0.2112, simple_loss=0.2605, pruned_loss=0.08094, over 4905.00 frames. ], tot_loss[loss=0.2461, simple_loss=0.2966, pruned_loss=0.09779, over 954317.19 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:55:01,985 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10835.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:55:12,302 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7499, 1.7218, 1.5919, 1.3191, 1.7779, 1.4310, 2.2675, 1.3045], + device='cuda:3'), covar=tensor([0.3360, 0.1319, 0.4418, 0.2632, 0.1555, 0.2011, 0.1341, 0.4209], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0351, 0.0435, 0.0369, 0.0404, 0.0375, 0.0401, 0.0415], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:55:45,708 INFO [finetune.py:976] (3/7) Epoch 2, batch 5150, loss[loss=0.2802, simple_loss=0.3244, pruned_loss=0.118, over 4742.00 frames. ], tot_loss[loss=0.2471, simple_loss=0.297, pruned_loss=0.09857, over 956519.49 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:55:47,700 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9267, 1.3573, 1.8086, 2.2494, 1.6686, 1.3660, 1.1234, 1.5745], + device='cuda:3'), covar=tensor([0.5868, 0.6744, 0.3141, 0.5391, 0.6580, 0.4961, 0.9571, 0.6070], + device='cuda:3'), in_proj_covar=tensor([0.0268, 0.0275, 0.0222, 0.0345, 0.0233, 0.0234, 0.0267, 0.0215], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 12:55:48,306 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1214, 2.0672, 1.8232, 1.7519, 2.1925, 1.6860, 2.8113, 1.6283], + device='cuda:3'), covar=tensor([0.4155, 0.1687, 0.4547, 0.3097, 0.1968, 0.2722, 0.1290, 0.4125], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0350, 0.0433, 0.0368, 0.0402, 0.0374, 0.0399, 0.0414], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:55:50,630 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10885.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:56:13,138 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.326e+02 2.074e+02 2.408e+02 2.882e+02 5.966e+02, threshold=4.815e+02, percent-clipped=3.0 +2023-04-26 12:56:18,493 INFO [finetune.py:976] (3/7) Epoch 2, batch 5200, loss[loss=0.251, simple_loss=0.3053, pruned_loss=0.09834, over 4843.00 frames. ], tot_loss[loss=0.2515, simple_loss=0.3019, pruned_loss=0.1005, over 955300.93 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:56:22,134 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2675, 1.5218, 1.1752, 1.4256, 1.2917, 1.1707, 1.2802, 1.0454], + device='cuda:3'), covar=tensor([0.1771, 0.1452, 0.1306, 0.1424, 0.3127, 0.1516, 0.1781, 0.2493], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0338, 0.0247, 0.0312, 0.0316, 0.0290, 0.0280, 0.0303], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:56:31,989 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10946.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:56:57,990 INFO [finetune.py:976] (3/7) Epoch 2, batch 5250, loss[loss=0.2567, simple_loss=0.3115, pruned_loss=0.101, over 4836.00 frames. ], tot_loss[loss=0.2524, simple_loss=0.3038, pruned_loss=0.1005, over 955574.14 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:57:40,211 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.533e+02 2.106e+02 2.454e+02 2.928e+02 5.017e+02, threshold=4.909e+02, percent-clipped=1.0 +2023-04-26 12:57:48,218 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 12:57:51,419 INFO [finetune.py:976] (3/7) Epoch 2, batch 5300, loss[loss=0.2166, simple_loss=0.2827, pruned_loss=0.07526, over 4849.00 frames. ], tot_loss[loss=0.2518, simple_loss=0.3034, pruned_loss=0.1001, over 952063.09 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:58:00,771 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1808, 0.5328, 0.8528, 1.5116, 1.3232, 1.0621, 1.0601, 1.0096], + device='cuda:3'), covar=tensor([3.5448, 4.7511, 5.9020, 5.9311, 3.9413, 5.6370, 5.7419, 3.8428], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0506, 0.0600, 0.0598, 0.0481, 0.0523, 0.0533, 0.0544], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:58:35,597 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0963, 2.9044, 2.3040, 2.5463, 2.1221, 2.4214, 2.6114, 1.9499], + device='cuda:3'), covar=tensor([0.2989, 0.2040, 0.1352, 0.1997, 0.3678, 0.1883, 0.2740, 0.3792], + device='cuda:3'), in_proj_covar=tensor([0.0320, 0.0341, 0.0250, 0.0315, 0.0319, 0.0292, 0.0282, 0.0306], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:58:45,113 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11065.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:58:57,625 INFO [finetune.py:976] (3/7) Epoch 2, batch 5350, loss[loss=0.2568, simple_loss=0.3134, pruned_loss=0.1001, over 4822.00 frames. ], tot_loss[loss=0.2513, simple_loss=0.3038, pruned_loss=0.09941, over 954100.64 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:59:08,260 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1204, 0.6877, 0.8976, 0.7175, 1.2731, 0.9926, 0.7749, 1.0265], + device='cuda:3'), covar=tensor([0.1811, 0.1765, 0.2445, 0.1982, 0.0957, 0.1640, 0.2171, 0.2110], + device='cuda:3'), in_proj_covar=tensor([0.0324, 0.0340, 0.0348, 0.0317, 0.0351, 0.0367, 0.0321, 0.0355], + device='cuda:3'), out_proj_covar=tensor([7.0678e-05, 7.2999e-05, 7.5300e-05, 6.6536e-05, 7.4786e-05, 8.0192e-05, + 7.0007e-05, 7.6717e-05], device='cuda:3') +2023-04-26 12:59:15,183 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4678, 1.2933, 1.6574, 1.6642, 1.6370, 1.3822, 1.5143, 1.4709], + device='cuda:3'), covar=tensor([2.8950, 4.0144, 4.7702, 5.1889, 3.0732, 5.1213, 5.2283, 4.0228], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0507, 0.0602, 0.0599, 0.0482, 0.0524, 0.0535, 0.0545], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 12:59:29,511 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11100.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:59:47,703 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11110.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:59:47,995 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 12:59:50,022 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11113.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:00:00,706 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.273e+02 1.923e+02 2.240e+02 2.858e+02 6.419e+02, threshold=4.480e+02, percent-clipped=4.0 +2023-04-26 13:00:09,381 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2393, 1.4597, 1.8865, 2.5836, 1.8381, 1.4271, 1.2825, 1.8159], + device='cuda:3'), covar=tensor([0.5490, 0.6883, 0.3220, 0.5817, 0.7222, 0.5497, 0.9263, 0.6764], + device='cuda:3'), in_proj_covar=tensor([0.0269, 0.0276, 0.0223, 0.0345, 0.0233, 0.0235, 0.0267, 0.0214], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:00:11,118 INFO [finetune.py:976] (3/7) Epoch 2, batch 5400, loss[loss=0.228, simple_loss=0.2781, pruned_loss=0.08894, over 3992.00 frames. ], tot_loss[loss=0.2486, simple_loss=0.3009, pruned_loss=0.09814, over 954393.34 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:00:12,411 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:00:32,458 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11158.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:00:37,255 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7497, 1.7801, 1.9727, 2.2242, 2.1821, 1.6876, 1.3043, 1.9149], + device='cuda:3'), covar=tensor([0.1213, 0.1132, 0.0644, 0.0656, 0.0717, 0.1117, 0.1326, 0.0697], + device='cuda:3'), in_proj_covar=tensor([0.0214, 0.0214, 0.0192, 0.0188, 0.0183, 0.0203, 0.0180, 0.0199], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:00:42,742 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11174.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:00:45,043 INFO [finetune.py:976] (3/7) Epoch 2, batch 5450, loss[loss=0.2633, simple_loss=0.3123, pruned_loss=0.1071, over 4894.00 frames. ], tot_loss[loss=0.245, simple_loss=0.297, pruned_loss=0.0965, over 953539.67 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:01:00,517 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3845, 1.6557, 1.5622, 2.2195, 1.8877, 2.1883, 1.5134, 4.5583], + device='cuda:3'), covar=tensor([0.0646, 0.0825, 0.0849, 0.1168, 0.0658, 0.0540, 0.0802, 0.0129], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0041, 0.0041, 0.0041, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 13:01:18,381 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.436e+02 1.901e+02 2.286e+02 2.895e+02 5.867e+02, threshold=4.573e+02, percent-clipped=4.0 +2023-04-26 13:01:23,720 INFO [finetune.py:976] (3/7) Epoch 2, batch 5500, loss[loss=0.2063, simple_loss=0.2663, pruned_loss=0.07315, over 4895.00 frames. ], tot_loss[loss=0.2423, simple_loss=0.294, pruned_loss=0.09532, over 954967.28 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:01:28,096 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11235.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:01:31,706 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11241.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:01:57,530 INFO [finetune.py:976] (3/7) Epoch 2, batch 5550, loss[loss=0.21, simple_loss=0.2507, pruned_loss=0.08465, over 3810.00 frames. ], tot_loss[loss=0.2441, simple_loss=0.2955, pruned_loss=0.09632, over 956141.69 frames. ], batch size: 16, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:02:01,940 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4748, 1.0610, 1.1894, 0.9945, 1.6878, 1.3416, 1.0015, 1.2073], + device='cuda:3'), covar=tensor([0.1687, 0.1444, 0.1850, 0.1504, 0.0818, 0.1265, 0.1797, 0.1788], + device='cuda:3'), in_proj_covar=tensor([0.0324, 0.0340, 0.0349, 0.0318, 0.0351, 0.0366, 0.0321, 0.0356], + device='cuda:3'), out_proj_covar=tensor([7.0628e-05, 7.2964e-05, 7.5510e-05, 6.6766e-05, 7.4778e-05, 8.0000e-05, + 6.9957e-05, 7.6957e-05], device='cuda:3') +2023-04-26 13:02:02,587 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.07 vs. limit=5.0 +2023-04-26 13:02:31,826 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3452, 3.1155, 0.9713, 1.7015, 1.6447, 2.3390, 1.9572, 1.0848], + device='cuda:3'), covar=tensor([0.1629, 0.1176, 0.2298, 0.1660, 0.1348, 0.1206, 0.1501, 0.2298], + device='cuda:3'), in_proj_covar=tensor([0.0125, 0.0272, 0.0152, 0.0133, 0.0144, 0.0165, 0.0129, 0.0134], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:02:35,254 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.528e+02 2.026e+02 2.364e+02 2.674e+02 3.741e+02, threshold=4.728e+02, percent-clipped=0.0 +2023-04-26 13:02:39,950 INFO [finetune.py:976] (3/7) Epoch 2, batch 5600, loss[loss=0.2294, simple_loss=0.2701, pruned_loss=0.09435, over 4751.00 frames. ], tot_loss[loss=0.2467, simple_loss=0.2996, pruned_loss=0.09692, over 957914.37 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:03:30,851 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.21 vs. limit=5.0 +2023-04-26 13:03:42,987 INFO [finetune.py:976] (3/7) Epoch 2, batch 5650, loss[loss=0.229, simple_loss=0.2892, pruned_loss=0.0844, over 4833.00 frames. ], tot_loss[loss=0.2484, simple_loss=0.3016, pruned_loss=0.09763, over 953834.49 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:04:07,968 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9903, 2.0829, 2.0728, 2.6834, 2.8961, 2.6646, 2.5007, 2.3219], + device='cuda:3'), covar=tensor([0.1321, 0.1856, 0.2063, 0.1645, 0.1182, 0.1651, 0.2921, 0.2153], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0339, 0.0349, 0.0317, 0.0351, 0.0365, 0.0321, 0.0354], + device='cuda:3'), out_proj_covar=tensor([7.0502e-05, 7.2865e-05, 7.5471e-05, 6.6633e-05, 7.4779e-05, 7.9832e-05, + 6.9942e-05, 7.6628e-05], device='cuda:3') +2023-04-26 13:04:13,905 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11400.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:04:27,170 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 13:04:31,185 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.867e+02 2.284e+02 2.863e+02 5.268e+02, threshold=4.568e+02, percent-clipped=1.0 +2023-04-26 13:04:35,905 INFO [finetune.py:976] (3/7) Epoch 2, batch 5700, loss[loss=0.2167, simple_loss=0.2659, pruned_loss=0.08372, over 4357.00 frames. ], tot_loss[loss=0.2462, simple_loss=0.2978, pruned_loss=0.09728, over 938410.68 frames. ], batch size: 19, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:04:37,186 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11430.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:04:47,962 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11448.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:05:07,855 INFO [finetune.py:976] (3/7) Epoch 3, batch 0, loss[loss=0.3208, simple_loss=0.3682, pruned_loss=0.1367, over 4737.00 frames. ], tot_loss[loss=0.3208, simple_loss=0.3682, pruned_loss=0.1367, over 4737.00 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:05:07,855 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 13:05:24,857 INFO [finetune.py:1010] (3/7) Epoch 3, validation: loss=0.1779, simple_loss=0.251, pruned_loss=0.05243, over 2265189.00 frames. +2023-04-26 13:05:24,858 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 13:05:42,225 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11478.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:06:01,719 INFO [finetune.py:976] (3/7) Epoch 3, batch 50, loss[loss=0.2423, simple_loss=0.2993, pruned_loss=0.09269, over 4924.00 frames. ], tot_loss[loss=0.2501, simple_loss=0.3033, pruned_loss=0.09844, over 216214.25 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:06:01,851 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4911, 2.0105, 1.5176, 1.2365, 1.1539, 1.1923, 1.5218, 1.1485], + device='cuda:3'), covar=tensor([0.2787, 0.2307, 0.2755, 0.3435, 0.4093, 0.3179, 0.2034, 0.3302], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0225, 0.0193, 0.0215, 0.0231, 0.0195, 0.0189, 0.0208], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:06:08,311 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 13:06:11,126 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.469e+02 1.955e+02 2.272e+02 2.719e+02 4.720e+02, threshold=4.545e+02, percent-clipped=1.0 +2023-04-26 13:06:17,277 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11530.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:06:18,550 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11532.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:06:24,402 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11541.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 13:06:33,178 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2731, 1.5430, 1.1706, 1.4563, 1.3448, 1.1383, 1.3069, 1.0594], + device='cuda:3'), covar=tensor([0.2131, 0.1579, 0.1464, 0.1474, 0.3569, 0.1835, 0.2049, 0.2825], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0338, 0.0247, 0.0309, 0.0316, 0.0289, 0.0277, 0.0302], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:06:34,900 INFO [finetune.py:976] (3/7) Epoch 3, batch 100, loss[loss=0.2506, simple_loss=0.2954, pruned_loss=0.1029, over 4806.00 frames. ], tot_loss[loss=0.2421, simple_loss=0.2945, pruned_loss=0.0948, over 380011.63 frames. ], batch size: 41, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:06:55,969 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11589.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:06:57,704 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3287, 1.4779, 1.4287, 1.5812, 1.5473, 1.7400, 1.5745, 1.5508], + device='cuda:3'), covar=tensor([1.5687, 2.9206, 2.4692, 2.0937, 2.4352, 3.8149, 3.0751, 2.5612], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0393, 0.0311, 0.0316, 0.0344, 0.0388, 0.0377, 0.0341], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:06:58,893 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11593.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:07:08,541 INFO [finetune.py:976] (3/7) Epoch 3, batch 150, loss[loss=0.2129, simple_loss=0.2749, pruned_loss=0.07546, over 4819.00 frames. ], tot_loss[loss=0.2396, simple_loss=0.2907, pruned_loss=0.09422, over 509874.01 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:07:18,014 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.434e+02 1.865e+02 2.222e+02 2.672e+02 4.139e+02, threshold=4.445e+02, percent-clipped=0.0 +2023-04-26 13:07:27,302 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11635.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:07:42,007 INFO [finetune.py:976] (3/7) Epoch 3, batch 200, loss[loss=0.2236, simple_loss=0.2702, pruned_loss=0.08852, over 4690.00 frames. ], tot_loss[loss=0.239, simple_loss=0.2898, pruned_loss=0.09412, over 610103.35 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:08:09,502 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-26 13:08:18,259 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-26 13:08:29,880 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11696.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:08:31,716 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11699.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:08:42,574 INFO [finetune.py:976] (3/7) Epoch 3, batch 250, loss[loss=0.2986, simple_loss=0.3487, pruned_loss=0.1243, over 4731.00 frames. ], tot_loss[loss=0.2426, simple_loss=0.2936, pruned_loss=0.09575, over 689203.13 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:09:02,120 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.384e+02 2.012e+02 2.336e+02 2.910e+02 4.662e+02, threshold=4.672e+02, percent-clipped=2.0 +2023-04-26 13:09:36,496 INFO [finetune.py:976] (3/7) Epoch 3, batch 300, loss[loss=0.2299, simple_loss=0.2877, pruned_loss=0.08606, over 4788.00 frames. ], tot_loss[loss=0.2455, simple_loss=0.2974, pruned_loss=0.09685, over 748439.11 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:09:40,554 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11760.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:10:10,602 INFO [finetune.py:976] (3/7) Epoch 3, batch 350, loss[loss=0.2332, simple_loss=0.2851, pruned_loss=0.09062, over 4218.00 frames. ], tot_loss[loss=0.245, simple_loss=0.2976, pruned_loss=0.09624, over 791809.90 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:10:18,385 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11816.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:10:21,168 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.314e+02 2.028e+02 2.307e+02 2.756e+02 7.160e+02, threshold=4.615e+02, percent-clipped=2.0 +2023-04-26 13:10:28,338 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11830.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:10:38,196 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11846.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:10:50,161 INFO [finetune.py:976] (3/7) Epoch 3, batch 400, loss[loss=0.2572, simple_loss=0.3145, pruned_loss=0.09996, over 4901.00 frames. ], tot_loss[loss=0.2471, simple_loss=0.3005, pruned_loss=0.09684, over 829041.61 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:11:20,862 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11877.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:11:21,408 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11878.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:11:33,329 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11888.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:11:44,733 INFO [finetune.py:976] (3/7) Epoch 3, batch 450, loss[loss=0.2507, simple_loss=0.3086, pruned_loss=0.09642, over 4904.00 frames. ], tot_loss[loss=0.2453, simple_loss=0.2994, pruned_loss=0.09554, over 859003.49 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:11:45,496 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11907.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:11:54,740 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.367e+02 1.957e+02 2.306e+02 2.695e+02 4.680e+02, threshold=4.613e+02, percent-clipped=1.0 +2023-04-26 13:11:54,854 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11920.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:06,614 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0726, 2.4353, 1.0933, 1.2278, 2.0151, 1.3126, 2.9558, 1.6206], + device='cuda:3'), covar=tensor([0.0643, 0.0621, 0.0772, 0.1235, 0.0455, 0.0990, 0.0287, 0.0659], + device='cuda:3'), in_proj_covar=tensor([0.0056, 0.0073, 0.0054, 0.0050, 0.0056, 0.0056, 0.0086, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 13:12:13,311 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11948.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:18,146 INFO [finetune.py:976] (3/7) Epoch 3, batch 500, loss[loss=0.2117, simple_loss=0.2665, pruned_loss=0.07841, over 4831.00 frames. ], tot_loss[loss=0.2421, simple_loss=0.2961, pruned_loss=0.0941, over 882379.31 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:12:36,367 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11981.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:38,125 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11984.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:42,329 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11991.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:52,574 INFO [finetune.py:976] (3/7) Epoch 3, batch 550, loss[loss=0.2266, simple_loss=0.2889, pruned_loss=0.08217, over 4758.00 frames. ], tot_loss[loss=0.2386, simple_loss=0.2916, pruned_loss=0.0928, over 899130.80 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:12:55,028 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12009.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:55,635 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6008, 1.5889, 0.6541, 1.2473, 1.6129, 1.5025, 1.3587, 1.3923], + device='cuda:3'), covar=tensor([0.0597, 0.0433, 0.0489, 0.0628, 0.0344, 0.0572, 0.0570, 0.0700], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0037, 0.0047, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 13:13:02,613 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.347e+02 1.896e+02 2.113e+02 2.653e+02 4.840e+02, threshold=4.226e+02, percent-clipped=1.0 +2023-04-26 13:13:19,326 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12045.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:13:30,905 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12055.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:13:31,437 INFO [finetune.py:976] (3/7) Epoch 3, batch 600, loss[loss=0.2857, simple_loss=0.347, pruned_loss=0.1122, over 4760.00 frames. ], tot_loss[loss=0.2405, simple_loss=0.2929, pruned_loss=0.09407, over 911345.12 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:13:41,335 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.8357, 4.7729, 3.3365, 5.5061, 4.8885, 4.6742, 2.6245, 4.6974], + device='cuda:3'), covar=tensor([0.1210, 0.0808, 0.2671, 0.0706, 0.2748, 0.1597, 0.4617, 0.1744], + device='cuda:3'), in_proj_covar=tensor([0.0254, 0.0229, 0.0270, 0.0322, 0.0314, 0.0266, 0.0281, 0.0282], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:14:01,471 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12076.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:14:35,133 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5006, 1.3524, 1.8026, 1.7081, 1.3602, 1.0981, 1.5519, 1.0108], + device='cuda:3'), covar=tensor([0.0948, 0.0927, 0.0586, 0.0890, 0.1043, 0.1487, 0.0799, 0.1242], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0079, 0.0075, 0.0071, 0.0085, 0.0096, 0.0088, 0.0079], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 13:14:37,616 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9728, 1.3776, 1.2284, 1.7307, 1.9406, 1.7322, 1.6089, 1.3849], + device='cuda:3'), covar=tensor([0.2022, 0.2114, 0.2176, 0.1925, 0.1466, 0.2195, 0.2729, 0.2157], + device='cuda:3'), in_proj_covar=tensor([0.0322, 0.0339, 0.0348, 0.0315, 0.0349, 0.0365, 0.0321, 0.0354], + device='cuda:3'), out_proj_covar=tensor([7.0136e-05, 7.2849e-05, 7.5477e-05, 6.6034e-05, 7.4414e-05, 7.9720e-05, + 6.9814e-05, 7.6582e-05], device='cuda:3') +2023-04-26 13:14:38,094 INFO [finetune.py:976] (3/7) Epoch 3, batch 650, loss[loss=0.2464, simple_loss=0.3051, pruned_loss=0.09386, over 4708.00 frames. ], tot_loss[loss=0.2428, simple_loss=0.2952, pruned_loss=0.09518, over 918767.04 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:14:47,756 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8789, 1.4831, 1.3252, 1.4096, 1.9752, 1.7511, 1.3865, 1.3099], + device='cuda:3'), covar=tensor([0.1656, 0.1817, 0.2413, 0.1722, 0.1171, 0.1745, 0.2488, 0.2282], + device='cuda:3'), in_proj_covar=tensor([0.0322, 0.0339, 0.0348, 0.0315, 0.0349, 0.0365, 0.0320, 0.0354], + device='cuda:3'), out_proj_covar=tensor([7.0107e-05, 7.2830e-05, 7.5509e-05, 6.6018e-05, 7.4423e-05, 7.9756e-05, + 6.9803e-05, 7.6571e-05], device='cuda:3') +2023-04-26 13:14:55,298 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.956e+02 2.261e+02 2.773e+02 5.813e+02, threshold=4.521e+02, percent-clipped=3.0 +2023-04-26 13:15:18,810 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12137.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:15:30,238 INFO [finetune.py:976] (3/7) Epoch 3, batch 700, loss[loss=0.2358, simple_loss=0.2904, pruned_loss=0.09053, over 4891.00 frames. ], tot_loss[loss=0.2451, simple_loss=0.2977, pruned_loss=0.09621, over 924077.66 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:15:40,476 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12172.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:15:52,191 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12188.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:16:00,821 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12202.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:16:03,162 INFO [finetune.py:976] (3/7) Epoch 3, batch 750, loss[loss=0.2565, simple_loss=0.3075, pruned_loss=0.1027, over 4761.00 frames. ], tot_loss[loss=0.2448, simple_loss=0.298, pruned_loss=0.09577, over 928716.49 frames. ], batch size: 27, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:16:11,580 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 2.083e+02 2.562e+02 2.927e+02 7.910e+02, threshold=5.125e+02, percent-clipped=5.0 +2023-04-26 13:16:16,502 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-26 13:16:23,274 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:16:42,545 INFO [finetune.py:976] (3/7) Epoch 3, batch 800, loss[loss=0.2171, simple_loss=0.2838, pruned_loss=0.07521, over 4736.00 frames. ], tot_loss[loss=0.2452, simple_loss=0.2983, pruned_loss=0.09606, over 933927.58 frames. ], batch size: 59, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:16:52,248 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9402, 2.8521, 2.3207, 3.3054, 2.8780, 2.9008, 1.1588, 2.8146], + device='cuda:3'), covar=tensor([0.1982, 0.1636, 0.3040, 0.2901, 0.2531, 0.2174, 0.5750, 0.2875], + device='cuda:3'), in_proj_covar=tensor([0.0255, 0.0229, 0.0270, 0.0323, 0.0315, 0.0266, 0.0281, 0.0283], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:17:00,572 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12276.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:08,098 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3461, 1.3366, 1.2684, 1.9270, 2.0871, 1.7685, 1.7227, 1.5157], + device='cuda:3'), covar=tensor([0.2148, 0.3590, 0.3824, 0.2553, 0.2643, 0.3685, 0.3506, 0.3054], + device='cuda:3'), in_proj_covar=tensor([0.0321, 0.0338, 0.0348, 0.0314, 0.0349, 0.0364, 0.0320, 0.0353], + device='cuda:3'), out_proj_covar=tensor([6.9893e-05, 7.2633e-05, 7.5492e-05, 6.5796e-05, 7.4367e-05, 7.9568e-05, + 6.9660e-05, 7.6240e-05], device='cuda:3') +2023-04-26 13:17:11,621 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12291.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:14,601 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0841, 1.4303, 1.9085, 2.3799, 1.7701, 1.3862, 1.2011, 1.7120], + device='cuda:3'), covar=tensor([0.5456, 0.6579, 0.3244, 0.4875, 0.6407, 0.4853, 0.8337, 0.5876], + device='cuda:3'), in_proj_covar=tensor([0.0269, 0.0275, 0.0223, 0.0343, 0.0232, 0.0235, 0.0265, 0.0212], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:17:19,901 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12304.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:21,019 INFO [finetune.py:976] (3/7) Epoch 3, batch 850, loss[loss=0.241, simple_loss=0.2857, pruned_loss=0.09819, over 4742.00 frames. ], tot_loss[loss=0.2428, simple_loss=0.2954, pruned_loss=0.09509, over 937439.62 frames. ], batch size: 27, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:17:29,560 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.971e+02 2.371e+02 2.612e+02 4.896e+02, threshold=4.741e+02, percent-clipped=0.0 +2023-04-26 13:17:43,089 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12339.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:43,714 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12340.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:54,306 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12355.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:54,807 INFO [finetune.py:976] (3/7) Epoch 3, batch 900, loss[loss=0.2037, simple_loss=0.2662, pruned_loss=0.07063, over 4816.00 frames. ], tot_loss[loss=0.2407, simple_loss=0.293, pruned_loss=0.09422, over 941551.61 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:18:03,688 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-26 13:18:10,314 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-26 13:18:26,968 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12403.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:18:28,775 INFO [finetune.py:976] (3/7) Epoch 3, batch 950, loss[loss=0.225, simple_loss=0.2811, pruned_loss=0.08446, over 4827.00 frames. ], tot_loss[loss=0.2405, simple_loss=0.2921, pruned_loss=0.09444, over 942598.78 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:18:37,307 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.210e+02 1.930e+02 2.195e+02 2.798e+02 5.251e+02, threshold=4.389e+02, percent-clipped=2.0 +2023-04-26 13:18:50,729 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12432.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:19:13,294 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12446.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:19:20,771 INFO [finetune.py:976] (3/7) Epoch 3, batch 1000, loss[loss=0.2404, simple_loss=0.2957, pruned_loss=0.09258, over 4891.00 frames. ], tot_loss[loss=0.2434, simple_loss=0.2951, pruned_loss=0.0959, over 944003.84 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:19:28,837 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3894, 0.9944, 1.3770, 1.5796, 1.4663, 1.3658, 1.3983, 1.3481], + device='cuda:3'), covar=tensor([2.3098, 3.1333, 3.4163, 4.0650, 2.5616, 3.6691, 3.8395, 2.8288], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0502, 0.0597, 0.0601, 0.0481, 0.0517, 0.0530, 0.0540], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:19:30,633 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12472.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:19:43,111 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8599, 1.2863, 1.7593, 2.1559, 1.5913, 1.3166, 1.1606, 1.4756], + device='cuda:3'), covar=tensor([0.5597, 0.6750, 0.3083, 0.5535, 0.6533, 0.4933, 0.8877, 0.6362], + device='cuda:3'), in_proj_covar=tensor([0.0270, 0.0274, 0.0223, 0.0343, 0.0232, 0.0235, 0.0265, 0.0212], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:20:03,632 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12502.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:20:06,941 INFO [finetune.py:976] (3/7) Epoch 3, batch 1050, loss[loss=0.2947, simple_loss=0.3441, pruned_loss=0.1226, over 4889.00 frames. ], tot_loss[loss=0.2457, simple_loss=0.2987, pruned_loss=0.09637, over 948214.59 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:20:07,669 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12507.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:20:25,976 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.380e+02 2.009e+02 2.317e+02 2.704e+02 5.500e+02, threshold=4.634e+02, percent-clipped=2.0 +2023-04-26 13:20:26,052 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12520.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:00,883 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6543, 4.4642, 0.9789, 2.1554, 2.5724, 2.8294, 2.6692, 1.1169], + device='cuda:3'), covar=tensor([0.1460, 0.0940, 0.2216, 0.1526, 0.0998, 0.1230, 0.1350, 0.1958], + device='cuda:3'), in_proj_covar=tensor([0.0125, 0.0270, 0.0151, 0.0131, 0.0142, 0.0165, 0.0129, 0.0133], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:21:02,676 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12550.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:12,635 INFO [finetune.py:976] (3/7) Epoch 3, batch 1100, loss[loss=0.235, simple_loss=0.2989, pruned_loss=0.08557, over 4820.00 frames. ], tot_loss[loss=0.2468, simple_loss=0.3005, pruned_loss=0.09652, over 951188.41 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:21:25,844 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12576.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:29,518 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9045, 1.2202, 1.7792, 2.0736, 1.5570, 1.2629, 1.0256, 1.4846], + device='cuda:3'), covar=tensor([0.5739, 0.6949, 0.3089, 0.5393, 0.6641, 0.5229, 0.8614, 0.5859], + device='cuda:3'), in_proj_covar=tensor([0.0270, 0.0274, 0.0224, 0.0343, 0.0232, 0.0235, 0.0265, 0.0211], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:21:44,469 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12604.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:44,756 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 13:21:46,066 INFO [finetune.py:976] (3/7) Epoch 3, batch 1150, loss[loss=0.2639, simple_loss=0.3114, pruned_loss=0.1083, over 4883.00 frames. ], tot_loss[loss=0.2458, simple_loss=0.3002, pruned_loss=0.09573, over 951972.19 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:21:55,573 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.367e+02 1.845e+02 2.160e+02 2.613e+02 5.486e+02, threshold=4.320e+02, percent-clipped=1.0 +2023-04-26 13:21:58,102 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12624.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:22:00,011 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3606, 1.5741, 1.2579, 1.5109, 1.3115, 1.2007, 1.4417, 1.1533], + device='cuda:3'), covar=tensor([0.1385, 0.1157, 0.1092, 0.1093, 0.2738, 0.1359, 0.1491, 0.1980], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0339, 0.0248, 0.0311, 0.0317, 0.0290, 0.0280, 0.0302], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:22:19,230 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12640.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:22:34,211 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12652.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:22:36,635 INFO [finetune.py:976] (3/7) Epoch 3, batch 1200, loss[loss=0.2684, simple_loss=0.3144, pruned_loss=0.1112, over 4169.00 frames. ], tot_loss[loss=0.2443, simple_loss=0.2984, pruned_loss=0.09504, over 951435.12 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:22:43,031 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9259, 1.2756, 1.7843, 2.2591, 1.5394, 1.3271, 1.1062, 1.4259], + device='cuda:3'), covar=tensor([0.5955, 0.7044, 0.3288, 0.5451, 0.6920, 0.5223, 0.9278, 0.6423], + device='cuda:3'), in_proj_covar=tensor([0.0270, 0.0274, 0.0224, 0.0343, 0.0233, 0.0236, 0.0265, 0.0211], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:22:57,556 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12688.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:23:09,931 INFO [finetune.py:976] (3/7) Epoch 3, batch 1250, loss[loss=0.2108, simple_loss=0.2545, pruned_loss=0.08357, over 4860.00 frames. ], tot_loss[loss=0.241, simple_loss=0.2945, pruned_loss=0.09375, over 950629.96 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:23:19,406 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.393e+02 1.929e+02 2.277e+02 2.741e+02 5.638e+02, threshold=4.554e+02, percent-clipped=3.0 +2023-04-26 13:23:26,759 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12732.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:23:42,758 INFO [finetune.py:976] (3/7) Epoch 3, batch 1300, loss[loss=0.2211, simple_loss=0.2887, pruned_loss=0.07671, over 4822.00 frames. ], tot_loss[loss=0.2383, simple_loss=0.2918, pruned_loss=0.0924, over 953723.42 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:23:47,613 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:24:09,812 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12780.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:24:30,038 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-04-26 13:24:35,997 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12802.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:24:39,387 INFO [finetune.py:976] (3/7) Epoch 3, batch 1350, loss[loss=0.2677, simple_loss=0.3061, pruned_loss=0.1146, over 4901.00 frames. ], tot_loss[loss=0.2385, simple_loss=0.2916, pruned_loss=0.09265, over 954392.39 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:24:48,912 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.232e+02 1.880e+02 2.141e+02 2.633e+02 4.297e+02, threshold=4.283e+02, percent-clipped=1.0 +2023-04-26 13:24:51,951 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12824.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:25:12,231 INFO [finetune.py:976] (3/7) Epoch 3, batch 1400, loss[loss=0.3119, simple_loss=0.3567, pruned_loss=0.1335, over 4817.00 frames. ], tot_loss[loss=0.2436, simple_loss=0.297, pruned_loss=0.09506, over 956832.88 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:25:20,043 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8271, 1.1475, 1.6746, 1.9859, 1.6318, 1.2641, 1.0987, 1.4079], + device='cuda:3'), covar=tensor([0.5320, 0.6479, 0.2882, 0.5284, 0.5608, 0.4285, 0.7388, 0.4962], + device='cuda:3'), in_proj_covar=tensor([0.0271, 0.0275, 0.0224, 0.0345, 0.0233, 0.0236, 0.0266, 0.0212], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:25:56,156 INFO [finetune.py:976] (3/7) Epoch 3, batch 1450, loss[loss=0.247, simple_loss=0.3163, pruned_loss=0.08882, over 4815.00 frames. ], tot_loss[loss=0.2435, simple_loss=0.2976, pruned_loss=0.09468, over 955533.59 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:26:17,005 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.305e+02 1.925e+02 2.434e+02 2.951e+02 7.906e+02, threshold=4.868e+02, percent-clipped=4.0 +2023-04-26 13:26:58,305 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 13:27:02,919 INFO [finetune.py:976] (3/7) Epoch 3, batch 1500, loss[loss=0.2045, simple_loss=0.2388, pruned_loss=0.08512, over 4405.00 frames. ], tot_loss[loss=0.246, simple_loss=0.3001, pruned_loss=0.09599, over 954112.80 frames. ], batch size: 19, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:27:41,573 INFO [finetune.py:976] (3/7) Epoch 3, batch 1550, loss[loss=0.2191, simple_loss=0.2664, pruned_loss=0.08586, over 4734.00 frames. ], tot_loss[loss=0.2438, simple_loss=0.2985, pruned_loss=0.09451, over 953275.17 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:28:02,727 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.270e+02 2.021e+02 2.296e+02 2.726e+02 4.661e+02, threshold=4.592e+02, percent-clipped=0.0 +2023-04-26 13:28:21,022 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 13:28:47,302 INFO [finetune.py:976] (3/7) Epoch 3, batch 1600, loss[loss=0.1913, simple_loss=0.2573, pruned_loss=0.06261, over 4890.00 frames. ], tot_loss[loss=0.2409, simple_loss=0.2957, pruned_loss=0.09308, over 953714.88 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:29:23,844 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13102.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:29:26,216 INFO [finetune.py:976] (3/7) Epoch 3, batch 1650, loss[loss=0.2041, simple_loss=0.2589, pruned_loss=0.07464, over 4935.00 frames. ], tot_loss[loss=0.2379, simple_loss=0.2923, pruned_loss=0.0918, over 954862.96 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:29:34,728 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=13119.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:29:35,235 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.899e+02 2.180e+02 2.609e+02 5.027e+02, threshold=4.359e+02, percent-clipped=1.0 +2023-04-26 13:29:55,914 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13150.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:29:59,486 INFO [finetune.py:976] (3/7) Epoch 3, batch 1700, loss[loss=0.2588, simple_loss=0.3015, pruned_loss=0.1081, over 4892.00 frames. ], tot_loss[loss=0.2377, simple_loss=0.2915, pruned_loss=0.09193, over 955955.53 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:30:19,359 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 13:30:27,395 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9635, 3.0021, 2.0086, 1.9210, 1.3424, 1.4899, 2.1809, 1.4695], + device='cuda:3'), covar=tensor([0.2225, 0.1921, 0.2214, 0.2706, 0.3512, 0.2563, 0.1703, 0.2785], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0224, 0.0191, 0.0216, 0.0229, 0.0194, 0.0186, 0.0206], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:30:31,773 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4821, 1.1983, 0.5878, 1.1387, 1.4982, 1.3467, 1.2224, 1.3002], + device='cuda:3'), covar=tensor([0.0606, 0.0483, 0.0525, 0.0621, 0.0338, 0.0576, 0.0581, 0.0718], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 13:30:33,506 INFO [finetune.py:976] (3/7) Epoch 3, batch 1750, loss[loss=0.2631, simple_loss=0.3157, pruned_loss=0.1053, over 4831.00 frames. ], tot_loss[loss=0.2392, simple_loss=0.2931, pruned_loss=0.09264, over 955267.37 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:30:43,120 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.834e+02 2.276e+02 2.863e+02 4.766e+02, threshold=4.553e+02, percent-clipped=3.0 +2023-04-26 13:30:47,831 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7081, 1.8828, 1.6183, 1.7209, 1.6343, 1.9471, 1.8295, 1.6111], + device='cuda:3'), covar=tensor([1.2955, 2.3093, 1.9595, 1.6697, 1.8517, 3.0358, 2.2897, 2.1369], + device='cuda:3'), in_proj_covar=tensor([0.0299, 0.0399, 0.0315, 0.0322, 0.0349, 0.0401, 0.0384, 0.0344], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:30:56,572 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9573, 1.2473, 3.3121, 3.0435, 2.9592, 3.2294, 3.2692, 2.9705], + device='cuda:3'), covar=tensor([0.7100, 0.5448, 0.1489, 0.2203, 0.1505, 0.1993, 0.1371, 0.1659], + device='cuda:3'), in_proj_covar=tensor([0.0328, 0.0312, 0.0440, 0.0445, 0.0375, 0.0427, 0.0338, 0.0394], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 13:31:02,352 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 13:31:07,408 INFO [finetune.py:976] (3/7) Epoch 3, batch 1800, loss[loss=0.238, simple_loss=0.2961, pruned_loss=0.08996, over 4848.00 frames. ], tot_loss[loss=0.2401, simple_loss=0.2952, pruned_loss=0.0925, over 954497.23 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:31:23,408 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-26 13:31:44,870 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.05 vs. limit=5.0 +2023-04-26 13:31:47,123 INFO [finetune.py:976] (3/7) Epoch 3, batch 1850, loss[loss=0.2627, simple_loss=0.3181, pruned_loss=0.1036, over 4819.00 frames. ], tot_loss[loss=0.2424, simple_loss=0.2974, pruned_loss=0.09372, over 954819.61 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:31:56,813 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.349e+02 1.904e+02 2.289e+02 2.801e+02 9.204e+02, threshold=4.579e+02, percent-clipped=2.0 +2023-04-26 13:32:19,081 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1031, 1.2877, 1.1889, 1.5914, 1.3232, 1.6295, 1.2353, 2.9667], + device='cuda:3'), covar=tensor([0.0874, 0.1108, 0.1128, 0.1537, 0.0936, 0.0748, 0.1059, 0.0277], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 13:32:28,657 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 13:32:30,328 INFO [finetune.py:976] (3/7) Epoch 3, batch 1900, loss[loss=0.2585, simple_loss=0.3156, pruned_loss=0.1007, over 4804.00 frames. ], tot_loss[loss=0.2423, simple_loss=0.2978, pruned_loss=0.09338, over 952897.91 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:33:26,631 INFO [finetune.py:976] (3/7) Epoch 3, batch 1950, loss[loss=0.1921, simple_loss=0.2489, pruned_loss=0.06769, over 4792.00 frames. ], tot_loss[loss=0.2407, simple_loss=0.2961, pruned_loss=0.09261, over 954257.31 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:33:29,863 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9081, 2.2399, 0.9453, 1.1792, 1.6589, 1.1837, 2.4449, 1.3636], + device='cuda:3'), covar=tensor([0.0696, 0.0589, 0.0694, 0.1291, 0.0435, 0.1020, 0.0311, 0.0729], + device='cuda:3'), in_proj_covar=tensor([0.0056, 0.0073, 0.0054, 0.0050, 0.0055, 0.0056, 0.0085, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 13:33:39,985 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13419.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:33:41,756 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.766e+02 2.138e+02 2.602e+02 4.545e+02, threshold=4.277e+02, percent-clipped=0.0 +2023-04-26 13:34:17,442 INFO [finetune.py:976] (3/7) Epoch 3, batch 2000, loss[loss=0.1881, simple_loss=0.2478, pruned_loss=0.0642, over 4735.00 frames. ], tot_loss[loss=0.2379, simple_loss=0.2927, pruned_loss=0.09152, over 953576.48 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:34:24,601 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:34:31,348 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5992, 1.8179, 1.7961, 2.2393, 2.5817, 2.3069, 1.9962, 1.8320], + device='cuda:3'), covar=tensor([0.1746, 0.2402, 0.2583, 0.2432, 0.1437, 0.2453, 0.3256, 0.2444], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0336, 0.0347, 0.0312, 0.0348, 0.0358, 0.0319, 0.0350], + device='cuda:3'), out_proj_covar=tensor([6.9305e-05, 7.2224e-05, 7.5279e-05, 6.5346e-05, 7.4108e-05, 7.8313e-05, + 6.9441e-05, 7.5436e-05], device='cuda:3') +2023-04-26 13:34:51,090 INFO [finetune.py:976] (3/7) Epoch 3, batch 2050, loss[loss=0.2029, simple_loss=0.269, pruned_loss=0.06844, over 4907.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.2888, pruned_loss=0.08987, over 956042.84 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:35:01,249 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.294e+02 1.882e+02 2.225e+02 2.584e+02 5.744e+02, threshold=4.451e+02, percent-clipped=3.0 +2023-04-26 13:35:24,229 INFO [finetune.py:976] (3/7) Epoch 3, batch 2100, loss[loss=0.2924, simple_loss=0.3395, pruned_loss=0.1226, over 4209.00 frames. ], tot_loss[loss=0.2348, simple_loss=0.289, pruned_loss=0.09027, over 955455.81 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:35:26,685 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9435, 2.8561, 2.2702, 3.3112, 2.9283, 2.8928, 1.2494, 2.8507], + device='cuda:3'), covar=tensor([0.2070, 0.1601, 0.3161, 0.2671, 0.3488, 0.2281, 0.5435, 0.2725], + device='cuda:3'), in_proj_covar=tensor([0.0257, 0.0229, 0.0271, 0.0322, 0.0316, 0.0268, 0.0282, 0.0284], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:35:30,125 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-26 13:35:43,771 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8258, 2.5465, 1.7029, 1.6511, 1.3103, 1.4264, 1.7212, 1.3019], + device='cuda:3'), covar=tensor([0.2436, 0.1904, 0.2309, 0.2742, 0.3586, 0.2727, 0.1899, 0.2835], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0226, 0.0191, 0.0216, 0.0230, 0.0195, 0.0186, 0.0207], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:35:47,655 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6774, 1.7686, 1.1642, 1.4109, 2.0159, 1.5619, 1.4681, 1.5517], + device='cuda:3'), covar=tensor([0.0484, 0.0365, 0.0402, 0.0522, 0.0276, 0.0507, 0.0467, 0.0578], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 13:35:50,112 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2847, 2.5770, 1.1970, 1.3390, 2.1862, 1.2676, 3.3893, 1.6824], + device='cuda:3'), covar=tensor([0.0654, 0.0606, 0.0806, 0.1293, 0.0471, 0.1003, 0.0259, 0.0654], + device='cuda:3'), in_proj_covar=tensor([0.0056, 0.0073, 0.0054, 0.0050, 0.0055, 0.0056, 0.0085, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 13:35:57,672 INFO [finetune.py:976] (3/7) Epoch 3, batch 2150, loss[loss=0.2464, simple_loss=0.304, pruned_loss=0.09438, over 4860.00 frames. ], tot_loss[loss=0.2393, simple_loss=0.2938, pruned_loss=0.09241, over 955355.52 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:36:07,903 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.340e+02 1.924e+02 2.294e+02 2.856e+02 4.572e+02, threshold=4.587e+02, percent-clipped=1.0 +2023-04-26 13:36:14,709 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4832, 3.7362, 0.7803, 2.0232, 2.0311, 2.4797, 2.2835, 1.1132], + device='cuda:3'), covar=tensor([0.1462, 0.0946, 0.2341, 0.1421, 0.1146, 0.1240, 0.1431, 0.2241], + device='cuda:3'), in_proj_covar=tensor([0.0125, 0.0272, 0.0152, 0.0132, 0.0143, 0.0166, 0.0129, 0.0133], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:36:30,854 INFO [finetune.py:976] (3/7) Epoch 3, batch 2200, loss[loss=0.2128, simple_loss=0.2734, pruned_loss=0.07606, over 4761.00 frames. ], tot_loss[loss=0.2403, simple_loss=0.2952, pruned_loss=0.09267, over 955960.93 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:37:21,694 INFO [finetune.py:976] (3/7) Epoch 3, batch 2250, loss[loss=0.204, simple_loss=0.276, pruned_loss=0.06602, over 4856.00 frames. ], tot_loss[loss=0.2409, simple_loss=0.2958, pruned_loss=0.09294, over 956135.43 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:37:31,818 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.241e+02 1.916e+02 2.386e+02 2.867e+02 7.645e+02, threshold=4.772e+02, percent-clipped=4.0 +2023-04-26 13:37:43,772 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=13737.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:38:02,148 INFO [finetune.py:976] (3/7) Epoch 3, batch 2300, loss[loss=0.1958, simple_loss=0.2617, pruned_loss=0.06493, over 4919.00 frames. ], tot_loss[loss=0.2404, simple_loss=0.2962, pruned_loss=0.09234, over 957028.14 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:38:57,791 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=13798.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:39:09,160 INFO [finetune.py:976] (3/7) Epoch 3, batch 2350, loss[loss=0.2734, simple_loss=0.3079, pruned_loss=0.1195, over 4799.00 frames. ], tot_loss[loss=0.2373, simple_loss=0.2926, pruned_loss=0.09102, over 955056.67 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:39:15,383 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1724, 1.3405, 1.4140, 1.4688, 1.4768, 1.1813, 0.8253, 1.3281], + device='cuda:3'), covar=tensor([0.0984, 0.1111, 0.0793, 0.0691, 0.0681, 0.1046, 0.1144, 0.0687], + device='cuda:3'), in_proj_covar=tensor([0.0210, 0.0209, 0.0189, 0.0184, 0.0182, 0.0200, 0.0175, 0.0196], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:39:37,981 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.317e+02 2.025e+02 2.480e+02 2.975e+02 5.089e+02, threshold=4.960e+02, percent-clipped=2.0 +2023-04-26 13:39:43,023 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5027, 0.9126, 0.4052, 1.1621, 1.0754, 1.3861, 1.2556, 1.2451], + device='cuda:3'), covar=tensor([0.0608, 0.0498, 0.0514, 0.0636, 0.0371, 0.0635, 0.0576, 0.0708], + device='cuda:3'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 13:40:20,272 INFO [finetune.py:976] (3/7) Epoch 3, batch 2400, loss[loss=0.2168, simple_loss=0.2706, pruned_loss=0.08148, over 4877.00 frames. ], tot_loss[loss=0.2356, simple_loss=0.2902, pruned_loss=0.09052, over 953389.27 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:40:22,709 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9670, 1.3510, 1.7555, 1.9237, 1.5936, 1.3321, 0.8522, 1.3996], + device='cuda:3'), covar=tensor([0.5388, 0.6203, 0.2636, 0.4455, 0.5585, 0.4504, 0.7782, 0.5223], + device='cuda:3'), in_proj_covar=tensor([0.0269, 0.0271, 0.0223, 0.0341, 0.0230, 0.0234, 0.0262, 0.0208], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:40:57,684 INFO [finetune.py:976] (3/7) Epoch 3, batch 2450, loss[loss=0.2212, simple_loss=0.2781, pruned_loss=0.0821, over 4819.00 frames. ], tot_loss[loss=0.2335, simple_loss=0.288, pruned_loss=0.0895, over 956228.76 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:41:09,356 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.373e+02 2.035e+02 2.305e+02 2.759e+02 5.668e+02, threshold=4.609e+02, percent-clipped=2.0 +2023-04-26 13:41:11,160 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5395, 1.9557, 1.5493, 1.8173, 1.4900, 1.6610, 1.6776, 1.2738], + device='cuda:3'), covar=tensor([0.2204, 0.1403, 0.1270, 0.1585, 0.3287, 0.1532, 0.1858, 0.2901], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0343, 0.0252, 0.0315, 0.0324, 0.0294, 0.0283, 0.0306], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:41:26,930 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6232, 2.9913, 0.9126, 1.4959, 2.3408, 1.4056, 4.4924, 2.3542], + device='cuda:3'), covar=tensor([0.0575, 0.1083, 0.1012, 0.1289, 0.0565, 0.0995, 0.0212, 0.0587], + device='cuda:3'), in_proj_covar=tensor([0.0056, 0.0074, 0.0054, 0.0051, 0.0056, 0.0056, 0.0086, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 13:41:31,084 INFO [finetune.py:976] (3/7) Epoch 3, batch 2500, loss[loss=0.2089, simple_loss=0.2821, pruned_loss=0.06785, over 4895.00 frames. ], tot_loss[loss=0.2359, simple_loss=0.2903, pruned_loss=0.09078, over 955084.13 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:41:33,063 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1226, 2.1520, 2.0029, 1.8084, 2.2795, 1.7362, 2.9143, 1.7581], + device='cuda:3'), covar=tensor([0.4020, 0.1737, 0.4352, 0.3074, 0.1911, 0.2992, 0.1166, 0.3994], + device='cuda:3'), in_proj_covar=tensor([0.0350, 0.0352, 0.0433, 0.0370, 0.0402, 0.0379, 0.0396, 0.0414], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:42:05,883 INFO [finetune.py:976] (3/7) Epoch 3, batch 2550, loss[loss=0.2674, simple_loss=0.3131, pruned_loss=0.1108, over 4897.00 frames. ], tot_loss[loss=0.2408, simple_loss=0.2954, pruned_loss=0.09306, over 954483.34 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:42:26,619 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.33 vs. limit=5.0 +2023-04-26 13:42:28,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.240e+02 1.992e+02 2.434e+02 2.841e+02 4.921e+02, threshold=4.868e+02, percent-clipped=1.0 +2023-04-26 13:43:13,249 INFO [finetune.py:976] (3/7) Epoch 3, batch 2600, loss[loss=0.2219, simple_loss=0.2846, pruned_loss=0.07965, over 4798.00 frames. ], tot_loss[loss=0.2418, simple_loss=0.2966, pruned_loss=0.09345, over 954590.24 frames. ], batch size: 41, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:43:37,606 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14072.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:43:57,176 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5539, 1.3689, 1.7491, 1.7998, 1.7911, 1.5251, 1.6419, 1.6137], + device='cuda:3'), covar=tensor([2.3958, 3.1498, 3.5912, 3.9711, 2.3760, 4.2700, 4.2273, 3.1295], + device='cuda:3'), in_proj_covar=tensor([0.0446, 0.0497, 0.0588, 0.0598, 0.0476, 0.0511, 0.0522, 0.0533], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:44:06,892 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14093.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 13:44:08,663 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.3527, 4.2594, 3.0869, 4.9725, 4.2974, 4.3665, 2.0364, 4.2446], + device='cuda:3'), covar=tensor([0.1650, 0.1040, 0.3247, 0.0899, 0.2239, 0.1636, 0.5153, 0.2060], + device='cuda:3'), in_proj_covar=tensor([0.0257, 0.0228, 0.0270, 0.0320, 0.0315, 0.0266, 0.0283, 0.0282], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:44:19,004 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14103.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:44:25,133 INFO [finetune.py:976] (3/7) Epoch 3, batch 2650, loss[loss=0.2314, simple_loss=0.2959, pruned_loss=0.08344, over 4819.00 frames. ], tot_loss[loss=0.2409, simple_loss=0.2963, pruned_loss=0.09274, over 955268.99 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:44:29,521 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9328, 1.0274, 5.0722, 4.6905, 4.4452, 4.7463, 4.4046, 4.4764], + device='cuda:3'), covar=tensor([0.7124, 0.6890, 0.1158, 0.1995, 0.1032, 0.1276, 0.1620, 0.1658], + device='cuda:3'), in_proj_covar=tensor([0.0327, 0.0311, 0.0437, 0.0443, 0.0372, 0.0423, 0.0335, 0.0391], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:44:40,404 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14120.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:44:47,127 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.256e+02 1.867e+02 2.323e+02 2.768e+02 1.192e+03, threshold=4.647e+02, percent-clipped=2.0 +2023-04-26 13:44:51,507 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 13:45:00,938 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14133.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:45:11,047 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-26 13:45:21,185 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-04-26 13:45:30,815 INFO [finetune.py:976] (3/7) Epoch 3, batch 2700, loss[loss=0.2106, simple_loss=0.2584, pruned_loss=0.08144, over 4795.00 frames. ], tot_loss[loss=0.2385, simple_loss=0.2941, pruned_loss=0.09141, over 955888.31 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:45:41,238 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14164.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:46:03,629 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14181.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:46:07,893 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-26 13:46:30,719 INFO [finetune.py:976] (3/7) Epoch 3, batch 2750, loss[loss=0.2444, simple_loss=0.2992, pruned_loss=0.09477, over 4874.00 frames. ], tot_loss[loss=0.2359, simple_loss=0.2908, pruned_loss=0.09044, over 955617.91 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:46:40,381 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.352e+02 1.823e+02 2.218e+02 2.622e+02 5.684e+02, threshold=4.435e+02, percent-clipped=2.0 +2023-04-26 13:46:46,386 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0407, 2.4932, 2.4375, 2.8577, 2.6385, 2.8681, 2.3747, 4.9069], + device='cuda:3'), covar=tensor([0.0511, 0.0593, 0.0604, 0.0943, 0.0506, 0.0398, 0.0588, 0.0124], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 13:47:03,042 INFO [finetune.py:976] (3/7) Epoch 3, batch 2800, loss[loss=0.1837, simple_loss=0.2448, pruned_loss=0.06129, over 4828.00 frames. ], tot_loss[loss=0.2322, simple_loss=0.2866, pruned_loss=0.08888, over 955921.64 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:47:35,420 INFO [finetune.py:976] (3/7) Epoch 3, batch 2850, loss[loss=0.1716, simple_loss=0.2266, pruned_loss=0.05831, over 4798.00 frames. ], tot_loss[loss=0.2321, simple_loss=0.286, pruned_loss=0.08911, over 955607.33 frames. ], batch size: 29, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:47:45,432 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.273e+02 1.919e+02 2.285e+02 2.703e+02 8.098e+02, threshold=4.570e+02, percent-clipped=1.0 +2023-04-26 13:48:06,154 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6477, 2.3783, 1.6603, 1.4381, 1.2419, 1.2729, 1.7025, 1.2525], + device='cuda:3'), covar=tensor([0.2231, 0.1865, 0.2246, 0.2809, 0.3409, 0.2592, 0.1695, 0.2669], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0223, 0.0188, 0.0213, 0.0227, 0.0193, 0.0183, 0.0203], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:48:08,421 INFO [finetune.py:976] (3/7) Epoch 3, batch 2900, loss[loss=0.2254, simple_loss=0.289, pruned_loss=0.08087, over 4930.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.2892, pruned_loss=0.09107, over 955746.67 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:48:33,689 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14393.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:48:41,107 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6182, 1.2040, 4.1719, 3.8617, 3.6741, 3.8863, 3.7290, 3.7354], + device='cuda:3'), covar=tensor([0.6601, 0.5751, 0.0815, 0.1407, 0.0983, 0.1569, 0.2131, 0.1309], + device='cuda:3'), in_proj_covar=tensor([0.0325, 0.0310, 0.0436, 0.0441, 0.0370, 0.0422, 0.0333, 0.0390], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:48:41,652 INFO [finetune.py:976] (3/7) Epoch 3, batch 2950, loss[loss=0.2471, simple_loss=0.3, pruned_loss=0.09712, over 4784.00 frames. ], tot_loss[loss=0.238, simple_loss=0.2922, pruned_loss=0.09188, over 955155.76 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:48:45,428 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8327, 4.2968, 0.8844, 2.0327, 2.2579, 2.7904, 2.5159, 0.9579], + device='cuda:3'), covar=tensor([0.1319, 0.0917, 0.2323, 0.1496, 0.1118, 0.1147, 0.1407, 0.2296], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0266, 0.0149, 0.0130, 0.0141, 0.0164, 0.0126, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:48:46,068 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14413.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:48:51,910 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.499e+02 2.004e+02 2.390e+02 2.807e+02 6.198e+02, threshold=4.780e+02, percent-clipped=3.0 +2023-04-26 13:48:55,646 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14428.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:49:04,517 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14441.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:49:13,398 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0311, 2.0116, 2.1866, 2.4084, 2.3981, 1.8831, 1.6523, 2.0852], + device='cuda:3'), covar=tensor([0.1011, 0.1000, 0.0634, 0.0694, 0.0617, 0.1146, 0.1035, 0.0644], + device='cuda:3'), in_proj_covar=tensor([0.0208, 0.0209, 0.0188, 0.0182, 0.0182, 0.0199, 0.0174, 0.0194], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:49:14,497 INFO [finetune.py:976] (3/7) Epoch 3, batch 3000, loss[loss=0.2848, simple_loss=0.3259, pruned_loss=0.1219, over 4214.00 frames. ], tot_loss[loss=0.239, simple_loss=0.2937, pruned_loss=0.09214, over 953558.40 frames. ], batch size: 66, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:49:14,497 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 13:49:18,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0741, 1.5789, 1.8603, 2.0100, 1.7454, 1.4939, 0.9781, 1.5903], + device='cuda:3'), covar=tensor([0.4683, 0.5921, 0.2575, 0.4282, 0.5286, 0.4120, 0.7539, 0.4911], + device='cuda:3'), in_proj_covar=tensor([0.0270, 0.0271, 0.0223, 0.0341, 0.0230, 0.0234, 0.0261, 0.0208], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:49:22,509 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4051, 1.2281, 1.6220, 1.5373, 1.3008, 1.1065, 1.4105, 1.0547], + device='cuda:3'), covar=tensor([0.0922, 0.0774, 0.0716, 0.0834, 0.1062, 0.1605, 0.0705, 0.0996], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0079, 0.0076, 0.0071, 0.0084, 0.0097, 0.0088, 0.0080], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 13:49:25,022 INFO [finetune.py:1010] (3/7) Epoch 3, validation: loss=0.1699, simple_loss=0.2433, pruned_loss=0.04821, over 2265189.00 frames. +2023-04-26 13:49:25,022 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 13:49:26,924 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14459.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:49:36,018 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14474.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:49:37,157 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14476.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:49:37,873 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-26 13:49:44,320 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14487.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:50:13,197 INFO [finetune.py:976] (3/7) Epoch 3, batch 3050, loss[loss=0.2485, simple_loss=0.3106, pruned_loss=0.09315, over 4809.00 frames. ], tot_loss[loss=0.2382, simple_loss=0.2938, pruned_loss=0.0913, over 954533.18 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:50:24,171 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.447e+02 2.059e+02 2.417e+02 2.833e+02 5.136e+02, threshold=4.834e+02, percent-clipped=1.0 +2023-04-26 13:50:24,941 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2557, 1.4254, 1.3825, 1.5419, 1.4428, 1.5989, 1.5181, 1.4929], + device='cuda:3'), covar=tensor([1.3094, 2.1939, 1.8001, 1.5848, 1.8352, 3.0536, 2.2098, 1.9638], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0400, 0.0316, 0.0322, 0.0349, 0.0403, 0.0384, 0.0343], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:50:43,060 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4824, 3.0645, 0.8763, 1.7152, 1.8425, 2.0524, 1.9551, 0.9509], + device='cuda:3'), covar=tensor([0.1274, 0.0954, 0.1975, 0.1338, 0.1024, 0.1159, 0.1424, 0.1820], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0269, 0.0150, 0.0131, 0.0141, 0.0165, 0.0127, 0.0131], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 13:50:52,337 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14548.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:51:03,206 INFO [finetune.py:976] (3/7) Epoch 3, batch 3100, loss[loss=0.2166, simple_loss=0.2519, pruned_loss=0.09064, over 4116.00 frames. ], tot_loss[loss=0.2365, simple_loss=0.292, pruned_loss=0.09051, over 954305.41 frames. ], batch size: 17, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:51:50,135 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9004, 2.4594, 2.0557, 2.2076, 1.7348, 1.9799, 2.2390, 1.6567], + device='cuda:3'), covar=tensor([0.2505, 0.1469, 0.1140, 0.1510, 0.3325, 0.1545, 0.2074, 0.3076], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0337, 0.0246, 0.0308, 0.0319, 0.0289, 0.0279, 0.0300], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:52:09,680 INFO [finetune.py:976] (3/7) Epoch 3, batch 3150, loss[loss=0.2368, simple_loss=0.2595, pruned_loss=0.107, over 4454.00 frames. ], tot_loss[loss=0.2345, simple_loss=0.2893, pruned_loss=0.08986, over 955303.62 frames. ], batch size: 19, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:52:30,357 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.439e+02 2.002e+02 2.381e+02 3.061e+02 6.553e+02, threshold=4.761e+02, percent-clipped=3.0 +2023-04-26 13:52:36,685 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1514, 1.4263, 1.3202, 1.8136, 1.5625, 1.6119, 1.4222, 2.5012], + device='cuda:3'), covar=tensor([0.0628, 0.0778, 0.0824, 0.1212, 0.0633, 0.0463, 0.0760, 0.0278], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0041, 0.0041, 0.0041, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 13:52:36,739 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7097, 0.9611, 1.2410, 1.4353, 1.4337, 1.6518, 1.2712, 1.3137], + device='cuda:3'), covar=tensor([1.1988, 1.9610, 1.6711, 1.4331, 1.6254, 2.6073, 1.9661, 1.7177], + device='cuda:3'), in_proj_covar=tensor([0.0302, 0.0399, 0.0316, 0.0322, 0.0349, 0.0402, 0.0384, 0.0342], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:52:51,310 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2011, 1.4122, 1.3752, 1.5272, 1.4771, 1.6618, 1.4823, 1.4488], + device='cuda:3'), covar=tensor([1.2619, 2.1554, 1.6869, 1.5349, 1.8298, 2.8090, 2.1386, 1.7424], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0401, 0.0317, 0.0323, 0.0350, 0.0404, 0.0385, 0.0343], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:52:51,772 INFO [finetune.py:976] (3/7) Epoch 3, batch 3200, loss[loss=0.207, simple_loss=0.2691, pruned_loss=0.07241, over 4846.00 frames. ], tot_loss[loss=0.23, simple_loss=0.2844, pruned_loss=0.08782, over 955616.10 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:53:25,510 INFO [finetune.py:976] (3/7) Epoch 3, batch 3250, loss[loss=0.2289, simple_loss=0.2928, pruned_loss=0.08251, over 4917.00 frames. ], tot_loss[loss=0.2305, simple_loss=0.2846, pruned_loss=0.08824, over 954005.61 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:53:37,684 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.942e+02 2.271e+02 2.763e+02 5.504e+02, threshold=4.542e+02, percent-clipped=2.0 +2023-04-26 13:53:41,981 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14728.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:53:49,334 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14740.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:53:59,447 INFO [finetune.py:976] (3/7) Epoch 3, batch 3300, loss[loss=0.1879, simple_loss=0.234, pruned_loss=0.07091, over 4079.00 frames. ], tot_loss[loss=0.2348, simple_loss=0.289, pruned_loss=0.09033, over 950888.55 frames. ], batch size: 17, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:54:01,388 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14759.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:54:09,427 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14769.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:54:14,673 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14776.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:54:14,700 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14776.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:54:30,604 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14801.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:54:33,538 INFO [finetune.py:976] (3/7) Epoch 3, batch 3350, loss[loss=0.2683, simple_loss=0.3089, pruned_loss=0.1138, over 4219.00 frames. ], tot_loss[loss=0.2368, simple_loss=0.2911, pruned_loss=0.0912, over 947746.36 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:54:34,179 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14807.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:54:39,094 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2962, 1.5389, 2.0444, 2.6523, 1.8948, 1.5712, 1.3308, 1.8426], + device='cuda:3'), covar=tensor([0.4792, 0.5763, 0.2715, 0.4813, 0.5925, 0.4233, 0.7519, 0.5270], + device='cuda:3'), in_proj_covar=tensor([0.0272, 0.0272, 0.0224, 0.0343, 0.0231, 0.0235, 0.0262, 0.0209], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:54:44,575 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.421e+02 1.946e+02 2.255e+02 2.721e+02 6.180e+02, threshold=4.510e+02, percent-clipped=3.0 +2023-04-26 13:54:46,354 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14824.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:54:49,846 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1359, 0.7373, 0.9259, 0.7366, 1.2630, 0.9619, 0.7999, 1.0141], + device='cuda:3'), covar=tensor([0.1826, 0.1911, 0.2070, 0.1721, 0.1052, 0.1567, 0.2037, 0.2113], + device='cuda:3'), in_proj_covar=tensor([0.0322, 0.0340, 0.0351, 0.0314, 0.0352, 0.0361, 0.0321, 0.0354], + device='cuda:3'), out_proj_covar=tensor([7.0024e-05, 7.3036e-05, 7.6115e-05, 6.5786e-05, 7.4921e-05, 7.8865e-05, + 6.9857e-05, 7.6359e-05], device='cuda:3') +2023-04-26 13:54:58,942 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14843.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:55:13,044 INFO [finetune.py:976] (3/7) Epoch 3, batch 3400, loss[loss=0.2915, simple_loss=0.3394, pruned_loss=0.1218, over 4921.00 frames. ], tot_loss[loss=0.2395, simple_loss=0.2938, pruned_loss=0.09255, over 948400.66 frames. ], batch size: 42, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:56:04,612 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-04-26 13:56:08,499 INFO [finetune.py:976] (3/7) Epoch 3, batch 3450, loss[loss=0.2085, simple_loss=0.2568, pruned_loss=0.08013, over 4794.00 frames. ], tot_loss[loss=0.2372, simple_loss=0.2916, pruned_loss=0.09142, over 949475.80 frames. ], batch size: 29, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:56:18,850 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.286e+02 1.992e+02 2.282e+02 2.746e+02 5.093e+02, threshold=4.564e+02, percent-clipped=2.0 +2023-04-26 13:56:27,252 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-26 13:56:42,395 INFO [finetune.py:976] (3/7) Epoch 3, batch 3500, loss[loss=0.2274, simple_loss=0.2835, pruned_loss=0.0856, over 4874.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.2894, pruned_loss=0.09015, over 952280.36 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:57:38,372 INFO [finetune.py:976] (3/7) Epoch 3, batch 3550, loss[loss=0.2212, simple_loss=0.2501, pruned_loss=0.0962, over 4068.00 frames. ], tot_loss[loss=0.2308, simple_loss=0.2854, pruned_loss=0.08811, over 953492.39 frames. ], batch size: 17, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:57:54,037 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.782e+02 2.249e+02 2.630e+02 4.772e+02, threshold=4.498e+02, percent-clipped=1.0 +2023-04-26 13:58:13,552 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-04-26 13:58:28,222 INFO [finetune.py:976] (3/7) Epoch 3, batch 3600, loss[loss=0.1916, simple_loss=0.2674, pruned_loss=0.05792, over 4899.00 frames. ], tot_loss[loss=0.2274, simple_loss=0.2825, pruned_loss=0.08616, over 955495.77 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:58:36,830 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15069.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:58:56,109 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15096.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:59:02,216 INFO [finetune.py:976] (3/7) Epoch 3, batch 3650, loss[loss=0.2556, simple_loss=0.3208, pruned_loss=0.09518, over 4907.00 frames. ], tot_loss[loss=0.2296, simple_loss=0.2848, pruned_loss=0.08715, over 954298.52 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:59:09,045 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15117.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:59:12,566 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.417e+02 2.035e+02 2.405e+02 3.001e+02 5.155e+02, threshold=4.810e+02, percent-clipped=2.0 +2023-04-26 13:59:28,017 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15143.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:59:36,236 INFO [finetune.py:976] (3/7) Epoch 3, batch 3700, loss[loss=0.24, simple_loss=0.2954, pruned_loss=0.09232, over 4925.00 frames. ], tot_loss[loss=0.2327, simple_loss=0.2885, pruned_loss=0.08845, over 954861.18 frames. ], batch size: 42, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:59:42,989 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7734, 2.7124, 2.0434, 2.2124, 1.8518, 2.0389, 2.3649, 1.6669], + device='cuda:3'), covar=tensor([0.3286, 0.1538, 0.1275, 0.2186, 0.3892, 0.1886, 0.2964, 0.3991], + device='cuda:3'), in_proj_covar=tensor([0.0319, 0.0337, 0.0247, 0.0310, 0.0321, 0.0291, 0.0280, 0.0302], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 13:59:49,507 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.7956, 4.7262, 3.1813, 5.4323, 4.7486, 4.7484, 2.1927, 4.7354], + device='cuda:3'), covar=tensor([0.1273, 0.0887, 0.2908, 0.0838, 0.2547, 0.1453, 0.5230, 0.1733], + device='cuda:3'), in_proj_covar=tensor([0.0255, 0.0227, 0.0268, 0.0320, 0.0313, 0.0264, 0.0281, 0.0280], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 13:59:59,376 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15191.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:00:02,747 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15195.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:00:10,004 INFO [finetune.py:976] (3/7) Epoch 3, batch 3750, loss[loss=0.2453, simple_loss=0.2999, pruned_loss=0.09529, over 4876.00 frames. ], tot_loss[loss=0.2354, simple_loss=0.2907, pruned_loss=0.09, over 954648.06 frames. ], batch size: 34, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 14:00:23,190 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5291, 1.5909, 0.8344, 1.1655, 1.8524, 1.4268, 1.2686, 1.3481], + device='cuda:3'), covar=tensor([0.0581, 0.0394, 0.0475, 0.0599, 0.0296, 0.0588, 0.0544, 0.0649], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 14:00:24,836 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.429e+02 1.976e+02 2.277e+02 2.735e+02 5.558e+02, threshold=4.554e+02, percent-clipped=2.0 +2023-04-26 14:00:55,185 INFO [finetune.py:976] (3/7) Epoch 3, batch 3800, loss[loss=0.2612, simple_loss=0.31, pruned_loss=0.1062, over 4821.00 frames. ], tot_loss[loss=0.2368, simple_loss=0.2928, pruned_loss=0.09041, over 955161.86 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 14:00:55,292 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15256.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:01:02,083 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2535, 1.5193, 1.2513, 1.4496, 1.3377, 1.1888, 1.3523, 1.1467], + device='cuda:3'), covar=tensor([0.1881, 0.1517, 0.1272, 0.1373, 0.3428, 0.1672, 0.1613, 0.2349], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0335, 0.0245, 0.0308, 0.0319, 0.0289, 0.0278, 0.0300], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:01:29,010 INFO [finetune.py:976] (3/7) Epoch 3, batch 3850, loss[loss=0.2301, simple_loss=0.2802, pruned_loss=0.08997, over 4901.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.2911, pruned_loss=0.08939, over 955409.69 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:01:38,746 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.876e+02 2.244e+02 2.688e+02 1.462e+03, threshold=4.488e+02, percent-clipped=2.0 +2023-04-26 14:02:01,731 INFO [finetune.py:976] (3/7) Epoch 3, batch 3900, loss[loss=0.2233, simple_loss=0.2797, pruned_loss=0.0834, over 4819.00 frames. ], tot_loss[loss=0.2322, simple_loss=0.2879, pruned_loss=0.08823, over 955604.40 frames. ], batch size: 41, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:02:50,622 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15396.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:03:03,039 INFO [finetune.py:976] (3/7) Epoch 3, batch 3950, loss[loss=0.2057, simple_loss=0.2632, pruned_loss=0.07415, over 4901.00 frames. ], tot_loss[loss=0.2269, simple_loss=0.2825, pruned_loss=0.08562, over 956812.03 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:03:24,417 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.749e+02 2.086e+02 2.518e+02 3.780e+02, threshold=4.171e+02, percent-clipped=0.0 +2023-04-26 14:03:54,151 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15444.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:04:08,306 INFO [finetune.py:976] (3/7) Epoch 3, batch 4000, loss[loss=0.2026, simple_loss=0.2617, pruned_loss=0.07176, over 4774.00 frames. ], tot_loss[loss=0.2251, simple_loss=0.2803, pruned_loss=0.08493, over 956482.29 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:05:04,590 INFO [finetune.py:976] (3/7) Epoch 3, batch 4050, loss[loss=0.2006, simple_loss=0.2707, pruned_loss=0.06527, over 4925.00 frames. ], tot_loss[loss=0.229, simple_loss=0.2846, pruned_loss=0.0867, over 956670.98 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:05:15,729 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.986e+02 2.201e+02 2.641e+02 4.376e+02, threshold=4.402e+02, percent-clipped=3.0 +2023-04-26 14:05:20,092 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6058, 3.9837, 0.8558, 2.0802, 2.2349, 2.6126, 2.3964, 0.9004], + device='cuda:3'), covar=tensor([0.1477, 0.1002, 0.2192, 0.1471, 0.1081, 0.1260, 0.1471, 0.2214], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0266, 0.0148, 0.0130, 0.0140, 0.0163, 0.0127, 0.0131], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:05:23,197 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 14:05:33,869 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15551.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:05:36,857 INFO [finetune.py:976] (3/7) Epoch 3, batch 4100, loss[loss=0.2397, simple_loss=0.2958, pruned_loss=0.0918, over 4828.00 frames. ], tot_loss[loss=0.2314, simple_loss=0.2875, pruned_loss=0.08767, over 955288.75 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:06:10,674 INFO [finetune.py:976] (3/7) Epoch 3, batch 4150, loss[loss=0.2812, simple_loss=0.323, pruned_loss=0.1197, over 4822.00 frames. ], tot_loss[loss=0.2309, simple_loss=0.2874, pruned_loss=0.08727, over 954991.87 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:06:23,083 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.264e+02 1.923e+02 2.257e+02 2.741e+02 5.009e+02, threshold=4.514e+02, percent-clipped=1.0 +2023-04-26 14:06:44,254 INFO [finetune.py:976] (3/7) Epoch 3, batch 4200, loss[loss=0.2242, simple_loss=0.2911, pruned_loss=0.07859, over 4789.00 frames. ], tot_loss[loss=0.2306, simple_loss=0.2877, pruned_loss=0.08672, over 954773.72 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:07:17,985 INFO [finetune.py:976] (3/7) Epoch 3, batch 4250, loss[loss=0.2448, simple_loss=0.3051, pruned_loss=0.09224, over 4775.00 frames. ], tot_loss[loss=0.2288, simple_loss=0.2856, pruned_loss=0.086, over 954698.03 frames. ], batch size: 29, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:07:29,536 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.926e+02 2.195e+02 2.674e+02 5.290e+02, threshold=4.390e+02, percent-clipped=1.0 +2023-04-26 14:07:33,803 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 14:07:50,580 INFO [finetune.py:976] (3/7) Epoch 3, batch 4300, loss[loss=0.2553, simple_loss=0.2906, pruned_loss=0.11, over 4284.00 frames. ], tot_loss[loss=0.2284, simple_loss=0.2842, pruned_loss=0.0863, over 954092.75 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:08:16,647 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-26 14:08:59,709 INFO [finetune.py:976] (3/7) Epoch 3, batch 4350, loss[loss=0.3052, simple_loss=0.3389, pruned_loss=0.1357, over 4834.00 frames. ], tot_loss[loss=0.225, simple_loss=0.2802, pruned_loss=0.08489, over 953761.13 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:09:08,071 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 14:09:21,816 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.851e+02 2.117e+02 2.553e+02 4.240e+02, threshold=4.233e+02, percent-clipped=0.0 +2023-04-26 14:10:01,714 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15851.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:10:04,718 INFO [finetune.py:976] (3/7) Epoch 3, batch 4400, loss[loss=0.2976, simple_loss=0.3391, pruned_loss=0.1281, over 4873.00 frames. ], tot_loss[loss=0.2282, simple_loss=0.2828, pruned_loss=0.08677, over 954735.42 frames. ], batch size: 34, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:10:39,595 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15883.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:10:49,611 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15899.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:10:53,840 INFO [finetune.py:976] (3/7) Epoch 3, batch 4450, loss[loss=0.2575, simple_loss=0.3125, pruned_loss=0.1012, over 4707.00 frames. ], tot_loss[loss=0.2315, simple_loss=0.2869, pruned_loss=0.08802, over 952085.14 frames. ], batch size: 59, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:11:04,060 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.361e+02 2.045e+02 2.492e+02 3.065e+02 6.089e+02, threshold=4.985e+02, percent-clipped=5.0 +2023-04-26 14:11:11,138 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5042, 2.3422, 2.7595, 2.9462, 2.7577, 2.1436, 1.9372, 2.4042], + device='cuda:3'), covar=tensor([0.1140, 0.0991, 0.0533, 0.0772, 0.0734, 0.1257, 0.1205, 0.0762], + device='cuda:3'), in_proj_covar=tensor([0.0211, 0.0212, 0.0189, 0.0185, 0.0184, 0.0201, 0.0175, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:11:20,075 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15944.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:11:27,286 INFO [finetune.py:976] (3/7) Epoch 3, batch 4500, loss[loss=0.2317, simple_loss=0.2963, pruned_loss=0.0835, over 4791.00 frames. ], tot_loss[loss=0.2339, simple_loss=0.2897, pruned_loss=0.08905, over 954794.11 frames. ], batch size: 51, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:11:53,527 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4001, 3.4980, 0.9879, 1.7969, 1.9524, 2.4000, 2.0430, 1.1251], + device='cuda:3'), covar=tensor([0.1559, 0.0878, 0.2058, 0.1430, 0.1074, 0.1131, 0.1443, 0.1983], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0266, 0.0148, 0.0130, 0.0140, 0.0163, 0.0127, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:11:59,089 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16001.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:12:02,031 INFO [finetune.py:976] (3/7) Epoch 3, batch 4550, loss[loss=0.2704, simple_loss=0.317, pruned_loss=0.1119, over 4806.00 frames. ], tot_loss[loss=0.236, simple_loss=0.2918, pruned_loss=0.09007, over 954321.28 frames. ], batch size: 41, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:12:11,702 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.282e+02 1.831e+02 2.180e+02 2.572e+02 4.192e+02, threshold=4.359e+02, percent-clipped=0.0 +2023-04-26 14:12:35,166 INFO [finetune.py:976] (3/7) Epoch 3, batch 4600, loss[loss=0.1852, simple_loss=0.2521, pruned_loss=0.05911, over 4864.00 frames. ], tot_loss[loss=0.2336, simple_loss=0.2898, pruned_loss=0.08866, over 954918.21 frames. ], batch size: 31, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:12:38,934 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16062.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:12:53,912 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16085.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:13:07,930 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 14:13:08,933 INFO [finetune.py:976] (3/7) Epoch 3, batch 4650, loss[loss=0.1887, simple_loss=0.2624, pruned_loss=0.05752, over 4909.00 frames. ], tot_loss[loss=0.231, simple_loss=0.287, pruned_loss=0.08756, over 955048.09 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:13:18,702 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.295e+02 1.883e+02 2.189e+02 2.586e+02 3.625e+02, threshold=4.377e+02, percent-clipped=0.0 +2023-04-26 14:13:47,600 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16146.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:13:49,897 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16149.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:13:53,994 INFO [finetune.py:976] (3/7) Epoch 3, batch 4700, loss[loss=0.2065, simple_loss=0.2532, pruned_loss=0.07995, over 4819.00 frames. ], tot_loss[loss=0.2286, simple_loss=0.284, pruned_loss=0.08663, over 956364.34 frames. ], batch size: 51, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:14:15,148 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1645, 1.5109, 1.9856, 2.4845, 1.9091, 1.5082, 1.2849, 1.7216], + device='cuda:3'), covar=tensor([0.4922, 0.5765, 0.2434, 0.4843, 0.5413, 0.4225, 0.7221, 0.4951], + device='cuda:3'), in_proj_covar=tensor([0.0271, 0.0270, 0.0223, 0.0340, 0.0228, 0.0233, 0.0257, 0.0205], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:14:38,783 INFO [finetune.py:976] (3/7) Epoch 3, batch 4750, loss[loss=0.3204, simple_loss=0.3388, pruned_loss=0.151, over 3938.00 frames. ], tot_loss[loss=0.2259, simple_loss=0.2812, pruned_loss=0.08535, over 956541.40 frames. ], batch size: 65, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:14:47,175 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16210.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:15:00,398 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.300e+02 1.819e+02 2.117e+02 2.422e+02 6.375e+02, threshold=4.235e+02, percent-clipped=3.0 +2023-04-26 14:15:10,809 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16239.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:15:23,779 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-26 14:15:35,035 INFO [finetune.py:976] (3/7) Epoch 3, batch 4800, loss[loss=0.2562, simple_loss=0.3264, pruned_loss=0.09305, over 4821.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2844, pruned_loss=0.08631, over 956971.15 frames. ], batch size: 51, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:16:42,025 INFO [finetune.py:976] (3/7) Epoch 3, batch 4850, loss[loss=0.2684, simple_loss=0.3099, pruned_loss=0.1135, over 4826.00 frames. ], tot_loss[loss=0.2312, simple_loss=0.2876, pruned_loss=0.08733, over 954730.43 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:16:52,894 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0575, 2.4294, 1.1187, 1.3619, 1.8643, 1.1954, 3.2227, 1.6755], + device='cuda:3'), covar=tensor([0.0709, 0.0689, 0.0803, 0.1232, 0.0555, 0.1069, 0.0264, 0.0661], + device='cuda:3'), in_proj_covar=tensor([0.0055, 0.0072, 0.0054, 0.0050, 0.0055, 0.0056, 0.0085, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 14:17:03,614 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.365e+02 1.946e+02 2.319e+02 2.886e+02 4.382e+02, threshold=4.637e+02, percent-clipped=1.0 +2023-04-26 14:17:31,156 INFO [finetune.py:976] (3/7) Epoch 3, batch 4900, loss[loss=0.3132, simple_loss=0.3518, pruned_loss=0.1373, over 4147.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.291, pruned_loss=0.08881, over 956008.82 frames. ], batch size: 65, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:17:32,328 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16357.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:17:42,321 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16371.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:18:04,092 INFO [finetune.py:976] (3/7) Epoch 3, batch 4950, loss[loss=0.2268, simple_loss=0.2861, pruned_loss=0.08381, over 4892.00 frames. ], tot_loss[loss=0.2353, simple_loss=0.2924, pruned_loss=0.08917, over 957504.27 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:18:16,360 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.286e+02 1.926e+02 2.230e+02 2.725e+02 6.217e+02, threshold=4.460e+02, percent-clipped=3.0 +2023-04-26 14:18:21,912 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16432.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:18:27,287 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16441.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:18:37,159 INFO [finetune.py:976] (3/7) Epoch 3, batch 5000, loss[loss=0.2202, simple_loss=0.2761, pruned_loss=0.08217, over 4771.00 frames. ], tot_loss[loss=0.2329, simple_loss=0.2895, pruned_loss=0.08818, over 956506.14 frames. ], batch size: 28, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:19:09,882 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16505.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:19:10,435 INFO [finetune.py:976] (3/7) Epoch 3, batch 5050, loss[loss=0.2357, simple_loss=0.31, pruned_loss=0.08065, over 4824.00 frames. ], tot_loss[loss=0.2326, simple_loss=0.2882, pruned_loss=0.08851, over 954816.76 frames. ], batch size: 41, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:19:23,710 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.221e+02 1.760e+02 2.084e+02 2.475e+02 5.733e+02, threshold=4.169e+02, percent-clipped=2.0 +2023-04-26 14:19:33,522 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16539.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:19:35,303 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3899, 1.3354, 4.1313, 3.8346, 3.6798, 3.9576, 3.9119, 3.6267], + device='cuda:3'), covar=tensor([0.7358, 0.5877, 0.1066, 0.1732, 0.0992, 0.1499, 0.1233, 0.1425], + device='cuda:3'), in_proj_covar=tensor([0.0328, 0.0315, 0.0440, 0.0442, 0.0374, 0.0427, 0.0335, 0.0395], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:3') +2023-04-26 14:19:43,694 INFO [finetune.py:976] (3/7) Epoch 3, batch 5100, loss[loss=0.2452, simple_loss=0.2826, pruned_loss=0.1039, over 4869.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.2846, pruned_loss=0.0871, over 956099.22 frames. ], batch size: 34, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:19:56,480 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5684, 2.0364, 1.4673, 1.2387, 1.1789, 1.2048, 1.4561, 1.1585], + device='cuda:3'), covar=tensor([0.2391, 0.1792, 0.2438, 0.2757, 0.3549, 0.2882, 0.1754, 0.2859], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0223, 0.0186, 0.0212, 0.0224, 0.0192, 0.0181, 0.0202], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:20:10,604 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16587.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:20:11,943 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 14:20:12,831 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-26 14:20:33,131 INFO [finetune.py:976] (3/7) Epoch 3, batch 5150, loss[loss=0.2842, simple_loss=0.3388, pruned_loss=0.1148, over 4822.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2834, pruned_loss=0.08661, over 952593.64 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:20:56,490 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.263e+02 1.945e+02 2.270e+02 2.709e+02 5.209e+02, threshold=4.540e+02, percent-clipped=2.0 +2023-04-26 14:21:30,004 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9658, 1.9692, 1.6048, 1.6157, 1.9781, 1.4441, 2.5713, 1.2674], + device='cuda:3'), covar=tensor([0.4348, 0.2069, 0.5684, 0.3610, 0.2395, 0.3115, 0.1649, 0.5290], + device='cuda:3'), in_proj_covar=tensor([0.0361, 0.0362, 0.0446, 0.0377, 0.0415, 0.0390, 0.0408, 0.0428], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:21:30,489 INFO [finetune.py:976] (3/7) Epoch 3, batch 5200, loss[loss=0.2416, simple_loss=0.311, pruned_loss=0.08609, over 4805.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2858, pruned_loss=0.08693, over 953590.07 frames. ], batch size: 41, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:21:31,198 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16657.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:22:02,898 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16705.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:22:03,453 INFO [finetune.py:976] (3/7) Epoch 3, batch 5250, loss[loss=0.2558, simple_loss=0.303, pruned_loss=0.1043, over 4759.00 frames. ], tot_loss[loss=0.2311, simple_loss=0.2879, pruned_loss=0.08714, over 953091.99 frames. ], batch size: 28, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:22:04,186 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9391, 2.4617, 1.9853, 2.2795, 1.7705, 1.9029, 2.0980, 1.6664], + device='cuda:3'), covar=tensor([0.2599, 0.1465, 0.1197, 0.1686, 0.3190, 0.1830, 0.2268, 0.3275], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0335, 0.0247, 0.0309, 0.0321, 0.0288, 0.0279, 0.0301], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:22:14,739 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.971e+02 2.303e+02 2.786e+02 5.327e+02, threshold=4.606e+02, percent-clipped=1.0 +2023-04-26 14:22:18,689 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16727.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:22:33,170 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16741.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:22:47,371 INFO [finetune.py:976] (3/7) Epoch 3, batch 5300, loss[loss=0.208, simple_loss=0.2678, pruned_loss=0.07408, over 4754.00 frames. ], tot_loss[loss=0.2323, simple_loss=0.2896, pruned_loss=0.08756, over 953805.70 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:23:10,432 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16789.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:23:20,238 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16805.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:23:20,766 INFO [finetune.py:976] (3/7) Epoch 3, batch 5350, loss[loss=0.1771, simple_loss=0.2453, pruned_loss=0.05445, over 4789.00 frames. ], tot_loss[loss=0.232, simple_loss=0.2892, pruned_loss=0.08743, over 952559.44 frames. ], batch size: 29, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:23:31,442 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.249e+02 1.831e+02 2.243e+02 2.677e+02 3.612e+02, threshold=4.486e+02, percent-clipped=0.0 +2023-04-26 14:23:33,815 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9219, 4.2781, 0.8482, 2.3666, 2.3612, 2.6720, 2.5253, 0.9550], + device='cuda:3'), covar=tensor([0.1284, 0.1118, 0.2120, 0.1271, 0.0993, 0.1237, 0.1418, 0.2193], + device='cuda:3'), in_proj_covar=tensor([0.0124, 0.0267, 0.0149, 0.0131, 0.0141, 0.0163, 0.0128, 0.0131], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:23:46,896 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-26 14:23:52,046 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16853.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:23:53,794 INFO [finetune.py:976] (3/7) Epoch 3, batch 5400, loss[loss=0.2865, simple_loss=0.3137, pruned_loss=0.1296, over 4166.00 frames. ], tot_loss[loss=0.2303, simple_loss=0.287, pruned_loss=0.08678, over 952772.21 frames. ], batch size: 65, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:24:01,174 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16868.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:24:12,191 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6123, 1.9883, 1.4689, 1.2207, 1.2156, 1.2230, 1.4049, 1.2008], + device='cuda:3'), covar=tensor([0.2286, 0.1932, 0.2269, 0.2698, 0.3409, 0.2681, 0.1767, 0.2765], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0222, 0.0186, 0.0212, 0.0224, 0.0191, 0.0180, 0.0202], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:24:21,311 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7852, 1.6339, 4.1825, 3.9063, 3.7119, 3.9895, 3.9394, 3.7718], + device='cuda:3'), covar=tensor([0.6466, 0.5258, 0.1088, 0.1625, 0.1096, 0.1543, 0.1175, 0.1371], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0310, 0.0432, 0.0436, 0.0368, 0.0420, 0.0329, 0.0390], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:24:27,334 INFO [finetune.py:976] (3/7) Epoch 3, batch 5450, loss[loss=0.2309, simple_loss=0.2849, pruned_loss=0.08852, over 4847.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2844, pruned_loss=0.0861, over 954455.67 frames. ], batch size: 44, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:24:27,446 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16906.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:24:37,663 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.358e+02 1.908e+02 2.254e+02 2.695e+02 6.070e+02, threshold=4.507e+02, percent-clipped=3.0 +2023-04-26 14:24:41,872 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16929.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 14:25:00,560 INFO [finetune.py:976] (3/7) Epoch 3, batch 5500, loss[loss=0.2051, simple_loss=0.2603, pruned_loss=0.07489, over 4798.00 frames. ], tot_loss[loss=0.2241, simple_loss=0.2802, pruned_loss=0.08404, over 956951.13 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:25:07,394 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16967.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 14:25:17,627 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-26 14:25:33,761 INFO [finetune.py:976] (3/7) Epoch 3, batch 5550, loss[loss=0.3083, simple_loss=0.3516, pruned_loss=0.1325, over 4907.00 frames. ], tot_loss[loss=0.2281, simple_loss=0.2841, pruned_loss=0.08606, over 958484.98 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:25:41,262 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-26 14:25:54,585 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.318e+02 1.795e+02 2.241e+02 2.790e+02 4.152e+02, threshold=4.483e+02, percent-clipped=0.0 +2023-04-26 14:26:02,725 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17027.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:26:13,508 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 14:26:29,085 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7954, 1.7313, 1.7619, 1.4404, 1.9201, 1.4287, 2.5309, 1.5354], + device='cuda:3'), covar=tensor([0.4580, 0.1982, 0.5386, 0.3743, 0.1947, 0.3052, 0.1702, 0.4964], + device='cuda:3'), in_proj_covar=tensor([0.0354, 0.0356, 0.0439, 0.0369, 0.0407, 0.0383, 0.0401, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:26:35,600 INFO [finetune.py:976] (3/7) Epoch 3, batch 5600, loss[loss=0.228, simple_loss=0.2917, pruned_loss=0.08215, over 4769.00 frames. ], tot_loss[loss=0.2309, simple_loss=0.2878, pruned_loss=0.08698, over 954507.98 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:26:47,147 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-04-26 14:26:59,010 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17075.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:27:17,769 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 14:27:38,803 INFO [finetune.py:976] (3/7) Epoch 3, batch 5650, loss[loss=0.2401, simple_loss=0.3096, pruned_loss=0.08531, over 4823.00 frames. ], tot_loss[loss=0.2319, simple_loss=0.2897, pruned_loss=0.08706, over 954746.59 frames. ], batch size: 40, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:27:55,218 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.353e+02 1.900e+02 2.198e+02 2.594e+02 5.346e+02, threshold=4.396e+02, percent-clipped=1.0 +2023-04-26 14:28:26,721 INFO [finetune.py:976] (3/7) Epoch 3, batch 5700, loss[loss=0.1729, simple_loss=0.2215, pruned_loss=0.06219, over 4276.00 frames. ], tot_loss[loss=0.2291, simple_loss=0.2848, pruned_loss=0.08668, over 933932.62 frames. ], batch size: 18, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:29:04,045 INFO [finetune.py:976] (3/7) Epoch 4, batch 0, loss[loss=0.2861, simple_loss=0.3361, pruned_loss=0.118, over 4828.00 frames. ], tot_loss[loss=0.2861, simple_loss=0.3361, pruned_loss=0.118, over 4828.00 frames. ], batch size: 49, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:29:04,045 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 14:29:10,880 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5905, 1.1824, 1.3984, 1.2733, 1.8183, 1.5154, 1.2158, 1.3839], + device='cuda:3'), covar=tensor([0.1772, 0.1537, 0.2216, 0.1754, 0.1242, 0.1625, 0.2623, 0.2357], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0334, 0.0348, 0.0309, 0.0345, 0.0353, 0.0314, 0.0351], + device='cuda:3'), out_proj_covar=tensor([6.7823e-05, 7.1689e-05, 7.5527e-05, 6.4718e-05, 7.3169e-05, 7.7032e-05, + 6.8408e-05, 7.5909e-05], device='cuda:3') +2023-04-26 14:29:26,720 INFO [finetune.py:1010] (3/7) Epoch 4, validation: loss=0.1686, simple_loss=0.2415, pruned_loss=0.04788, over 2265189.00 frames. +2023-04-26 14:29:26,720 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 14:29:31,431 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17189.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:29:33,611 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-26 14:29:52,776 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.687e+02 2.032e+02 2.492e+02 4.364e+02, threshold=4.064e+02, percent-clipped=0.0 +2023-04-26 14:29:53,450 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17224.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:29:57,823 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2445, 1.6386, 2.0848, 2.3416, 1.9171, 1.6312, 1.2514, 1.7771], + device='cuda:3'), covar=tensor([0.4629, 0.5704, 0.2542, 0.4233, 0.5116, 0.4101, 0.6524, 0.4321], + device='cuda:3'), in_proj_covar=tensor([0.0276, 0.0272, 0.0226, 0.0343, 0.0230, 0.0236, 0.0259, 0.0206], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:29:58,842 INFO [finetune.py:976] (3/7) Epoch 4, batch 50, loss[loss=0.2484, simple_loss=0.2999, pruned_loss=0.09848, over 4835.00 frames. ], tot_loss[loss=0.2361, simple_loss=0.2907, pruned_loss=0.09077, over 215314.64 frames. ], batch size: 44, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:30:11,289 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17250.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:30:18,607 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17262.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:30:29,503 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0087, 2.4608, 1.0374, 1.3234, 1.7627, 1.2221, 3.0640, 1.6172], + device='cuda:3'), covar=tensor([0.0682, 0.0571, 0.0784, 0.1281, 0.0528, 0.1073, 0.0259, 0.0675], + device='cuda:3'), in_proj_covar=tensor([0.0055, 0.0072, 0.0054, 0.0050, 0.0055, 0.0055, 0.0084, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 14:30:31,853 INFO [finetune.py:976] (3/7) Epoch 4, batch 100, loss[loss=0.1959, simple_loss=0.2488, pruned_loss=0.07147, over 4765.00 frames. ], tot_loss[loss=0.2253, simple_loss=0.2803, pruned_loss=0.08515, over 378656.71 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:30:46,888 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5435, 1.1613, 1.2227, 1.2593, 1.8184, 1.4272, 1.1802, 1.2266], + device='cuda:3'), covar=tensor([0.2008, 0.1627, 0.2330, 0.1517, 0.0807, 0.1945, 0.2413, 0.2133], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0337, 0.0350, 0.0312, 0.0347, 0.0356, 0.0317, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.8254e-05, 7.2157e-05, 7.5943e-05, 6.5325e-05, 7.3530e-05, 7.7595e-05, + 6.8947e-05, 7.6492e-05], device='cuda:3') +2023-04-26 14:30:52,370 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1470, 2.1756, 2.0322, 1.8255, 2.4228, 1.7308, 2.9096, 1.7306], + device='cuda:3'), covar=tensor([0.4078, 0.1952, 0.4416, 0.3460, 0.1816, 0.2994, 0.1522, 0.4371], + device='cuda:3'), in_proj_covar=tensor([0.0352, 0.0354, 0.0439, 0.0369, 0.0405, 0.0381, 0.0398, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:30:53,214 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.16 vs. limit=5.0 +2023-04-26 14:30:56,646 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4819, 3.5380, 0.8200, 1.9571, 1.8829, 2.4762, 2.0947, 1.0690], + device='cuda:3'), covar=tensor([0.1410, 0.0913, 0.2176, 0.1323, 0.1094, 0.1119, 0.1434, 0.1919], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0264, 0.0149, 0.0130, 0.0140, 0.0162, 0.0127, 0.0129], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:30:58,845 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.312e+02 1.837e+02 2.342e+02 2.858e+02 5.078e+02, threshold=4.684e+02, percent-clipped=3.0 +2023-04-26 14:31:04,970 INFO [finetune.py:976] (3/7) Epoch 4, batch 150, loss[loss=0.1843, simple_loss=0.2524, pruned_loss=0.05808, over 4812.00 frames. ], tot_loss[loss=0.2216, simple_loss=0.2763, pruned_loss=0.08345, over 508135.64 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:31:18,609 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2407, 2.5325, 0.8643, 1.5712, 1.5634, 1.9923, 1.7087, 0.8794], + device='cuda:3'), covar=tensor([0.1510, 0.1311, 0.1973, 0.1435, 0.1212, 0.1026, 0.1651, 0.1637], + device='cuda:3'), in_proj_covar=tensor([0.0124, 0.0265, 0.0149, 0.0130, 0.0141, 0.0162, 0.0127, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:31:24,778 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17362.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:31:38,033 INFO [finetune.py:976] (3/7) Epoch 4, batch 200, loss[loss=0.237, simple_loss=0.289, pruned_loss=0.09248, over 4840.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2742, pruned_loss=0.08268, over 607979.66 frames. ], batch size: 49, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:31:38,167 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17383.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:31:39,441 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8620, 2.6788, 1.8051, 1.8034, 1.3824, 1.3983, 1.9869, 1.2804], + device='cuda:3'), covar=tensor([0.2183, 0.1923, 0.2193, 0.2652, 0.3329, 0.2524, 0.1617, 0.2786], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0222, 0.0186, 0.0212, 0.0224, 0.0192, 0.0180, 0.0201], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:32:05,047 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 2.027e+02 2.314e+02 2.814e+02 1.019e+03, threshold=4.629e+02, percent-clipped=4.0 +2023-04-26 14:32:05,200 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17423.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:32:11,142 INFO [finetune.py:976] (3/7) Epoch 4, batch 250, loss[loss=0.176, simple_loss=0.2343, pruned_loss=0.05892, over 4800.00 frames. ], tot_loss[loss=0.2245, simple_loss=0.28, pruned_loss=0.08448, over 686461.82 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:32:25,164 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17444.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:33:16,610 INFO [finetune.py:976] (3/7) Epoch 4, batch 300, loss[loss=0.2314, simple_loss=0.2805, pruned_loss=0.09112, over 4866.00 frames. ], tot_loss[loss=0.2302, simple_loss=0.287, pruned_loss=0.08674, over 749250.54 frames. ], batch size: 31, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:34:04,408 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.938e+02 2.299e+02 2.709e+02 4.777e+02, threshold=4.598e+02, percent-clipped=1.0 +2023-04-26 14:34:10,651 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17524.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:34:22,587 INFO [finetune.py:976] (3/7) Epoch 4, batch 350, loss[loss=0.2499, simple_loss=0.2951, pruned_loss=0.1024, over 4716.00 frames. ], tot_loss[loss=0.2322, simple_loss=0.2895, pruned_loss=0.08741, over 795734.29 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:34:36,259 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17545.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:34:58,558 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0546, 1.1523, 1.3275, 1.4708, 1.4593, 1.6200, 1.4052, 1.4306], + device='cuda:3'), covar=tensor([0.9938, 1.6855, 1.4478, 1.2945, 1.4614, 2.3350, 1.7775, 1.4921], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0400, 0.0319, 0.0324, 0.0350, 0.0411, 0.0386, 0.0341], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:34:59,121 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17562.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:35:10,346 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17572.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:35:17,457 INFO [finetune.py:976] (3/7) Epoch 4, batch 400, loss[loss=0.1768, simple_loss=0.2418, pruned_loss=0.05593, over 4818.00 frames. ], tot_loss[loss=0.2312, simple_loss=0.2888, pruned_loss=0.08684, over 830949.61 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:35:24,670 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 14:35:36,594 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8330, 2.6030, 1.9108, 1.7130, 1.3945, 1.4134, 1.9949, 1.3321], + device='cuda:3'), covar=tensor([0.2177, 0.1995, 0.2074, 0.2683, 0.3214, 0.2444, 0.1577, 0.2725], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0221, 0.0185, 0.0210, 0.0222, 0.0191, 0.0179, 0.0201], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:35:37,142 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17610.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:35:45,055 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.309e+02 1.823e+02 2.162e+02 2.599e+02 8.047e+02, threshold=4.324e+02, percent-clipped=1.0 +2023-04-26 14:35:51,211 INFO [finetune.py:976] (3/7) Epoch 4, batch 450, loss[loss=0.2248, simple_loss=0.28, pruned_loss=0.08486, over 4885.00 frames. ], tot_loss[loss=0.229, simple_loss=0.2866, pruned_loss=0.08564, over 859525.12 frames. ], batch size: 35, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:36:00,451 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0747, 2.4636, 0.9107, 1.4627, 1.4914, 1.9314, 1.6367, 0.8810], + device='cuda:3'), covar=tensor([0.1457, 0.1224, 0.1671, 0.1354, 0.1120, 0.0965, 0.1483, 0.1659], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0266, 0.0149, 0.0129, 0.0140, 0.0162, 0.0127, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:36:20,903 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8097, 3.6908, 1.3257, 2.1558, 2.1542, 2.7594, 2.2899, 1.3278], + device='cuda:3'), covar=tensor([0.1392, 0.1277, 0.2019, 0.1301, 0.1099, 0.1172, 0.1445, 0.1828], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0266, 0.0149, 0.0129, 0.0141, 0.0162, 0.0127, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:36:25,012 INFO [finetune.py:976] (3/7) Epoch 4, batch 500, loss[loss=0.2345, simple_loss=0.2878, pruned_loss=0.09058, over 4895.00 frames. ], tot_loss[loss=0.227, simple_loss=0.2834, pruned_loss=0.08535, over 879697.55 frames. ], batch size: 35, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:36:35,732 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17700.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:36:45,650 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 14:36:48,885 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9889, 1.1887, 5.2119, 4.8902, 4.5210, 4.9082, 4.5614, 4.5692], + device='cuda:3'), covar=tensor([0.6782, 0.6773, 0.0975, 0.1675, 0.1104, 0.1102, 0.1223, 0.1387], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0314, 0.0433, 0.0435, 0.0370, 0.0421, 0.0331, 0.0389], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:36:49,494 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17718.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:36:52,487 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.778e+02 2.181e+02 2.733e+02 7.061e+02, threshold=4.362e+02, percent-clipped=3.0 +2023-04-26 14:36:56,857 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17730.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:36:58,591 INFO [finetune.py:976] (3/7) Epoch 4, batch 550, loss[loss=0.2321, simple_loss=0.2784, pruned_loss=0.09293, over 4827.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.2801, pruned_loss=0.08427, over 895963.29 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:37:02,831 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17739.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:37:10,741 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8786, 2.9026, 2.3011, 3.2966, 2.8669, 2.9092, 1.1672, 2.8165], + device='cuda:3'), covar=tensor([0.2141, 0.1601, 0.3485, 0.3295, 0.3715, 0.2187, 0.6011, 0.2863], + device='cuda:3'), in_proj_covar=tensor([0.0252, 0.0224, 0.0265, 0.0319, 0.0312, 0.0262, 0.0278, 0.0280], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:37:18,639 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17761.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:37:43,912 INFO [finetune.py:976] (3/7) Epoch 4, batch 600, loss[loss=0.2504, simple_loss=0.3128, pruned_loss=0.09406, over 4811.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2811, pruned_loss=0.08406, over 908185.29 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:37:49,341 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17791.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:37:50,842 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 14:37:51,777 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1207, 1.2830, 1.3411, 1.4984, 1.4429, 1.6331, 1.4693, 1.4269], + device='cuda:3'), covar=tensor([0.9983, 1.6037, 1.4368, 1.2328, 1.4570, 2.3288, 1.7159, 1.4406], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0402, 0.0320, 0.0325, 0.0351, 0.0413, 0.0387, 0.0343], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:38:12,407 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3632, 2.1896, 2.3368, 2.6886, 2.4768, 2.1433, 1.7958, 2.2931], + device='cuda:3'), covar=tensor([0.0945, 0.0984, 0.0607, 0.0566, 0.0682, 0.1016, 0.1009, 0.0626], + device='cuda:3'), in_proj_covar=tensor([0.0209, 0.0212, 0.0189, 0.0183, 0.0185, 0.0200, 0.0173, 0.0193], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:38:23,100 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.336e+02 1.937e+02 2.249e+02 2.763e+02 6.989e+02, threshold=4.497e+02, percent-clipped=2.0 +2023-04-26 14:38:34,866 INFO [finetune.py:976] (3/7) Epoch 4, batch 650, loss[loss=0.2376, simple_loss=0.3009, pruned_loss=0.08717, over 4829.00 frames. ], tot_loss[loss=0.2277, simple_loss=0.2845, pruned_loss=0.08547, over 917578.13 frames. ], batch size: 47, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:38:52,795 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17845.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:39:09,741 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6514, 1.9931, 1.7241, 1.8746, 1.7140, 1.6149, 1.7154, 1.3156], + device='cuda:3'), covar=tensor([0.2008, 0.1436, 0.1140, 0.1448, 0.3316, 0.1605, 0.2066, 0.3025], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0335, 0.0246, 0.0308, 0.0322, 0.0287, 0.0277, 0.0300], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:39:22,793 INFO [finetune.py:976] (3/7) Epoch 4, batch 700, loss[loss=0.183, simple_loss=0.2428, pruned_loss=0.06165, over 3963.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2869, pruned_loss=0.08635, over 923899.51 frames. ], batch size: 17, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:39:28,941 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17893.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:40:07,834 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.332e+02 1.898e+02 2.217e+02 2.664e+02 7.094e+02, threshold=4.434e+02, percent-clipped=3.0 +2023-04-26 14:40:16,784 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1925, 1.3699, 1.3482, 1.5020, 1.4177, 1.6310, 1.5300, 1.4835], + device='cuda:3'), covar=tensor([1.0368, 1.7299, 1.4658, 1.3132, 1.5145, 2.3658, 1.6971, 1.4937], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0402, 0.0320, 0.0325, 0.0351, 0.0412, 0.0387, 0.0342], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:40:19,664 INFO [finetune.py:976] (3/7) Epoch 4, batch 750, loss[loss=0.3065, simple_loss=0.3335, pruned_loss=0.1398, over 4823.00 frames. ], tot_loss[loss=0.2303, simple_loss=0.2876, pruned_loss=0.08653, over 931197.47 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:40:42,460 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-04-26 14:40:53,851 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17960.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:41:24,412 INFO [finetune.py:976] (3/7) Epoch 4, batch 800, loss[loss=0.2443, simple_loss=0.283, pruned_loss=0.1028, over 4803.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.2869, pruned_loss=0.08588, over 936571.12 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:41:48,958 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18018.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:41:51,292 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18021.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:41:52,409 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.866e+02 2.231e+02 2.895e+02 6.246e+02, threshold=4.462e+02, percent-clipped=2.0 +2023-04-26 14:41:58,981 INFO [finetune.py:976] (3/7) Epoch 4, batch 850, loss[loss=0.211, simple_loss=0.2607, pruned_loss=0.08062, over 4828.00 frames. ], tot_loss[loss=0.2271, simple_loss=0.2838, pruned_loss=0.08519, over 940818.33 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:42:02,668 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18039.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 14:42:13,542 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18056.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:42:20,540 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18066.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:42:32,274 INFO [finetune.py:976] (3/7) Epoch 4, batch 900, loss[loss=0.1936, simple_loss=0.2433, pruned_loss=0.07191, over 4808.00 frames. ], tot_loss[loss=0.2238, simple_loss=0.2806, pruned_loss=0.08351, over 944522.31 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:42:34,141 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18086.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:42:34,724 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18087.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:42:58,117 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.952e+02 2.266e+02 2.638e+02 7.997e+02, threshold=4.531e+02, percent-clipped=4.0 +2023-04-26 14:43:05,178 INFO [finetune.py:976] (3/7) Epoch 4, batch 950, loss[loss=0.1846, simple_loss=0.2446, pruned_loss=0.06227, over 4786.00 frames. ], tot_loss[loss=0.2212, simple_loss=0.2781, pruned_loss=0.08213, over 947366.38 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:43:44,171 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18174.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:43:44,769 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18175.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:43:53,136 INFO [finetune.py:976] (3/7) Epoch 4, batch 1000, loss[loss=0.2703, simple_loss=0.3322, pruned_loss=0.1042, over 4752.00 frames. ], tot_loss[loss=0.2248, simple_loss=0.282, pruned_loss=0.08381, over 948676.77 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:44:03,854 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4395, 2.3950, 1.8102, 2.1036, 2.3708, 1.7739, 3.0301, 1.4906], + device='cuda:3'), covar=tensor([0.3989, 0.1611, 0.4900, 0.3320, 0.1908, 0.2867, 0.1817, 0.4853], + device='cuda:3'), in_proj_covar=tensor([0.0353, 0.0357, 0.0441, 0.0369, 0.0406, 0.0382, 0.0399, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:44:14,198 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1161, 1.3243, 1.3858, 1.5450, 1.4437, 1.6432, 1.4439, 1.4626], + device='cuda:3'), covar=tensor([1.0905, 1.7448, 1.3820, 1.2366, 1.5252, 2.3193, 1.7266, 1.5321], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0403, 0.0320, 0.0326, 0.0351, 0.0413, 0.0387, 0.0342], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:44:30,497 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.196e+02 1.905e+02 2.183e+02 2.540e+02 5.721e+02, threshold=4.367e+02, percent-clipped=1.0 +2023-04-26 14:44:38,563 INFO [finetune.py:976] (3/7) Epoch 4, batch 1050, loss[loss=0.2686, simple_loss=0.3227, pruned_loss=0.1072, over 4207.00 frames. ], tot_loss[loss=0.227, simple_loss=0.2849, pruned_loss=0.08461, over 948001.59 frames. ], batch size: 65, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:44:39,845 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18235.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:44:40,480 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:44:43,484 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6298, 1.8293, 1.5947, 1.8034, 1.5885, 1.9493, 1.6795, 1.6768], + device='cuda:3'), covar=tensor([1.0747, 1.8673, 1.5771, 1.3148, 1.6236, 2.3167, 2.1025, 1.6928], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0402, 0.0320, 0.0325, 0.0350, 0.0412, 0.0386, 0.0341], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:45:12,450 INFO [finetune.py:976] (3/7) Epoch 4, batch 1100, loss[loss=0.2427, simple_loss=0.3006, pruned_loss=0.09239, over 4913.00 frames. ], tot_loss[loss=0.2286, simple_loss=0.2862, pruned_loss=0.08554, over 946173.99 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:45:37,843 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6244, 1.5001, 0.7321, 1.2529, 1.7908, 1.4989, 1.3580, 1.3489], + device='cuda:3'), covar=tensor([0.0549, 0.0428, 0.0448, 0.0638, 0.0283, 0.0582, 0.0561, 0.0678], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 14:45:50,512 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18316.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:46:00,505 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.371e+02 2.039e+02 2.386e+02 2.843e+02 4.542e+02, threshold=4.773e+02, percent-clipped=1.0 +2023-04-26 14:46:08,162 INFO [finetune.py:976] (3/7) Epoch 4, batch 1150, loss[loss=0.238, simple_loss=0.2935, pruned_loss=0.09127, over 4922.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2863, pruned_loss=0.08533, over 950009.89 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:46:12,667 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-26 14:46:22,667 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8365, 1.2956, 1.3689, 1.5118, 2.0186, 1.7163, 1.4139, 1.3796], + device='cuda:3'), covar=tensor([0.1654, 0.1820, 0.2020, 0.1449, 0.0850, 0.1677, 0.2132, 0.1879], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0336, 0.0350, 0.0309, 0.0346, 0.0353, 0.0315, 0.0350], + device='cuda:3'), out_proj_covar=tensor([6.8239e-05, 7.1970e-05, 7.5824e-05, 6.4693e-05, 7.3481e-05, 7.7062e-05, + 6.8502e-05, 7.5637e-05], device='cuda:3') +2023-04-26 14:46:23,838 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18356.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:46:33,595 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1121, 1.0708, 1.2564, 1.2052, 1.0669, 0.9282, 1.0237, 0.6205], + device='cuda:3'), covar=tensor([0.0711, 0.0743, 0.0639, 0.0752, 0.0947, 0.1659, 0.0576, 0.1136], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0078, 0.0077, 0.0071, 0.0083, 0.0098, 0.0087, 0.0080], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 14:46:33,980 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 14:46:40,855 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8542, 2.7733, 2.3230, 3.2673, 2.8452, 2.8318, 1.3192, 2.7565], + device='cuda:3'), covar=tensor([0.2385, 0.1662, 0.3241, 0.3059, 0.3263, 0.2429, 0.5570, 0.3042], + device='cuda:3'), in_proj_covar=tensor([0.0254, 0.0226, 0.0264, 0.0318, 0.0312, 0.0261, 0.0278, 0.0281], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:46:52,608 INFO [finetune.py:976] (3/7) Epoch 4, batch 1200, loss[loss=0.2277, simple_loss=0.2855, pruned_loss=0.08496, over 4908.00 frames. ], tot_loss[loss=0.2274, simple_loss=0.2846, pruned_loss=0.08512, over 951193.35 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:46:55,526 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18386.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:47:18,676 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18404.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:47:36,009 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.337e+02 1.888e+02 2.199e+02 2.566e+02 6.503e+02, threshold=4.398e+02, percent-clipped=1.0 +2023-04-26 14:47:42,651 INFO [finetune.py:976] (3/7) Epoch 4, batch 1250, loss[loss=0.1665, simple_loss=0.2437, pruned_loss=0.0446, over 4934.00 frames. ], tot_loss[loss=0.2249, simple_loss=0.2821, pruned_loss=0.08389, over 949760.88 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:47:43,323 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18434.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:48:15,942 INFO [finetune.py:976] (3/7) Epoch 4, batch 1300, loss[loss=0.1601, simple_loss=0.2228, pruned_loss=0.04871, over 4704.00 frames. ], tot_loss[loss=0.2202, simple_loss=0.2775, pruned_loss=0.08151, over 948386.19 frames. ], batch size: 23, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:48:17,444 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 14:48:21,734 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18490.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:48:32,742 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8859, 3.9122, 2.8367, 4.5163, 3.8619, 3.9499, 1.9537, 3.8034], + device='cuda:3'), covar=tensor([0.1607, 0.1087, 0.3307, 0.1280, 0.4044, 0.1630, 0.4930, 0.2099], + device='cuda:3'), in_proj_covar=tensor([0.0251, 0.0223, 0.0262, 0.0315, 0.0310, 0.0260, 0.0275, 0.0279], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:48:34,019 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18508.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:48:43,001 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.221e+02 1.895e+02 2.151e+02 2.623e+02 4.474e+02, threshold=4.301e+02, percent-clipped=1.0 +2023-04-26 14:48:47,802 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18530.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:48:48,424 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18531.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:48:49,575 INFO [finetune.py:976] (3/7) Epoch 4, batch 1350, loss[loss=0.2161, simple_loss=0.2931, pruned_loss=0.0696, over 4907.00 frames. ], tot_loss[loss=0.221, simple_loss=0.2777, pruned_loss=0.08213, over 948872.35 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:49:08,064 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:49:32,001 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18569.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:49:45,981 INFO [finetune.py:976] (3/7) Epoch 4, batch 1400, loss[loss=0.2197, simple_loss=0.2743, pruned_loss=0.0826, over 4768.00 frames. ], tot_loss[loss=0.2255, simple_loss=0.2829, pruned_loss=0.08408, over 951133.21 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:50:09,124 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18616.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:50:13,226 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.222e+02 1.882e+02 2.176e+02 2.822e+02 8.396e+02, threshold=4.353e+02, percent-clipped=2.0 +2023-04-26 14:50:19,724 INFO [finetune.py:976] (3/7) Epoch 4, batch 1450, loss[loss=0.2914, simple_loss=0.3449, pruned_loss=0.119, over 4892.00 frames. ], tot_loss[loss=0.2263, simple_loss=0.2844, pruned_loss=0.08411, over 952833.47 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:50:40,788 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18664.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:50:52,607 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18674.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:51:03,068 INFO [finetune.py:976] (3/7) Epoch 4, batch 1500, loss[loss=0.2462, simple_loss=0.31, pruned_loss=0.09124, over 4882.00 frames. ], tot_loss[loss=0.2263, simple_loss=0.2847, pruned_loss=0.08395, over 954226.12 frames. ], batch size: 43, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:51:56,301 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3540, 3.4502, 0.8732, 1.8158, 1.8670, 2.4444, 2.0274, 1.0839], + device='cuda:3'), covar=tensor([0.1376, 0.0883, 0.2056, 0.1360, 0.1037, 0.1091, 0.1430, 0.2110], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0266, 0.0150, 0.0130, 0.0141, 0.0163, 0.0127, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 14:51:57,996 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.902e+02 2.274e+02 2.776e+02 4.497e+02, threshold=4.548e+02, percent-clipped=1.0 +2023-04-26 14:52:08,248 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.93 vs. limit=2.0 +2023-04-26 14:52:09,747 INFO [finetune.py:976] (3/7) Epoch 4, batch 1550, loss[loss=0.235, simple_loss=0.2902, pruned_loss=0.08992, over 4801.00 frames. ], tot_loss[loss=0.2249, simple_loss=0.2834, pruned_loss=0.08318, over 950678.36 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:52:11,084 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18735.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:52:50,477 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0964, 1.4482, 1.9745, 2.4321, 1.8215, 1.4521, 1.2636, 1.6670], + device='cuda:3'), covar=tensor([0.5402, 0.5940, 0.2752, 0.4426, 0.5534, 0.4371, 0.7553, 0.4879], + device='cuda:3'), in_proj_covar=tensor([0.0273, 0.0267, 0.0222, 0.0337, 0.0226, 0.0232, 0.0254, 0.0202], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:53:10,644 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9738, 1.3661, 1.8123, 2.0894, 1.6941, 1.3451, 1.0745, 1.5334], + device='cuda:3'), covar=tensor([0.4969, 0.5647, 0.2752, 0.4077, 0.5082, 0.4336, 0.7170, 0.4593], + device='cuda:3'), in_proj_covar=tensor([0.0274, 0.0268, 0.0223, 0.0338, 0.0226, 0.0233, 0.0254, 0.0203], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:53:12,328 INFO [finetune.py:976] (3/7) Epoch 4, batch 1600, loss[loss=0.2363, simple_loss=0.2982, pruned_loss=0.08722, over 4920.00 frames. ], tot_loss[loss=0.2228, simple_loss=0.281, pruned_loss=0.08232, over 950788.45 frames. ], batch size: 43, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:53:23,466 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0944, 1.0208, 1.2163, 1.1735, 1.0388, 0.8954, 0.9658, 0.5236], + device='cuda:3'), covar=tensor([0.0792, 0.0799, 0.0736, 0.0890, 0.0983, 0.1774, 0.0665, 0.1202], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0078, 0.0077, 0.0071, 0.0083, 0.0098, 0.0087, 0.0080], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 14:53:44,153 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.258e+02 2.049e+02 2.445e+02 2.756e+02 4.210e+02, threshold=4.889e+02, percent-clipped=0.0 +2023-04-26 14:53:46,066 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.9503, 3.8497, 2.8100, 4.5564, 3.9308, 3.9751, 2.1000, 3.9039], + device='cuda:3'), covar=tensor([0.1598, 0.1210, 0.3573, 0.1320, 0.3333, 0.1857, 0.5031, 0.2139], + device='cuda:3'), in_proj_covar=tensor([0.0254, 0.0226, 0.0265, 0.0318, 0.0312, 0.0262, 0.0278, 0.0282], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 14:53:48,493 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18830.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:53:49,093 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18831.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:53:50,211 INFO [finetune.py:976] (3/7) Epoch 4, batch 1650, loss[loss=0.1841, simple_loss=0.2425, pruned_loss=0.06283, over 4857.00 frames. ], tot_loss[loss=0.2199, simple_loss=0.278, pruned_loss=0.0809, over 953792.84 frames. ], batch size: 47, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:53:58,588 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18846.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:54:07,540 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3374, 0.6349, 1.0844, 1.6734, 1.4992, 1.2533, 1.2719, 1.2929], + device='cuda:3'), covar=tensor([1.2527, 1.6941, 1.7593, 2.2202, 1.3645, 2.0060, 2.0382, 1.6343], + device='cuda:3'), in_proj_covar=tensor([0.0435, 0.0480, 0.0570, 0.0587, 0.0466, 0.0493, 0.0507, 0.0515], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:54:11,009 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18864.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 14:54:20,474 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18878.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:54:21,070 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18879.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:54:23,388 INFO [finetune.py:976] (3/7) Epoch 4, batch 1700, loss[loss=0.2265, simple_loss=0.287, pruned_loss=0.08305, over 4826.00 frames. ], tot_loss[loss=0.2192, simple_loss=0.2768, pruned_loss=0.08081, over 954317.81 frames. ], batch size: 30, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:54:43,157 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-26 14:55:17,073 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.323e+02 1.925e+02 2.287e+02 2.733e+02 5.363e+02, threshold=4.574e+02, percent-clipped=1.0 +2023-04-26 14:55:21,954 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18930.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:55:23,643 INFO [finetune.py:976] (3/7) Epoch 4, batch 1750, loss[loss=0.2178, simple_loss=0.2851, pruned_loss=0.07529, over 4897.00 frames. ], tot_loss[loss=0.2222, simple_loss=0.2797, pruned_loss=0.08239, over 951216.37 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:55:36,459 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7675, 1.2308, 4.9292, 4.5869, 4.3050, 4.6869, 4.3636, 4.4369], + device='cuda:3'), covar=tensor([0.6857, 0.6092, 0.0975, 0.1770, 0.1103, 0.1371, 0.1354, 0.1367], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0307, 0.0426, 0.0429, 0.0364, 0.0416, 0.0329, 0.0383], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 14:55:56,823 INFO [finetune.py:976] (3/7) Epoch 4, batch 1800, loss[loss=0.2481, simple_loss=0.3068, pruned_loss=0.09464, over 4908.00 frames. ], tot_loss[loss=0.2249, simple_loss=0.2832, pruned_loss=0.08331, over 952758.46 frames. ], batch size: 36, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:56:08,183 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18991.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:56:35,228 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 14:56:44,258 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.520e+02 1.971e+02 2.444e+02 3.087e+02 4.882e+02, threshold=4.887e+02, percent-clipped=3.0 +2023-04-26 14:56:44,774 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-04-26 14:56:49,159 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19030.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:56:50,954 INFO [finetune.py:976] (3/7) Epoch 4, batch 1850, loss[loss=0.1945, simple_loss=0.2605, pruned_loss=0.06421, over 4883.00 frames. ], tot_loss[loss=0.2269, simple_loss=0.2849, pruned_loss=0.08448, over 951806.21 frames. ], batch size: 32, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:57:01,807 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2124, 2.3974, 1.1628, 1.5038, 2.0742, 1.4478, 3.1522, 1.7053], + device='cuda:3'), covar=tensor([0.0477, 0.0538, 0.0649, 0.1046, 0.0396, 0.0778, 0.0239, 0.0539], + device='cuda:3'), in_proj_covar=tensor([0.0055, 0.0071, 0.0053, 0.0050, 0.0055, 0.0055, 0.0084, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 14:57:24,086 INFO [finetune.py:976] (3/7) Epoch 4, batch 1900, loss[loss=0.2512, simple_loss=0.3165, pruned_loss=0.09298, over 4735.00 frames. ], tot_loss[loss=0.2278, simple_loss=0.2865, pruned_loss=0.08451, over 951706.57 frames. ], batch size: 59, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:57:50,671 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.323e+02 1.880e+02 2.172e+02 2.548e+02 4.187e+02, threshold=4.344e+02, percent-clipped=0.0 +2023-04-26 14:57:54,925 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19129.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:57:57,175 INFO [finetune.py:976] (3/7) Epoch 4, batch 1950, loss[loss=0.2218, simple_loss=0.2598, pruned_loss=0.09189, over 4491.00 frames. ], tot_loss[loss=0.2249, simple_loss=0.2835, pruned_loss=0.08311, over 952470.20 frames. ], batch size: 19, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:58:05,517 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19146.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:58:17,028 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19164.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:58:30,344 INFO [finetune.py:976] (3/7) Epoch 4, batch 2000, loss[loss=0.2511, simple_loss=0.2991, pruned_loss=0.1015, over 4812.00 frames. ], tot_loss[loss=0.2228, simple_loss=0.2809, pruned_loss=0.08231, over 953515.88 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:58:40,464 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19190.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:58:48,412 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19194.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:59:11,328 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19212.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:59:24,209 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.742e+02 2.062e+02 2.571e+02 4.374e+02, threshold=4.123e+02, percent-clipped=1.0 +2023-04-26 14:59:36,725 INFO [finetune.py:976] (3/7) Epoch 4, batch 2050, loss[loss=0.2066, simple_loss=0.2618, pruned_loss=0.07571, over 4806.00 frames. ], tot_loss[loss=0.2194, simple_loss=0.2769, pruned_loss=0.081, over 951694.61 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:00:00,870 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-26 15:00:04,259 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5222, 3.5408, 0.8934, 1.8847, 1.8860, 2.4118, 2.0556, 1.1008], + device='cuda:3'), covar=tensor([0.1348, 0.0867, 0.2211, 0.1425, 0.1147, 0.1149, 0.1487, 0.2050], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0265, 0.0148, 0.0129, 0.0140, 0.0162, 0.0126, 0.0129], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:00:15,254 INFO [finetune.py:976] (3/7) Epoch 4, batch 2100, loss[loss=0.2013, simple_loss=0.2593, pruned_loss=0.07167, over 4837.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2774, pruned_loss=0.08113, over 952520.82 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:00:17,154 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19286.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:01:01,584 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.821e+02 2.202e+02 2.671e+02 7.978e+02, threshold=4.404e+02, percent-clipped=3.0 +2023-04-26 15:01:07,020 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19330.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:01:14,462 INFO [finetune.py:976] (3/7) Epoch 4, batch 2150, loss[loss=0.2193, simple_loss=0.2828, pruned_loss=0.07783, over 4773.00 frames. ], tot_loss[loss=0.2229, simple_loss=0.281, pruned_loss=0.08243, over 955086.67 frames. ], batch size: 54, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:02:07,704 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19378.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:02:16,425 INFO [finetune.py:976] (3/7) Epoch 4, batch 2200, loss[loss=0.2305, simple_loss=0.2916, pruned_loss=0.08465, over 4928.00 frames. ], tot_loss[loss=0.2238, simple_loss=0.2826, pruned_loss=0.08248, over 956393.25 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:02:30,495 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19394.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:03:06,441 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.252e+02 1.815e+02 2.166e+02 2.683e+02 4.330e+02, threshold=4.331e+02, percent-clipped=0.0 +2023-04-26 15:03:18,527 INFO [finetune.py:976] (3/7) Epoch 4, batch 2250, loss[loss=0.2123, simple_loss=0.2585, pruned_loss=0.08301, over 4025.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2847, pruned_loss=0.0834, over 956807.52 frames. ], batch size: 17, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:03:51,639 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19455.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:04:06,324 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.19 vs. limit=5.0 +2023-04-26 15:04:21,505 INFO [finetune.py:976] (3/7) Epoch 4, batch 2300, loss[loss=0.2082, simple_loss=0.2742, pruned_loss=0.07111, over 4763.00 frames. ], tot_loss[loss=0.2262, simple_loss=0.2856, pruned_loss=0.08339, over 956448.90 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:04:23,293 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19485.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 15:04:58,954 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.751e+02 2.123e+02 2.573e+02 6.035e+02, threshold=4.246e+02, percent-clipped=1.0 +2023-04-26 15:05:02,275 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-26 15:05:05,528 INFO [finetune.py:976] (3/7) Epoch 4, batch 2350, loss[loss=0.2111, simple_loss=0.2564, pruned_loss=0.08288, over 4828.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2821, pruned_loss=0.08233, over 955923.29 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:05:39,105 INFO [finetune.py:976] (3/7) Epoch 4, batch 2400, loss[loss=0.241, simple_loss=0.3078, pruned_loss=0.08713, over 4824.00 frames. ], tot_loss[loss=0.2203, simple_loss=0.2785, pruned_loss=0.08104, over 954969.12 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:05:41,045 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19586.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:05:41,224 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-26 15:06:00,702 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5321, 3.4903, 2.6496, 4.1230, 3.5696, 3.5699, 1.6334, 3.4294], + device='cuda:3'), covar=tensor([0.1858, 0.1233, 0.3324, 0.2215, 0.3000, 0.2002, 0.5603, 0.2807], + device='cuda:3'), in_proj_covar=tensor([0.0254, 0.0226, 0.0264, 0.0318, 0.0312, 0.0262, 0.0279, 0.0283], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 15:06:03,642 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19618.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:06:06,614 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.798e+02 2.118e+02 2.540e+02 5.281e+02, threshold=4.235e+02, percent-clipped=1.0 +2023-04-26 15:06:12,779 INFO [finetune.py:976] (3/7) Epoch 4, batch 2450, loss[loss=0.1765, simple_loss=0.2436, pruned_loss=0.05465, over 4797.00 frames. ], tot_loss[loss=0.2169, simple_loss=0.2746, pruned_loss=0.07962, over 954377.82 frames. ], batch size: 25, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:06:13,450 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19634.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:06:17,036 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5021, 3.6007, 0.9981, 2.0338, 1.9636, 2.5593, 2.0909, 1.1060], + device='cuda:3'), covar=tensor([0.1374, 0.0920, 0.1967, 0.1338, 0.1086, 0.1014, 0.1434, 0.1969], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0264, 0.0148, 0.0129, 0.0139, 0.0161, 0.0125, 0.0129], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:06:41,798 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9713, 1.5972, 1.9496, 2.4752, 1.9346, 1.5336, 1.5523, 1.7583], + device='cuda:3'), covar=tensor([0.3697, 0.4420, 0.2030, 0.3037, 0.4022, 0.3370, 0.5465, 0.3864], + device='cuda:3'), in_proj_covar=tensor([0.0272, 0.0266, 0.0221, 0.0336, 0.0224, 0.0231, 0.0251, 0.0201], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 15:06:44,223 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19679.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:06:46,564 INFO [finetune.py:976] (3/7) Epoch 4, batch 2500, loss[loss=0.2276, simple_loss=0.3003, pruned_loss=0.07744, over 4742.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2778, pruned_loss=0.08159, over 954664.01 frames. ], batch size: 54, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:06:57,008 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-26 15:07:31,981 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.364e+02 1.985e+02 2.362e+02 2.892e+02 4.639e+02, threshold=4.724e+02, percent-clipped=3.0 +2023-04-26 15:07:41,036 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6065, 2.1479, 1.5574, 1.3876, 1.2371, 1.2623, 1.5647, 1.1157], + device='cuda:3'), covar=tensor([0.2213, 0.1830, 0.2099, 0.2456, 0.3333, 0.2477, 0.1673, 0.2669], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0222, 0.0183, 0.0211, 0.0222, 0.0189, 0.0177, 0.0200], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:07:43,944 INFO [finetune.py:976] (3/7) Epoch 4, batch 2550, loss[loss=0.2022, simple_loss=0.2775, pruned_loss=0.06347, over 4896.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2827, pruned_loss=0.08326, over 956007.66 frames. ], batch size: 37, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:08:07,138 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19750.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:08:50,442 INFO [finetune.py:976] (3/7) Epoch 4, batch 2600, loss[loss=0.2185, simple_loss=0.2666, pruned_loss=0.08524, over 4860.00 frames. ], tot_loss[loss=0.2249, simple_loss=0.2831, pruned_loss=0.08335, over 956211.25 frames. ], batch size: 31, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:08:51,171 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4367, 1.1681, 0.5796, 1.1322, 1.1089, 1.3513, 1.2270, 1.1642], + device='cuda:3'), covar=tensor([0.0532, 0.0418, 0.0484, 0.0581, 0.0339, 0.0524, 0.0525, 0.0612], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 15:08:57,303 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19785.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 15:09:23,390 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.319e+02 1.754e+02 2.126e+02 2.494e+02 3.908e+02, threshold=4.253e+02, percent-clipped=0.0 +2023-04-26 15:09:29,904 INFO [finetune.py:976] (3/7) Epoch 4, batch 2650, loss[loss=0.2249, simple_loss=0.3015, pruned_loss=0.0741, over 4854.00 frames. ], tot_loss[loss=0.2272, simple_loss=0.2857, pruned_loss=0.08429, over 956598.86 frames. ], batch size: 44, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:09:29,969 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19833.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 15:10:35,940 INFO [finetune.py:976] (3/7) Epoch 4, batch 2700, loss[loss=0.2086, simple_loss=0.2464, pruned_loss=0.08541, over 4225.00 frames. ], tot_loss[loss=0.2263, simple_loss=0.2845, pruned_loss=0.084, over 955964.30 frames. ], batch size: 18, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:10:45,476 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4752, 0.7977, 1.1400, 1.0802, 1.5765, 1.2657, 1.0508, 1.0816], + device='cuda:3'), covar=tensor([0.1645, 0.1854, 0.2371, 0.1557, 0.0920, 0.1639, 0.2075, 0.2160], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0338, 0.0353, 0.0311, 0.0352, 0.0355, 0.0317, 0.0355], + device='cuda:3'), out_proj_covar=tensor([6.8590e-05, 7.2623e-05, 7.6526e-05, 6.5199e-05, 7.4796e-05, 7.7314e-05, + 6.8928e-05, 7.6728e-05], device='cuda:3') +2023-04-26 15:11:06,178 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7895, 1.2251, 1.3715, 1.4260, 2.0171, 1.6556, 1.3554, 1.3467], + device='cuda:3'), covar=tensor([0.1774, 0.1797, 0.2186, 0.1327, 0.0951, 0.1529, 0.2143, 0.2173], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0338, 0.0352, 0.0311, 0.0351, 0.0354, 0.0316, 0.0355], + device='cuda:3'), out_proj_covar=tensor([6.8421e-05, 7.2494e-05, 7.6248e-05, 6.5066e-05, 7.4673e-05, 7.7157e-05, + 6.8761e-05, 7.6622e-05], device='cuda:3') +2023-04-26 15:11:14,328 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 15:11:15,310 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.807e+02 2.180e+02 2.734e+02 4.615e+02, threshold=4.361e+02, percent-clipped=4.0 +2023-04-26 15:11:18,845 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 15:11:21,337 INFO [finetune.py:976] (3/7) Epoch 4, batch 2750, loss[loss=0.2208, simple_loss=0.2663, pruned_loss=0.0877, over 4824.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2806, pruned_loss=0.08229, over 956176.73 frames. ], batch size: 49, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:11:49,856 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19974.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:11:55,248 INFO [finetune.py:976] (3/7) Epoch 4, batch 2800, loss[loss=0.1974, simple_loss=0.2519, pruned_loss=0.0715, over 4895.00 frames. ], tot_loss[loss=0.2189, simple_loss=0.2765, pruned_loss=0.0806, over 955999.32 frames. ], batch size: 32, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:12:21,322 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20020.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:12:23,470 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.778e+02 2.116e+02 2.453e+02 4.080e+02, threshold=4.233e+02, percent-clipped=0.0 +2023-04-26 15:12:30,050 INFO [finetune.py:976] (3/7) Epoch 4, batch 2850, loss[loss=0.2416, simple_loss=0.3013, pruned_loss=0.09093, over 4813.00 frames. ], tot_loss[loss=0.2181, simple_loss=0.2753, pruned_loss=0.08043, over 955060.69 frames. ], batch size: 41, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:12:30,428 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.18 vs. limit=5.0 +2023-04-26 15:12:40,870 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20050.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:02,648 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20081.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:03,758 INFO [finetune.py:976] (3/7) Epoch 4, batch 2900, loss[loss=0.2365, simple_loss=0.299, pruned_loss=0.087, over 4819.00 frames. ], tot_loss[loss=0.2207, simple_loss=0.2783, pruned_loss=0.08157, over 952647.17 frames. ], batch size: 40, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:13:07,522 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20089.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:13,380 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20098.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:22,345 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20112.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:27,053 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3825, 2.8781, 1.0218, 1.4354, 2.2819, 1.5882, 4.1943, 2.0423], + device='cuda:3'), covar=tensor([0.0684, 0.1000, 0.0982, 0.1372, 0.0594, 0.1001, 0.0260, 0.0653], + device='cuda:3'), in_proj_covar=tensor([0.0055, 0.0071, 0.0053, 0.0050, 0.0055, 0.0055, 0.0084, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 15:13:29,803 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.335e+02 1.818e+02 2.275e+02 2.799e+02 4.560e+02, threshold=4.551e+02, percent-clipped=2.0 +2023-04-26 15:13:47,514 INFO [finetune.py:976] (3/7) Epoch 4, batch 2950, loss[loss=0.2686, simple_loss=0.3248, pruned_loss=0.1062, over 4794.00 frames. ], tot_loss[loss=0.2232, simple_loss=0.2812, pruned_loss=0.08258, over 953145.58 frames. ], batch size: 51, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:14:09,930 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20150.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:14:35,351 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20173.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:14:54,003 INFO [finetune.py:976] (3/7) Epoch 4, batch 3000, loss[loss=0.257, simple_loss=0.3056, pruned_loss=0.1042, over 4894.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2829, pruned_loss=0.08317, over 953883.14 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:14:54,004 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 15:15:10,816 INFO [finetune.py:1010] (3/7) Epoch 4, validation: loss=0.1632, simple_loss=0.2363, pruned_loss=0.04509, over 2265189.00 frames. +2023-04-26 15:15:10,816 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 15:15:17,699 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4542, 3.4792, 1.0520, 1.8723, 1.8901, 2.3295, 2.0427, 1.0338], + device='cuda:3'), covar=tensor([0.1614, 0.1459, 0.2268, 0.1591, 0.1260, 0.1393, 0.1524, 0.2233], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0264, 0.0148, 0.0129, 0.0138, 0.0161, 0.0125, 0.0128], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:15:35,631 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20204.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:15:47,601 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.833e+02 2.218e+02 2.696e+02 4.809e+02, threshold=4.435e+02, percent-clipped=1.0 +2023-04-26 15:16:05,213 INFO [finetune.py:976] (3/7) Epoch 4, batch 3050, loss[loss=0.2381, simple_loss=0.2991, pruned_loss=0.08858, over 4818.00 frames. ], tot_loss[loss=0.2264, simple_loss=0.2846, pruned_loss=0.08407, over 954245.39 frames. ], batch size: 30, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:16:35,263 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7211, 2.0668, 5.6286, 5.2454, 4.9304, 5.1764, 4.8711, 4.9846], + device='cuda:3'), covar=tensor([0.5544, 0.5415, 0.0899, 0.1814, 0.0927, 0.1099, 0.1026, 0.1641], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0308, 0.0425, 0.0431, 0.0364, 0.0416, 0.0326, 0.0386], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:16:43,900 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:16:49,424 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20274.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:16:55,395 INFO [finetune.py:976] (3/7) Epoch 4, batch 3100, loss[loss=0.1889, simple_loss=0.2463, pruned_loss=0.06571, over 4811.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2817, pruned_loss=0.08254, over 954742.96 frames. ], batch size: 41, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:17:02,963 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6532, 1.5335, 1.7521, 1.9509, 1.9539, 1.5308, 1.1983, 1.6928], + device='cuda:3'), covar=tensor([0.0938, 0.1200, 0.0672, 0.0611, 0.0642, 0.1012, 0.1029, 0.0666], + device='cuda:3'), in_proj_covar=tensor([0.0208, 0.0209, 0.0186, 0.0181, 0.0182, 0.0198, 0.0171, 0.0193], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:17:21,575 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20322.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:17:22,124 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.364e+02 1.832e+02 2.083e+02 2.367e+02 4.071e+02, threshold=4.166e+02, percent-clipped=0.0 +2023-04-26 15:17:24,400 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-26 15:17:28,240 INFO [finetune.py:976] (3/7) Epoch 4, batch 3150, loss[loss=0.2173, simple_loss=0.2611, pruned_loss=0.0868, over 4809.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2783, pruned_loss=0.08137, over 956618.40 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:17:57,215 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20376.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:01,417 INFO [finetune.py:976] (3/7) Epoch 4, batch 3200, loss[loss=0.1889, simple_loss=0.2428, pruned_loss=0.06746, over 4107.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2757, pruned_loss=0.08019, over 956763.33 frames. ], batch size: 17, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:18:04,543 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20388.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:25,094 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20418.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:28,521 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 1.902e+02 2.131e+02 2.743e+02 4.123e+02, threshold=4.262e+02, percent-clipped=0.0 +2023-04-26 15:18:33,018 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9454, 1.1095, 1.3452, 1.4837, 1.4030, 1.6286, 1.4196, 1.4406], + device='cuda:3'), covar=tensor([0.9639, 1.3567, 1.1720, 1.0454, 1.3100, 1.9961, 1.3587, 1.2378], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0400, 0.0321, 0.0327, 0.0351, 0.0412, 0.0386, 0.0339], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 15:18:34,645 INFO [finetune.py:976] (3/7) Epoch 4, batch 3250, loss[loss=0.2264, simple_loss=0.3003, pruned_loss=0.07627, over 4826.00 frames. ], tot_loss[loss=0.2188, simple_loss=0.2764, pruned_loss=0.08055, over 955647.22 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 64.0 +2023-04-26 15:18:42,468 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20445.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:45,886 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20449.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:19:04,240 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20468.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:19:11,493 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20479.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:19:13,810 INFO [finetune.py:976] (3/7) Epoch 4, batch 3300, loss[loss=0.2526, simple_loss=0.3142, pruned_loss=0.09543, over 4945.00 frames. ], tot_loss[loss=0.2228, simple_loss=0.2809, pruned_loss=0.08236, over 956713.20 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:19:33,453 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.04 vs. limit=5.0 +2023-04-26 15:19:47,324 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.894e+02 2.274e+02 2.897e+02 7.817e+02, threshold=4.548e+02, percent-clipped=5.0 +2023-04-26 15:19:58,661 INFO [finetune.py:976] (3/7) Epoch 4, batch 3350, loss[loss=0.2725, simple_loss=0.3107, pruned_loss=0.1171, over 4143.00 frames. ], tot_loss[loss=0.2248, simple_loss=0.2835, pruned_loss=0.08305, over 955633.26 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:20:39,485 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20560.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:20:55,812 INFO [finetune.py:976] (3/7) Epoch 4, batch 3400, loss[loss=0.1967, simple_loss=0.2683, pruned_loss=0.06254, over 4775.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2839, pruned_loss=0.08264, over 955791.40 frames. ], batch size: 51, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:21:04,540 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20588.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:21:29,470 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.806e+02 2.102e+02 2.435e+02 3.817e+02, threshold=4.203e+02, percent-clipped=0.0 +2023-04-26 15:21:34,361 INFO [finetune.py:976] (3/7) Epoch 4, batch 3450, loss[loss=0.2244, simple_loss=0.2838, pruned_loss=0.08249, over 4933.00 frames. ], tot_loss[loss=0.2253, simple_loss=0.2847, pruned_loss=0.08288, over 957385.87 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:21:44,634 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20649.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:22:19,739 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20676.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:22:23,908 INFO [finetune.py:976] (3/7) Epoch 4, batch 3500, loss[loss=0.2554, simple_loss=0.308, pruned_loss=0.1014, over 4828.00 frames. ], tot_loss[loss=0.2227, simple_loss=0.2812, pruned_loss=0.08213, over 954527.35 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:22:58,021 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20724.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:22:58,550 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.751e+02 2.119e+02 2.772e+02 4.921e+02, threshold=4.238e+02, percent-clipped=1.0 +2023-04-26 15:23:03,442 INFO [finetune.py:976] (3/7) Epoch 4, batch 3550, loss[loss=0.2423, simple_loss=0.2833, pruned_loss=0.1006, over 4863.00 frames. ], tot_loss[loss=0.2216, simple_loss=0.2791, pruned_loss=0.08209, over 956088.61 frames. ], batch size: 34, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:23:13,727 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0282, 2.5923, 1.0948, 1.2670, 2.0550, 1.2275, 3.3897, 1.5832], + device='cuda:3'), covar=tensor([0.0703, 0.0771, 0.0860, 0.1383, 0.0499, 0.1034, 0.0257, 0.0688], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0071, 0.0053, 0.0050, 0.0055, 0.0055, 0.0083, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 15:23:15,953 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20744.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:23:16,590 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20745.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:23:44,324 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.01 vs. limit=5.0 +2023-04-26 15:23:48,062 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20768.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:23:58,632 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20774.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:24:04,024 INFO [finetune.py:976] (3/7) Epoch 4, batch 3600, loss[loss=0.26, simple_loss=0.3098, pruned_loss=0.1051, over 4886.00 frames. ], tot_loss[loss=0.22, simple_loss=0.2767, pruned_loss=0.08162, over 953198.00 frames. ], batch size: 34, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:24:10,256 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20793.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:24:31,211 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20816.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:24:38,615 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.726e+02 2.072e+02 2.545e+02 4.959e+02, threshold=4.144e+02, percent-clipped=1.0 +2023-04-26 15:24:43,526 INFO [finetune.py:976] (3/7) Epoch 4, batch 3650, loss[loss=0.2724, simple_loss=0.3262, pruned_loss=0.1093, over 4906.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.279, pruned_loss=0.082, over 953969.47 frames. ], batch size: 43, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:25:00,521 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20860.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:25:05,272 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4068, 0.9190, 0.3286, 1.1031, 1.1733, 1.3031, 1.1795, 1.1514], + device='cuda:3'), covar=tensor([0.0532, 0.0444, 0.0511, 0.0606, 0.0321, 0.0560, 0.0534, 0.0621], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 15:25:16,800 INFO [finetune.py:976] (3/7) Epoch 4, batch 3700, loss[loss=0.2326, simple_loss=0.2854, pruned_loss=0.08989, over 4924.00 frames. ], tot_loss[loss=0.2251, simple_loss=0.2828, pruned_loss=0.08368, over 952777.90 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:25:32,434 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20908.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:25:50,198 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.325e+02 1.979e+02 2.308e+02 2.768e+02 4.690e+02, threshold=4.617e+02, percent-clipped=4.0 +2023-04-26 15:25:53,233 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-04-26 15:26:01,985 INFO [finetune.py:976] (3/7) Epoch 4, batch 3750, loss[loss=0.2711, simple_loss=0.319, pruned_loss=0.1116, over 4814.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2863, pruned_loss=0.08538, over 953353.22 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:26:02,123 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8944, 1.4211, 1.4503, 1.4993, 2.0632, 1.6338, 1.3637, 1.4582], + device='cuda:3'), covar=tensor([0.1724, 0.1704, 0.2229, 0.1474, 0.0917, 0.2189, 0.2386, 0.1826], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0338, 0.0354, 0.0311, 0.0351, 0.0353, 0.0316, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.8439e-05, 7.2583e-05, 7.6686e-05, 6.5023e-05, 7.4559e-05, 7.6945e-05, + 6.8646e-05, 7.6354e-05], device='cuda:3') +2023-04-26 15:26:12,694 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 15:26:14,230 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20944.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:27:02,461 INFO [finetune.py:976] (3/7) Epoch 4, batch 3800, loss[loss=0.1901, simple_loss=0.257, pruned_loss=0.06155, over 4762.00 frames. ], tot_loss[loss=0.2268, simple_loss=0.285, pruned_loss=0.08431, over 953396.19 frames. ], batch size: 28, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:27:19,859 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2690, 1.2527, 1.3718, 1.5514, 1.5251, 1.2183, 0.9499, 1.3732], + device='cuda:3'), covar=tensor([0.0925, 0.1378, 0.0879, 0.0675, 0.0730, 0.0967, 0.1044, 0.0684], + device='cuda:3'), in_proj_covar=tensor([0.0210, 0.0211, 0.0188, 0.0183, 0.0183, 0.0200, 0.0172, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:27:22,329 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0178, 1.7542, 1.9805, 2.4327, 2.2670, 1.8210, 1.6213, 2.0451], + device='cuda:3'), covar=tensor([0.0928, 0.1129, 0.0661, 0.0525, 0.0620, 0.0962, 0.1028, 0.0617], + device='cuda:3'), in_proj_covar=tensor([0.0210, 0.0210, 0.0188, 0.0183, 0.0183, 0.0200, 0.0172, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:27:46,870 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.703e+02 2.110e+02 2.594e+02 5.214e+02, threshold=4.219e+02, percent-clipped=1.0 +2023-04-26 15:28:05,804 INFO [finetune.py:976] (3/7) Epoch 4, batch 3850, loss[loss=0.2136, simple_loss=0.2681, pruned_loss=0.07949, over 4676.00 frames. ], tot_loss[loss=0.2241, simple_loss=0.2831, pruned_loss=0.08255, over 954755.01 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:28:07,622 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8917, 1.7915, 1.9181, 2.2728, 2.1451, 1.7049, 1.4846, 1.9077], + device='cuda:3'), covar=tensor([0.1036, 0.1177, 0.0770, 0.0595, 0.0651, 0.1079, 0.1091, 0.0728], + device='cuda:3'), in_proj_covar=tensor([0.0210, 0.0211, 0.0188, 0.0183, 0.0183, 0.0200, 0.0172, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:28:18,891 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21044.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:28:29,650 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21052.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:28:54,709 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21074.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:11,949 INFO [finetune.py:976] (3/7) Epoch 4, batch 3900, loss[loss=0.1987, simple_loss=0.2556, pruned_loss=0.07085, over 4770.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2808, pruned_loss=0.08222, over 955507.07 frames. ], batch size: 28, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:29:18,472 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21092.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:21,608 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7400, 1.9994, 1.7355, 1.9207, 1.7145, 2.1083, 1.7823, 1.7793], + device='cuda:3'), covar=tensor([0.9495, 1.5773, 1.3959, 1.2539, 1.3789, 2.0968, 1.7353, 1.5169], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0401, 0.0322, 0.0328, 0.0351, 0.0415, 0.0387, 0.0342], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 15:29:28,203 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3665, 3.0724, 0.8712, 1.4243, 2.3157, 1.4990, 4.3670, 2.2014], + device='cuda:3'), covar=tensor([0.0704, 0.0682, 0.0983, 0.1383, 0.0574, 0.1028, 0.0218, 0.0615], + device='cuda:3'), in_proj_covar=tensor([0.0055, 0.0071, 0.0053, 0.0050, 0.0055, 0.0055, 0.0084, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 15:29:30,664 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0736, 1.5340, 5.4231, 4.9965, 4.6430, 5.0432, 4.7755, 4.7600], + device='cuda:3'), covar=tensor([0.6303, 0.5951, 0.0966, 0.1904, 0.1069, 0.1411, 0.1012, 0.1512], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0305, 0.0420, 0.0429, 0.0361, 0.0414, 0.0325, 0.0383], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:29:31,295 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21113.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:34,794 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21118.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:37,129 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21122.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:38,882 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.817e+02 2.053e+02 2.461e+02 7.075e+02, threshold=4.105e+02, percent-clipped=2.0 +2023-04-26 15:29:43,719 INFO [finetune.py:976] (3/7) Epoch 4, batch 3950, loss[loss=0.2073, simple_loss=0.2652, pruned_loss=0.07473, over 4832.00 frames. ], tot_loss[loss=0.2185, simple_loss=0.2766, pruned_loss=0.0802, over 954647.27 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:30:02,749 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7249, 2.3693, 1.6467, 1.5337, 1.3113, 1.3467, 1.7274, 1.2771], + device='cuda:3'), covar=tensor([0.2198, 0.1809, 0.2337, 0.2751, 0.3328, 0.2603, 0.1562, 0.2868], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0221, 0.0181, 0.0210, 0.0219, 0.0188, 0.0175, 0.0198], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:30:14,124 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21179.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:30:16,453 INFO [finetune.py:976] (3/7) Epoch 4, batch 4000, loss[loss=0.2141, simple_loss=0.2738, pruned_loss=0.07719, over 4916.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2754, pruned_loss=0.08006, over 955879.40 frames. ], batch size: 37, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:30:38,888 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5450, 1.1298, 4.4183, 4.0720, 3.8438, 4.1112, 4.0223, 3.8314], + device='cuda:3'), covar=tensor([0.7307, 0.6506, 0.1022, 0.1940, 0.1090, 0.1713, 0.1634, 0.1506], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0307, 0.0420, 0.0430, 0.0362, 0.0416, 0.0325, 0.0385], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:30:45,241 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.385e+02 1.961e+02 2.328e+02 2.721e+02 5.358e+02, threshold=4.656e+02, percent-clipped=3.0 +2023-04-26 15:30:50,170 INFO [finetune.py:976] (3/7) Epoch 4, batch 4050, loss[loss=0.2208, simple_loss=0.2985, pruned_loss=0.07149, over 4813.00 frames. ], tot_loss[loss=0.2212, simple_loss=0.2794, pruned_loss=0.08153, over 956478.33 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:30:56,603 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6908, 1.4535, 3.8897, 3.6276, 3.4467, 3.5806, 3.5294, 3.4497], + device='cuda:3'), covar=tensor([0.6482, 0.5455, 0.1079, 0.1788, 0.1031, 0.1687, 0.2850, 0.1372], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0306, 0.0419, 0.0428, 0.0361, 0.0414, 0.0324, 0.0383], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:30:58,875 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21244.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:02,473 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7487, 0.9532, 1.1585, 1.3173, 1.3207, 1.5287, 1.2216, 1.2303], + device='cuda:3'), covar=tensor([0.8079, 1.1243, 0.9887, 0.9374, 1.0970, 1.6832, 1.2114, 1.0632], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0398, 0.0319, 0.0326, 0.0348, 0.0411, 0.0384, 0.0339], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 15:31:07,570 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-26 15:31:23,491 INFO [finetune.py:976] (3/7) Epoch 4, batch 4100, loss[loss=0.2152, simple_loss=0.2688, pruned_loss=0.08081, over 4058.00 frames. ], tot_loss[loss=0.2239, simple_loss=0.2827, pruned_loss=0.08256, over 955388.92 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:31:30,026 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21292.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:43,176 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21312.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:43,820 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6675, 3.4798, 2.5406, 2.7265, 2.0587, 2.1275, 2.6807, 2.0627], + device='cuda:3'), covar=tensor([0.1718, 0.1551, 0.1827, 0.2015, 0.2842, 0.2095, 0.1326, 0.2234], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0220, 0.0181, 0.0209, 0.0219, 0.0187, 0.0174, 0.0198], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:31:43,844 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3131, 2.3706, 2.5433, 2.5991, 2.5137, 2.2259, 2.3530, 2.3408], + device='cuda:3'), covar=tensor([1.2804, 1.6309, 1.9079, 1.9399, 1.3172, 2.1967, 2.1771, 1.6326], + device='cuda:3'), in_proj_covar=tensor([0.0427, 0.0470, 0.0558, 0.0576, 0.0459, 0.0484, 0.0498, 0.0501], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:31:50,914 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.243e+02 1.937e+02 2.313e+02 2.737e+02 5.004e+02, threshold=4.625e+02, percent-clipped=1.0 +2023-04-26 15:31:52,900 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21328.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:56,304 INFO [finetune.py:976] (3/7) Epoch 4, batch 4150, loss[loss=0.2146, simple_loss=0.2874, pruned_loss=0.07093, over 4789.00 frames. ], tot_loss[loss=0.2244, simple_loss=0.2836, pruned_loss=0.08258, over 955314.97 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:32:12,382 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 15:32:40,030 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21373.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:32:46,014 INFO [finetune.py:976] (3/7) Epoch 4, batch 4200, loss[loss=0.2079, simple_loss=0.2607, pruned_loss=0.07759, over 4782.00 frames. ], tot_loss[loss=0.2252, simple_loss=0.2846, pruned_loss=0.08292, over 956151.87 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:32:50,234 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21389.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:32:59,824 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.4059, 1.4490, 1.4300, 1.1165, 1.4017, 1.2557, 1.7826, 1.2989], + device='cuda:3'), covar=tensor([0.3758, 0.1599, 0.4836, 0.2622, 0.1630, 0.1923, 0.1641, 0.4528], + device='cuda:3'), in_proj_covar=tensor([0.0357, 0.0358, 0.0443, 0.0372, 0.0405, 0.0385, 0.0399, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:33:21,610 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21408.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:33:42,773 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.624e+02 1.995e+02 2.490e+02 4.928e+02, threshold=3.989e+02, percent-clipped=1.0 +2023-04-26 15:33:52,193 INFO [finetune.py:976] (3/7) Epoch 4, batch 4250, loss[loss=0.2436, simple_loss=0.294, pruned_loss=0.0966, over 4861.00 frames. ], tot_loss[loss=0.222, simple_loss=0.2812, pruned_loss=0.0814, over 954753.63 frames. ], batch size: 34, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:34:03,817 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 15:34:30,946 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21474.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:34:41,946 INFO [finetune.py:976] (3/7) Epoch 4, batch 4300, loss[loss=0.1901, simple_loss=0.2566, pruned_loss=0.06182, over 4804.00 frames. ], tot_loss[loss=0.2196, simple_loss=0.278, pruned_loss=0.08063, over 955146.15 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:35:24,193 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7801, 1.6266, 1.9796, 2.0580, 1.9895, 1.6804, 1.7725, 1.9012], + device='cuda:3'), covar=tensor([1.3764, 1.8402, 2.0247, 1.9777, 1.3759, 2.4139, 2.2461, 1.7744], + device='cuda:3'), in_proj_covar=tensor([0.0427, 0.0470, 0.0557, 0.0577, 0.0459, 0.0484, 0.0498, 0.0501], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:35:27,096 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.707e+02 2.018e+02 2.403e+02 4.975e+02, threshold=4.035e+02, percent-clipped=3.0 +2023-04-26 15:35:31,900 INFO [finetune.py:976] (3/7) Epoch 4, batch 4350, loss[loss=0.1966, simple_loss=0.254, pruned_loss=0.06964, over 4829.00 frames. ], tot_loss[loss=0.2169, simple_loss=0.275, pruned_loss=0.07936, over 955559.39 frames. ], batch size: 30, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:35:54,180 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.25 vs. limit=5.0 +2023-04-26 15:36:02,097 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-26 15:36:16,127 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0360, 1.8654, 2.0899, 2.3941, 2.3093, 1.8273, 1.5495, 2.1014], + device='cuda:3'), covar=tensor([0.0879, 0.1189, 0.0605, 0.0599, 0.0584, 0.0984, 0.0980, 0.0565], + device='cuda:3'), in_proj_covar=tensor([0.0206, 0.0206, 0.0183, 0.0179, 0.0179, 0.0195, 0.0168, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:36:37,729 INFO [finetune.py:976] (3/7) Epoch 4, batch 4400, loss[loss=0.279, simple_loss=0.3393, pruned_loss=0.1094, over 4805.00 frames. ], tot_loss[loss=0.217, simple_loss=0.2752, pruned_loss=0.07941, over 955124.15 frames. ], batch size: 41, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:36:51,047 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21595.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:09,365 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21613.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:17,115 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.848e+02 2.167e+02 2.564e+02 6.946e+02, threshold=4.335e+02, percent-clipped=3.0 +2023-04-26 15:37:22,022 INFO [finetune.py:976] (3/7) Epoch 4, batch 4450, loss[loss=0.2382, simple_loss=0.3005, pruned_loss=0.08798, over 4814.00 frames. ], tot_loss[loss=0.2208, simple_loss=0.2792, pruned_loss=0.0812, over 955040.34 frames. ], batch size: 40, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:37:28,954 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.86 vs. limit=5.0 +2023-04-26 15:37:36,746 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21656.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:40,850 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.04 vs. limit=5.0 +2023-04-26 15:37:46,469 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21668.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:50,684 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21674.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:53,170 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5153, 0.9514, 1.3714, 1.8176, 1.6239, 1.4245, 1.4335, 1.4951], + device='cuda:3'), covar=tensor([1.2299, 1.6260, 1.7452, 2.0062, 1.4542, 1.9312, 1.9098, 1.4640], + device='cuda:3'), in_proj_covar=tensor([0.0427, 0.0468, 0.0555, 0.0574, 0.0458, 0.0483, 0.0496, 0.0500], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:37:56,117 INFO [finetune.py:976] (3/7) Epoch 4, batch 4500, loss[loss=0.1848, simple_loss=0.2319, pruned_loss=0.06883, over 4231.00 frames. ], tot_loss[loss=0.2209, simple_loss=0.2805, pruned_loss=0.0807, over 954399.66 frames. ], batch size: 18, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:37:56,780 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21684.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:11,962 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21708.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:24,949 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.777e+02 2.074e+02 2.624e+02 7.543e+02, threshold=4.148e+02, percent-clipped=2.0 +2023-04-26 15:38:29,906 INFO [finetune.py:976] (3/7) Epoch 4, batch 4550, loss[loss=0.1973, simple_loss=0.2616, pruned_loss=0.06649, over 4901.00 frames. ], tot_loss[loss=0.2219, simple_loss=0.2814, pruned_loss=0.08117, over 955814.46 frames. ], batch size: 37, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:38:44,652 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21756.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:48,420 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7794, 1.3884, 1.3294, 1.5709, 2.0376, 1.7101, 1.3647, 1.2912], + device='cuda:3'), covar=tensor([0.2094, 0.2040, 0.2571, 0.1643, 0.0971, 0.1935, 0.2830, 0.2567], + device='cuda:3'), in_proj_covar=tensor([0.0319, 0.0339, 0.0357, 0.0312, 0.0352, 0.0354, 0.0317, 0.0357], + device='cuda:3'), out_proj_covar=tensor([6.8724e-05, 7.2673e-05, 7.7337e-05, 6.5197e-05, 7.4800e-05, 7.7127e-05, + 6.8853e-05, 7.7057e-05], device='cuda:3') +2023-04-26 15:38:49,063 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:58,619 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21774.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:39:04,077 INFO [finetune.py:976] (3/7) Epoch 4, batch 4600, loss[loss=0.1874, simple_loss=0.25, pruned_loss=0.06238, over 4829.00 frames. ], tot_loss[loss=0.2209, simple_loss=0.2804, pruned_loss=0.08072, over 954875.92 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:39:30,457 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21822.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:39:36,926 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21824.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:39:37,415 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.786e+02 2.120e+02 2.552e+02 5.015e+02, threshold=4.240e+02, percent-clipped=1.0 +2023-04-26 15:39:48,032 INFO [finetune.py:976] (3/7) Epoch 4, batch 4650, loss[loss=0.2061, simple_loss=0.247, pruned_loss=0.08254, over 4064.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2784, pruned_loss=0.08054, over 954473.51 frames. ], batch size: 17, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:40:04,074 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21850.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:40:40,673 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0410, 0.9884, 1.2361, 1.1646, 0.9928, 0.8797, 0.9936, 0.5444], + device='cuda:3'), covar=tensor([0.0830, 0.0751, 0.0690, 0.0785, 0.0979, 0.1579, 0.0600, 0.1151], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0077, 0.0075, 0.0069, 0.0080, 0.0096, 0.0084, 0.0078], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:40:49,526 INFO [finetune.py:976] (3/7) Epoch 4, batch 4700, loss[loss=0.2054, simple_loss=0.2553, pruned_loss=0.07775, over 4713.00 frames. ], tot_loss[loss=0.2172, simple_loss=0.2753, pruned_loss=0.07949, over 955072.63 frames. ], batch size: 59, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:41:12,779 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21911.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:21,737 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.201e+02 1.822e+02 2.098e+02 2.414e+02 4.975e+02, threshold=4.197e+02, percent-clipped=3.0 +2023-04-26 15:41:28,677 INFO [finetune.py:976] (3/7) Epoch 4, batch 4750, loss[loss=0.1851, simple_loss=0.2322, pruned_loss=0.06906, over 4376.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2735, pruned_loss=0.07912, over 951622.81 frames. ], batch size: 19, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:41:40,253 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21951.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:45,927 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 15:41:46,319 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9053, 1.2300, 3.2837, 3.0248, 2.9593, 3.1311, 3.1407, 2.8802], + device='cuda:3'), covar=tensor([0.7577, 0.5665, 0.1720, 0.2472, 0.1653, 0.2278, 0.1934, 0.2011], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0306, 0.0422, 0.0426, 0.0363, 0.0415, 0.0325, 0.0382], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:41:50,661 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21968.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:51,756 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21969.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:52,442 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:53,703 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2638, 1.4556, 1.4081, 1.6334, 1.5057, 1.6705, 1.5302, 1.5918], + device='cuda:3'), covar=tensor([0.8420, 1.3336, 1.1672, 0.9986, 1.1394, 1.8370, 1.3603, 1.1446], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0398, 0.0319, 0.0326, 0.0349, 0.0412, 0.0384, 0.0338], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 15:42:02,273 INFO [finetune.py:976] (3/7) Epoch 4, batch 4800, loss[loss=0.2324, simple_loss=0.2846, pruned_loss=0.09011, over 4812.00 frames. ], tot_loss[loss=0.2186, simple_loss=0.2762, pruned_loss=0.08049, over 950251.66 frames. ], batch size: 25, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:42:03,019 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21984.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:39,327 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22016.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:42,034 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-04-26 15:42:50,415 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.187e+02 1.864e+02 2.140e+02 2.595e+02 5.123e+02, threshold=4.279e+02, percent-clipped=1.0 +2023-04-26 15:42:50,820 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-26 15:42:54,228 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22031.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:55,829 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22032.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:56,372 INFO [finetune.py:976] (3/7) Epoch 4, batch 4850, loss[loss=0.1725, simple_loss=0.2433, pruned_loss=0.05082, over 4926.00 frames. ], tot_loss[loss=0.2201, simple_loss=0.2782, pruned_loss=0.08097, over 948757.94 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:43:45,554 INFO [finetune.py:976] (3/7) Epoch 4, batch 4900, loss[loss=0.2055, simple_loss=0.2824, pruned_loss=0.06431, over 4888.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2802, pruned_loss=0.08159, over 950351.39 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:44:09,604 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8108, 3.7155, 1.2481, 2.1061, 2.2158, 2.6889, 2.2676, 1.3872], + device='cuda:3'), covar=tensor([0.1238, 0.0785, 0.1899, 0.1200, 0.0980, 0.1030, 0.1424, 0.1774], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0260, 0.0146, 0.0127, 0.0137, 0.0159, 0.0124, 0.0127], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:44:15,730 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22119.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:44:19,288 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.258e+02 1.858e+02 2.268e+02 2.601e+02 5.071e+02, threshold=4.535e+02, percent-clipped=2.0 +2023-04-26 15:44:22,972 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:44:24,746 INFO [finetune.py:976] (3/7) Epoch 4, batch 4950, loss[loss=0.1834, simple_loss=0.258, pruned_loss=0.05439, over 4785.00 frames. ], tot_loss[loss=0.2204, simple_loss=0.2799, pruned_loss=0.08048, over 951326.50 frames. ], batch size: 29, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:44:53,086 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7783, 2.4322, 1.7645, 1.7960, 1.3905, 1.4859, 1.8226, 1.3733], + device='cuda:3'), covar=tensor([0.1603, 0.1599, 0.1752, 0.2087, 0.2792, 0.2061, 0.1263, 0.2178], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0222, 0.0181, 0.0211, 0.0220, 0.0189, 0.0175, 0.0199], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:44:53,665 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22176.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:44:58,358 INFO [finetune.py:976] (3/7) Epoch 4, batch 5000, loss[loss=0.2211, simple_loss=0.2752, pruned_loss=0.08351, over 4909.00 frames. ], tot_loss[loss=0.2194, simple_loss=0.2785, pruned_loss=0.08012, over 953534.82 frames. ], batch size: 36, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:45:04,335 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22191.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:15,020 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22206.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:20,502 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22215.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:31,709 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.348e+02 1.793e+02 1.980e+02 2.407e+02 3.890e+02, threshold=3.961e+02, percent-clipped=0.0 +2023-04-26 15:45:42,309 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.86 vs. limit=5.0 +2023-04-26 15:45:42,551 INFO [finetune.py:976] (3/7) Epoch 4, batch 5050, loss[loss=0.223, simple_loss=0.279, pruned_loss=0.0835, over 4812.00 frames. ], tot_loss[loss=0.2186, simple_loss=0.2769, pruned_loss=0.08012, over 954921.45 frames. ], batch size: 40, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:45:45,124 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:45,134 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:48,285 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-26 15:46:06,549 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22251.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:28,144 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22269.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:32,420 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22276.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:42,366 INFO [finetune.py:976] (3/7) Epoch 4, batch 5100, loss[loss=0.1777, simple_loss=0.2388, pruned_loss=0.05825, over 4811.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2725, pruned_loss=0.07834, over 952336.45 frames. ], batch size: 25, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:46:52,535 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22298.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:53,601 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22299.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:47:06,008 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22317.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:47:10,808 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.713e+02 2.102e+02 2.590e+02 4.893e+02, threshold=4.205e+02, percent-clipped=2.0 +2023-04-26 15:47:11,504 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22326.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:47:15,715 INFO [finetune.py:976] (3/7) Epoch 4, batch 5150, loss[loss=0.2882, simple_loss=0.3361, pruned_loss=0.1202, over 4902.00 frames. ], tot_loss[loss=0.2172, simple_loss=0.2743, pruned_loss=0.0801, over 951709.63 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:47:15,824 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1154, 1.5071, 1.3941, 1.7721, 1.5629, 1.9130, 1.4312, 3.3728], + device='cuda:3'), covar=tensor([0.0713, 0.0804, 0.0834, 0.1298, 0.0726, 0.0495, 0.0777, 0.0199], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0042, 0.0041, 0.0040, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 15:47:54,331 INFO [finetune.py:976] (3/7) Epoch 4, batch 5200, loss[loss=0.2643, simple_loss=0.3231, pruned_loss=0.1027, over 4850.00 frames. ], tot_loss[loss=0.2208, simple_loss=0.2778, pruned_loss=0.08194, over 949836.52 frames. ], batch size: 44, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:48:01,251 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22385.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:48:13,480 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 15:48:30,908 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22419.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:48:34,446 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.341e+02 1.984e+02 2.270e+02 2.624e+02 8.595e+02, threshold=4.540e+02, percent-clipped=4.0 +2023-04-26 15:48:39,405 INFO [finetune.py:976] (3/7) Epoch 4, batch 5250, loss[loss=0.1645, simple_loss=0.2258, pruned_loss=0.05158, over 4889.00 frames. ], tot_loss[loss=0.2213, simple_loss=0.279, pruned_loss=0.08181, over 948558.29 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:48:47,955 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22446.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:49:09,422 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:49:25,135 INFO [finetune.py:976] (3/7) Epoch 4, batch 5300, loss[loss=0.2116, simple_loss=0.2795, pruned_loss=0.07182, over 4834.00 frames. ], tot_loss[loss=0.2214, simple_loss=0.2798, pruned_loss=0.08151, over 950316.51 frames. ], batch size: 47, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:49:32,904 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22486.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:49:33,170 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-26 15:49:56,903 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:50:15,635 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.302e+02 1.836e+02 2.188e+02 2.701e+02 5.070e+02, threshold=4.376e+02, percent-clipped=2.0 +2023-04-26 15:50:19,981 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22532.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:50:20,501 INFO [finetune.py:976] (3/7) Epoch 4, batch 5350, loss[loss=0.2404, simple_loss=0.2972, pruned_loss=0.09179, over 4891.00 frames. ], tot_loss[loss=0.2204, simple_loss=0.2795, pruned_loss=0.08064, over 950225.29 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:50:33,852 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22554.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:50:46,634 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22571.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:50:53,879 INFO [finetune.py:976] (3/7) Epoch 4, batch 5400, loss[loss=0.2115, simple_loss=0.2739, pruned_loss=0.07453, over 4892.00 frames. ], tot_loss[loss=0.2169, simple_loss=0.2761, pruned_loss=0.07884, over 951556.15 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:50:59,987 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22593.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:51:38,543 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.846e+02 2.198e+02 2.575e+02 4.196e+02, threshold=4.395e+02, percent-clipped=1.0 +2023-04-26 15:51:39,240 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22626.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:51:47,649 INFO [finetune.py:976] (3/7) Epoch 4, batch 5450, loss[loss=0.186, simple_loss=0.2492, pruned_loss=0.06146, over 4898.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2738, pruned_loss=0.07838, over 952678.05 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:51:50,827 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6145, 3.6000, 1.1946, 1.7396, 1.9855, 2.4159, 1.9520, 1.0109], + device='cuda:3'), covar=tensor([0.1375, 0.0961, 0.2029, 0.1561, 0.1157, 0.1231, 0.1615, 0.2111], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0264, 0.0148, 0.0129, 0.0139, 0.0162, 0.0126, 0.0129], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:51:51,448 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22639.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:52:32,210 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-04-26 15:52:39,601 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22674.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:52:41,631 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-26 15:52:50,912 INFO [finetune.py:976] (3/7) Epoch 4, batch 5500, loss[loss=0.2318, simple_loss=0.2835, pruned_loss=0.09006, over 4741.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2713, pruned_loss=0.07718, over 955473.11 frames. ], batch size: 59, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:52:51,665 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6328, 1.5899, 0.6721, 1.3379, 1.9420, 1.5145, 1.4169, 1.4447], + device='cuda:3'), covar=tensor([0.0555, 0.0425, 0.0440, 0.0613, 0.0285, 0.0596, 0.0566, 0.0647], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 15:53:12,633 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22700.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 15:53:34,589 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.236e+02 2.045e+02 2.332e+02 2.799e+02 5.603e+02, threshold=4.665e+02, percent-clipped=2.0 +2023-04-26 15:53:36,767 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-26 15:53:40,509 INFO [finetune.py:976] (3/7) Epoch 4, batch 5550, loss[loss=0.2429, simple_loss=0.2855, pruned_loss=0.1001, over 4723.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2738, pruned_loss=0.07845, over 956915.66 frames. ], batch size: 59, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:53:45,457 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22741.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:05,988 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.55 vs. limit=5.0 +2023-04-26 15:54:11,071 INFO [finetune.py:976] (3/7) Epoch 4, batch 5600, loss[loss=0.2554, simple_loss=0.3194, pruned_loss=0.09572, over 4831.00 frames. ], tot_loss[loss=0.2204, simple_loss=0.2796, pruned_loss=0.08063, over 956025.16 frames. ], batch size: 47, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:54:12,895 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22786.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:18,054 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22794.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:37,297 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.226e+02 1.792e+02 2.075e+02 2.605e+02 4.713e+02, threshold=4.150e+02, percent-clipped=1.0 +2023-04-26 15:54:37,400 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22825.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:41,474 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22832.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:42,023 INFO [finetune.py:976] (3/7) Epoch 4, batch 5650, loss[loss=0.1972, simple_loss=0.2631, pruned_loss=0.06563, over 4712.00 frames. ], tot_loss[loss=0.2214, simple_loss=0.2816, pruned_loss=0.08058, over 954539.48 frames. ], batch size: 23, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:54:42,646 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22834.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:55,189 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22855.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:55:15,927 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22871.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:55:26,407 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 15:55:27,439 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22880.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:55:29,216 INFO [finetune.py:976] (3/7) Epoch 4, batch 5700, loss[loss=0.2257, simple_loss=0.2639, pruned_loss=0.09376, over 4166.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2766, pruned_loss=0.07958, over 936409.04 frames. ], batch size: 18, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:55:31,109 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22886.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:55:40,647 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22893.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:56:19,324 INFO [finetune.py:976] (3/7) Epoch 5, batch 0, loss[loss=0.1933, simple_loss=0.2738, pruned_loss=0.05641, over 4793.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2738, pruned_loss=0.05641, over 4793.00 frames. ], batch size: 29, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:56:19,324 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 15:56:29,354 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5133, 1.3116, 1.6996, 1.6324, 1.4001, 1.2329, 1.4837, 0.9052], + device='cuda:3'), covar=tensor([0.0734, 0.0925, 0.0818, 0.0858, 0.0864, 0.1454, 0.0692, 0.1143], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0077, 0.0075, 0.0069, 0.0080, 0.0096, 0.0083, 0.0078], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 15:56:30,076 INFO [finetune.py:1010] (3/7) Epoch 5, validation: loss=0.1632, simple_loss=0.2369, pruned_loss=0.04473, over 2265189.00 frames. +2023-04-26 15:56:30,077 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 15:56:35,030 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22919.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:56:38,535 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.609e+02 1.975e+02 2.369e+02 5.506e+02, threshold=3.950e+02, percent-clipped=1.0 +2023-04-26 15:56:48,362 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22941.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:56:50,260 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7635, 1.8076, 1.6588, 1.4336, 1.8893, 1.3773, 2.5350, 1.5201], + device='cuda:3'), covar=tensor([0.5399, 0.2137, 0.7194, 0.4066, 0.2465, 0.3642, 0.1788, 0.5152], + device='cuda:3'), in_proj_covar=tensor([0.0356, 0.0357, 0.0441, 0.0371, 0.0403, 0.0385, 0.0399, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:57:02,116 INFO [finetune.py:976] (3/7) Epoch 5, batch 50, loss[loss=0.2335, simple_loss=0.2906, pruned_loss=0.08823, over 4819.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2812, pruned_loss=0.08112, over 216266.24 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:57:30,164 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22995.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 15:57:52,330 INFO [finetune.py:976] (3/7) Epoch 5, batch 100, loss[loss=0.1829, simple_loss=0.243, pruned_loss=0.06146, over 4798.00 frames. ], tot_loss[loss=0.2156, simple_loss=0.2731, pruned_loss=0.07909, over 379844.78 frames. ], batch size: 45, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:58:02,669 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.276e+01 1.854e+02 2.138e+02 2.662e+02 6.636e+02, threshold=4.277e+02, percent-clipped=4.0 +2023-04-26 15:58:12,551 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23041.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:58:14,014 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.38 vs. limit=5.0 +2023-04-26 15:58:23,832 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7335, 1.9945, 1.8186, 1.9823, 1.6095, 1.6765, 1.8214, 1.4770], + device='cuda:3'), covar=tensor([0.1598, 0.1081, 0.0786, 0.0958, 0.2423, 0.1085, 0.1491, 0.1899], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0332, 0.0240, 0.0304, 0.0322, 0.0287, 0.0274, 0.0296], + device='cuda:3'), out_proj_covar=tensor([1.2721e-04, 1.3501e-04, 9.7905e-05, 1.2251e-04, 1.3299e-04, 1.1594e-04, + 1.1308e-04, 1.1954e-04], device='cuda:3') +2023-04-26 15:58:25,549 INFO [finetune.py:976] (3/7) Epoch 5, batch 150, loss[loss=0.2257, simple_loss=0.2732, pruned_loss=0.08913, over 4751.00 frames. ], tot_loss[loss=0.2113, simple_loss=0.2684, pruned_loss=0.07713, over 506859.74 frames. ], batch size: 23, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:58:44,620 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23089.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:58:57,342 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23108.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:58:59,552 INFO [finetune.py:976] (3/7) Epoch 5, batch 200, loss[loss=0.1577, simple_loss=0.2193, pruned_loss=0.04808, over 4776.00 frames. ], tot_loss[loss=0.2102, simple_loss=0.268, pruned_loss=0.07618, over 606686.45 frames. ], batch size: 28, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:59:10,064 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.278e+02 1.706e+02 2.083e+02 2.458e+02 7.322e+02, threshold=4.166e+02, percent-clipped=1.0 +2023-04-26 15:59:15,651 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1540, 1.2402, 1.3444, 1.4947, 1.5117, 1.1423, 0.9709, 1.2989], + device='cuda:3'), covar=tensor([0.1129, 0.1422, 0.0900, 0.0758, 0.0801, 0.1172, 0.1120, 0.0882], + device='cuda:3'), in_proj_covar=tensor([0.0210, 0.0210, 0.0187, 0.0182, 0.0183, 0.0199, 0.0171, 0.0194], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 15:59:25,401 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23150.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:59:29,852 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-26 15:59:33,031 INFO [finetune.py:976] (3/7) Epoch 5, batch 250, loss[loss=0.173, simple_loss=0.2329, pruned_loss=0.05652, over 4752.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2729, pruned_loss=0.07891, over 682377.27 frames. ], batch size: 23, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:59:38,561 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23169.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:59:47,773 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23181.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:00:00,762 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3948, 0.9767, 1.4044, 1.6966, 1.4829, 1.3170, 1.3910, 1.4309], + device='cuda:3'), covar=tensor([1.4890, 1.9785, 2.2858, 2.8513, 1.7347, 2.3832, 2.4894, 1.8919], + device='cuda:3'), in_proj_covar=tensor([0.0427, 0.0466, 0.0555, 0.0575, 0.0459, 0.0484, 0.0496, 0.0498], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:00:03,129 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5336, 1.6919, 1.7016, 1.9085, 1.7660, 1.9273, 1.6180, 3.2938], + device='cuda:3'), covar=tensor([0.0567, 0.0652, 0.0700, 0.1023, 0.0550, 0.0720, 0.0685, 0.0182], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0040, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 16:00:06,066 INFO [finetune.py:976] (3/7) Epoch 5, batch 300, loss[loss=0.2183, simple_loss=0.2859, pruned_loss=0.0753, over 4872.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2776, pruned_loss=0.08094, over 743439.14 frames. ], batch size: 34, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:00:10,743 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23217.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:00:14,404 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0804, 2.5054, 0.8328, 1.3705, 1.4563, 1.8177, 1.5729, 0.9216], + device='cuda:3'), covar=tensor([0.1519, 0.1123, 0.1876, 0.1539, 0.1251, 0.1032, 0.1568, 0.1724], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0265, 0.0149, 0.0130, 0.0140, 0.0162, 0.0126, 0.0129], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:00:15,563 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.275e+02 1.925e+02 2.300e+02 2.815e+02 4.609e+02, threshold=4.600e+02, percent-clipped=3.0 +2023-04-26 16:00:18,575 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23229.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:00:34,117 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6963, 1.5485, 1.8375, 2.0791, 1.9233, 1.6041, 1.7049, 1.7622], + device='cuda:3'), covar=tensor([1.4717, 1.8400, 2.3432, 2.2979, 1.5848, 2.7170, 2.6361, 2.0369], + device='cuda:3'), in_proj_covar=tensor([0.0426, 0.0465, 0.0554, 0.0574, 0.0458, 0.0483, 0.0495, 0.0498], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:00:46,946 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 16:00:55,702 INFO [finetune.py:976] (3/7) Epoch 5, batch 350, loss[loss=0.2012, simple_loss=0.2612, pruned_loss=0.07057, over 4838.00 frames. ], tot_loss[loss=0.2216, simple_loss=0.2801, pruned_loss=0.08151, over 792257.16 frames. ], batch size: 47, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:01:18,185 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23278.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:01:31,063 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2149, 1.4894, 1.4637, 1.8659, 1.6831, 1.7455, 1.4706, 3.1217], + device='cuda:3'), covar=tensor([0.0663, 0.0802, 0.0782, 0.1162, 0.0649, 0.0512, 0.0717, 0.0213], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0040, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 16:01:31,701 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23290.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:01:40,985 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23295.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:01:55,884 INFO [finetune.py:976] (3/7) Epoch 5, batch 400, loss[loss=0.2275, simple_loss=0.2919, pruned_loss=0.08161, over 4914.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.2807, pruned_loss=0.08116, over 828431.22 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:02:05,328 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.686e+02 2.112e+02 2.575e+02 6.256e+02, threshold=4.223e+02, percent-clipped=1.0 +2023-04-26 16:02:07,177 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6928, 2.0045, 1.7417, 1.8833, 1.5700, 1.5737, 1.7097, 1.5068], + device='cuda:3'), covar=tensor([0.2117, 0.1377, 0.1051, 0.1469, 0.3218, 0.1417, 0.2000, 0.2471], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0331, 0.0241, 0.0305, 0.0324, 0.0287, 0.0274, 0.0296], + device='cuda:3'), out_proj_covar=tensor([1.2722e-04, 1.3467e-04, 9.8027e-05, 1.2283e-04, 1.3352e-04, 1.1572e-04, + 1.1292e-04, 1.1961e-04], device='cuda:3') +2023-04-26 16:02:18,849 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23343.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:02:29,822 INFO [finetune.py:976] (3/7) Epoch 5, batch 450, loss[loss=0.2086, simple_loss=0.2703, pruned_loss=0.07348, over 4756.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2788, pruned_loss=0.08038, over 857497.11 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:02:39,770 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 16:02:49,265 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4895, 0.9958, 1.4663, 1.7907, 1.5716, 1.3752, 1.4544, 1.5584], + device='cuda:3'), covar=tensor([1.5052, 1.9808, 2.3458, 2.9792, 1.7950, 2.6684, 2.5429, 1.9619], + device='cuda:3'), in_proj_covar=tensor([0.0425, 0.0464, 0.0553, 0.0572, 0.0457, 0.0481, 0.0493, 0.0496], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:03:12,684 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-26 16:03:14,682 INFO [finetune.py:976] (3/7) Epoch 5, batch 500, loss[loss=0.222, simple_loss=0.2715, pruned_loss=0.08625, over 4830.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.2756, pruned_loss=0.07932, over 878665.60 frames. ], batch size: 41, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:03:29,846 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.283e+02 1.686e+02 2.056e+02 2.772e+02 5.396e+02, threshold=4.112e+02, percent-clipped=3.0 +2023-04-26 16:03:47,540 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23450.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:03:54,138 INFO [finetune.py:976] (3/7) Epoch 5, batch 550, loss[loss=0.2003, simple_loss=0.2609, pruned_loss=0.06985, over 4892.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.271, pruned_loss=0.07733, over 895405.23 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:03:56,063 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23464.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:07,414 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23481.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:09,299 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7659, 2.1834, 1.0704, 1.6084, 2.2248, 1.7375, 1.6505, 1.7532], + device='cuda:3'), covar=tensor([0.0569, 0.0374, 0.0366, 0.0573, 0.0269, 0.0550, 0.0576, 0.0608], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 16:04:12,226 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23488.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:19,707 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23498.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:27,585 INFO [finetune.py:976] (3/7) Epoch 5, batch 600, loss[loss=0.2013, simple_loss=0.2607, pruned_loss=0.07091, over 4911.00 frames. ], tot_loss[loss=0.2115, simple_loss=0.2703, pruned_loss=0.07636, over 909516.50 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:04:32,585 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6106, 1.1666, 1.2972, 1.2033, 1.7652, 1.4532, 1.0544, 1.2181], + device='cuda:3'), covar=tensor([0.1607, 0.1468, 0.1989, 0.1426, 0.0820, 0.1505, 0.1989, 0.1900], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0332, 0.0348, 0.0305, 0.0343, 0.0344, 0.0312, 0.0351], + device='cuda:3'), out_proj_covar=tensor([6.7297e-05, 7.1079e-05, 7.5404e-05, 6.3846e-05, 7.2725e-05, 7.4917e-05, + 6.7584e-05, 7.5722e-05], device='cuda:3') +2023-04-26 16:04:36,131 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.343e+02 1.963e+02 2.277e+02 2.707e+02 6.010e+02, threshold=4.553e+02, percent-clipped=1.0 +2023-04-26 16:04:39,605 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23529.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:53,882 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23549.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 16:05:01,028 INFO [finetune.py:976] (3/7) Epoch 5, batch 650, loss[loss=0.2058, simple_loss=0.2603, pruned_loss=0.07561, over 4771.00 frames. ], tot_loss[loss=0.2144, simple_loss=0.2736, pruned_loss=0.07762, over 916038.64 frames. ], batch size: 27, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:05:08,398 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23573.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:05:16,708 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23585.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:05:29,209 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5651, 1.6097, 0.7352, 1.2619, 1.7769, 1.4750, 1.3752, 1.4183], + device='cuda:3'), covar=tensor([0.0534, 0.0402, 0.0433, 0.0605, 0.0311, 0.0561, 0.0517, 0.0623], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 16:05:32,191 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3048, 3.0035, 0.9651, 1.5723, 2.3396, 1.5032, 4.2038, 2.0483], + device='cuda:3'), covar=tensor([0.0723, 0.0886, 0.1045, 0.1361, 0.0543, 0.1057, 0.0235, 0.0666], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0071, 0.0053, 0.0049, 0.0054, 0.0055, 0.0083, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 16:05:34,423 INFO [finetune.py:976] (3/7) Epoch 5, batch 700, loss[loss=0.2323, simple_loss=0.2852, pruned_loss=0.08971, over 4870.00 frames. ], tot_loss[loss=0.2174, simple_loss=0.2767, pruned_loss=0.07906, over 924608.16 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:05:42,880 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.928e+02 2.468e+02 2.939e+02 6.493e+02, threshold=4.936e+02, percent-clipped=4.0 +2023-04-26 16:06:15,515 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4879, 1.2144, 4.2423, 3.9784, 3.7656, 3.9870, 3.9231, 3.6785], + device='cuda:3'), covar=tensor([0.6575, 0.5859, 0.0921, 0.1535, 0.1121, 0.1482, 0.1602, 0.1509], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0307, 0.0421, 0.0427, 0.0363, 0.0416, 0.0324, 0.0383], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:06:19,708 INFO [finetune.py:976] (3/7) Epoch 5, batch 750, loss[loss=0.2353, simple_loss=0.2968, pruned_loss=0.08691, over 4844.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2774, pruned_loss=0.07908, over 930986.76 frames. ], batch size: 44, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:07:12,391 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 16:07:13,809 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7830, 1.3298, 1.4393, 1.3369, 1.9915, 1.6762, 1.1913, 1.3906], + device='cuda:3'), covar=tensor([0.1893, 0.1663, 0.2048, 0.1489, 0.0875, 0.1626, 0.3096, 0.2457], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0335, 0.0351, 0.0308, 0.0345, 0.0347, 0.0313, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.7724e-05, 7.1822e-05, 7.6146e-05, 6.4464e-05, 7.3066e-05, 7.5598e-05, + 6.7886e-05, 7.6390e-05], device='cuda:3') +2023-04-26 16:07:26,345 INFO [finetune.py:976] (3/7) Epoch 5, batch 800, loss[loss=0.2083, simple_loss=0.2738, pruned_loss=0.0714, over 4896.00 frames. ], tot_loss[loss=0.2185, simple_loss=0.2784, pruned_loss=0.07928, over 938052.14 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:07:34,822 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.218e+02 1.744e+02 2.077e+02 2.568e+02 5.488e+02, threshold=4.154e+02, percent-clipped=2.0 +2023-04-26 16:07:35,531 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0074, 1.1834, 5.0957, 4.7550, 4.5459, 4.7804, 4.4967, 4.4233], + device='cuda:3'), covar=tensor([0.6441, 0.6425, 0.0956, 0.1686, 0.0920, 0.1228, 0.1386, 0.1565], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0306, 0.0421, 0.0427, 0.0362, 0.0415, 0.0323, 0.0381], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:07:39,546 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-26 16:08:00,117 INFO [finetune.py:976] (3/7) Epoch 5, batch 850, loss[loss=0.1915, simple_loss=0.2642, pruned_loss=0.05934, over 4773.00 frames. ], tot_loss[loss=0.2161, simple_loss=0.2761, pruned_loss=0.078, over 943161.53 frames. ], batch size: 26, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:08:02,022 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23764.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:08:39,280 INFO [finetune.py:976] (3/7) Epoch 5, batch 900, loss[loss=0.2283, simple_loss=0.2916, pruned_loss=0.08243, over 4727.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2744, pruned_loss=0.07764, over 947050.33 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:08:40,386 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23812.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:08:50,421 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9189, 2.8143, 2.3257, 3.3127, 2.9027, 2.9374, 1.1025, 2.8330], + device='cuda:3'), covar=tensor([0.1860, 0.1597, 0.2862, 0.2691, 0.2898, 0.2185, 0.5350, 0.2538], + device='cuda:3'), in_proj_covar=tensor([0.0248, 0.0222, 0.0260, 0.0315, 0.0309, 0.0259, 0.0277, 0.0278], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 16:08:54,020 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.680e+02 2.075e+02 2.492e+02 8.869e+02, threshold=4.150e+02, percent-clipped=5.0 +2023-04-26 16:09:23,842 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23844.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:09:46,929 INFO [finetune.py:976] (3/7) Epoch 5, batch 950, loss[loss=0.1867, simple_loss=0.2442, pruned_loss=0.06456, over 4038.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2723, pruned_loss=0.07665, over 950529.29 frames. ], batch size: 65, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:09:56,932 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23873.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:10:14,841 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23885.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:10:44,366 INFO [finetune.py:976] (3/7) Epoch 5, batch 1000, loss[loss=0.2002, simple_loss=0.2632, pruned_loss=0.06858, over 4821.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2741, pruned_loss=0.07713, over 952443.32 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:10:52,028 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23921.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:10:54,330 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.295e+02 1.932e+02 2.275e+02 2.774e+02 5.327e+02, threshold=4.551e+02, percent-clipped=3.0 +2023-04-26 16:10:59,243 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23933.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:11:17,666 INFO [finetune.py:976] (3/7) Epoch 5, batch 1050, loss[loss=0.1988, simple_loss=0.2618, pruned_loss=0.06788, over 4757.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2747, pruned_loss=0.07656, over 951675.82 frames. ], batch size: 26, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:11:26,089 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1409, 2.6351, 1.0160, 1.5332, 2.0770, 1.2780, 3.7615, 2.1818], + device='cuda:3'), covar=tensor([0.0704, 0.0733, 0.0913, 0.1425, 0.0614, 0.1134, 0.0286, 0.0625], + device='cuda:3'), in_proj_covar=tensor([0.0055, 0.0071, 0.0053, 0.0050, 0.0055, 0.0056, 0.0084, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 16:12:02,464 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6081, 2.1194, 1.4932, 1.2857, 1.1762, 1.2165, 1.5093, 1.1213], + device='cuda:3'), covar=tensor([0.2155, 0.1505, 0.2142, 0.2342, 0.3168, 0.2480, 0.1549, 0.2611], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0221, 0.0180, 0.0209, 0.0217, 0.0189, 0.0173, 0.0197], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:12:31,173 INFO [finetune.py:976] (3/7) Epoch 5, batch 1100, loss[loss=0.2322, simple_loss=0.2917, pruned_loss=0.0864, over 4913.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2755, pruned_loss=0.07681, over 950426.88 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:12:45,513 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.175e+02 1.787e+02 2.210e+02 2.620e+02 5.269e+02, threshold=4.419e+02, percent-clipped=3.0 +2023-04-26 16:12:53,367 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24037.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:12:56,959 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9058, 1.4389, 5.3336, 4.9081, 4.5952, 5.0463, 4.6168, 4.6773], + device='cuda:3'), covar=tensor([0.6489, 0.5870, 0.0933, 0.1881, 0.1130, 0.1008, 0.1035, 0.1544], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0306, 0.0421, 0.0430, 0.0364, 0.0416, 0.0325, 0.0384], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:13:03,522 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9078, 1.1746, 3.2799, 2.9946, 2.9793, 3.1815, 3.1898, 2.9022], + device='cuda:3'), covar=tensor([0.7239, 0.5293, 0.1342, 0.2258, 0.1418, 0.1674, 0.1657, 0.1786], + device='cuda:3'), in_proj_covar=tensor([0.0319, 0.0306, 0.0421, 0.0430, 0.0364, 0.0416, 0.0325, 0.0385], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:13:04,981 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-04-26 16:13:08,675 INFO [finetune.py:976] (3/7) Epoch 5, batch 1150, loss[loss=0.1993, simple_loss=0.2586, pruned_loss=0.06998, over 4807.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2774, pruned_loss=0.07785, over 951759.72 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:13:33,483 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24098.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:13:37,342 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 16:13:42,163 INFO [finetune.py:976] (3/7) Epoch 5, batch 1200, loss[loss=0.246, simple_loss=0.3004, pruned_loss=0.09581, over 4873.00 frames. ], tot_loss[loss=0.2158, simple_loss=0.2761, pruned_loss=0.07772, over 951195.00 frames. ], batch size: 34, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:13:52,169 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.779e+02 2.090e+02 2.399e+02 4.332e+02, threshold=4.179e+02, percent-clipped=0.0 +2023-04-26 16:14:04,710 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24144.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 16:14:15,789 INFO [finetune.py:976] (3/7) Epoch 5, batch 1250, loss[loss=0.1995, simple_loss=0.26, pruned_loss=0.06954, over 4780.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2735, pruned_loss=0.07693, over 952326.42 frames. ], batch size: 28, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:14:18,929 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1049, 2.1379, 1.8886, 1.7495, 2.3908, 1.7410, 2.7973, 1.7477], + device='cuda:3'), covar=tensor([0.4290, 0.1932, 0.4652, 0.3757, 0.1706, 0.2947, 0.1446, 0.4134], + device='cuda:3'), in_proj_covar=tensor([0.0350, 0.0356, 0.0438, 0.0370, 0.0399, 0.0384, 0.0396, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:14:42,784 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24192.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:14:51,500 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.12 vs. limit=5.0 +2023-04-26 16:15:00,605 INFO [finetune.py:976] (3/7) Epoch 5, batch 1300, loss[loss=0.1877, simple_loss=0.246, pruned_loss=0.06468, over 4792.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.2698, pruned_loss=0.07503, over 953020.27 frames. ], batch size: 51, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:15:10,272 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 16:15:10,677 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.317e+02 1.846e+02 2.149e+02 2.713e+02 6.103e+02, threshold=4.299e+02, percent-clipped=1.0 +2023-04-26 16:15:41,943 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5466, 1.1034, 1.2743, 1.2115, 1.7501, 1.3790, 1.1105, 1.2712], + device='cuda:3'), covar=tensor([0.1753, 0.1612, 0.2109, 0.1384, 0.0866, 0.1535, 0.2362, 0.2327], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0336, 0.0352, 0.0310, 0.0348, 0.0348, 0.0313, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.8139e-05, 7.2225e-05, 7.6353e-05, 6.4871e-05, 7.3765e-05, 7.5792e-05, + 6.7899e-05, 7.6456e-05], device='cuda:3') +2023-04-26 16:15:49,985 INFO [finetune.py:976] (3/7) Epoch 5, batch 1350, loss[loss=0.2153, simple_loss=0.2684, pruned_loss=0.08105, over 4720.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2693, pruned_loss=0.07465, over 952773.02 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:16:00,985 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-04-26 16:16:24,900 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2376, 2.9839, 1.2681, 1.5332, 2.1326, 1.5134, 3.5368, 1.9038], + device='cuda:3'), covar=tensor([0.0625, 0.0798, 0.0791, 0.1067, 0.0446, 0.0848, 0.0194, 0.0565], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0071, 0.0053, 0.0050, 0.0054, 0.0055, 0.0083, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 16:16:29,538 INFO [finetune.py:976] (3/7) Epoch 5, batch 1400, loss[loss=0.2362, simple_loss=0.3023, pruned_loss=0.08506, over 4874.00 frames. ], tot_loss[loss=0.2124, simple_loss=0.2723, pruned_loss=0.07618, over 951251.04 frames. ], batch size: 34, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:16:38,505 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.222e+02 1.836e+02 2.129e+02 2.470e+02 6.262e+02, threshold=4.259e+02, percent-clipped=1.0 +2023-04-26 16:17:09,954 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-26 16:17:19,153 INFO [finetune.py:976] (3/7) Epoch 5, batch 1450, loss[loss=0.2312, simple_loss=0.2887, pruned_loss=0.08687, over 4828.00 frames. ], tot_loss[loss=0.2152, simple_loss=0.2752, pruned_loss=0.07757, over 952171.91 frames. ], batch size: 30, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:17:29,453 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7170, 2.3770, 1.7435, 1.6569, 1.3096, 1.3954, 1.8557, 1.2812], + device='cuda:3'), covar=tensor([0.1858, 0.1637, 0.1840, 0.2237, 0.2801, 0.2288, 0.1261, 0.2409], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0221, 0.0180, 0.0209, 0.0217, 0.0189, 0.0172, 0.0196], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:18:00,682 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24393.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:18:13,898 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1207, 1.4694, 5.3404, 5.0206, 4.6054, 5.0934, 4.6576, 4.6641], + device='cuda:3'), covar=tensor([0.6696, 0.5942, 0.0888, 0.1515, 0.0945, 0.1245, 0.1268, 0.1419], + device='cuda:3'), in_proj_covar=tensor([0.0321, 0.0308, 0.0426, 0.0432, 0.0366, 0.0418, 0.0326, 0.0387], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:18:16,238 INFO [finetune.py:976] (3/7) Epoch 5, batch 1500, loss[loss=0.2236, simple_loss=0.2829, pruned_loss=0.08214, over 4771.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2779, pruned_loss=0.0788, over 953809.31 frames. ], batch size: 26, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:18:25,729 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.847e+02 2.193e+02 2.526e+02 4.286e+02, threshold=4.386e+02, percent-clipped=1.0 +2023-04-26 16:18:49,530 INFO [finetune.py:976] (3/7) Epoch 5, batch 1550, loss[loss=0.2189, simple_loss=0.2855, pruned_loss=0.07616, over 4769.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2785, pruned_loss=0.07893, over 954325.34 frames. ], batch size: 26, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:19:00,416 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3800, 1.2863, 1.7070, 1.6592, 1.4014, 1.1120, 1.5190, 1.0512], + device='cuda:3'), covar=tensor([0.0921, 0.0853, 0.0541, 0.0741, 0.0933, 0.1229, 0.0716, 0.0974], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0078, 0.0076, 0.0070, 0.0081, 0.0097, 0.0084, 0.0079], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:3') +2023-04-26 16:19:22,930 INFO [finetune.py:976] (3/7) Epoch 5, batch 1600, loss[loss=0.2101, simple_loss=0.2706, pruned_loss=0.07483, over 4819.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2758, pruned_loss=0.07741, over 954023.05 frames. ], batch size: 41, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:19:32,011 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.836e+02 2.156e+02 2.625e+02 3.904e+02, threshold=4.311e+02, percent-clipped=0.0 +2023-04-26 16:19:48,182 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24547.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:19:56,589 INFO [finetune.py:976] (3/7) Epoch 5, batch 1650, loss[loss=0.216, simple_loss=0.2764, pruned_loss=0.07782, over 4701.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.2732, pruned_loss=0.07685, over 955913.94 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:20:00,579 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 16:20:01,617 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24569.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:20:28,448 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24608.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:20:30,138 INFO [finetune.py:976] (3/7) Epoch 5, batch 1700, loss[loss=0.2493, simple_loss=0.3085, pruned_loss=0.09504, over 4831.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2705, pruned_loss=0.07575, over 955200.83 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:20:38,586 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.758e+02 2.083e+02 2.406e+02 5.348e+02, threshold=4.165e+02, percent-clipped=1.0 +2023-04-26 16:20:42,274 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24630.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:20:42,344 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 16:21:03,215 INFO [finetune.py:976] (3/7) Epoch 5, batch 1750, loss[loss=0.3351, simple_loss=0.3665, pruned_loss=0.1519, over 4768.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2739, pruned_loss=0.07791, over 956111.04 frames. ], batch size: 59, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:21:35,457 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24693.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:21:40,138 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-26 16:21:48,679 INFO [finetune.py:976] (3/7) Epoch 5, batch 1800, loss[loss=0.2542, simple_loss=0.3131, pruned_loss=0.09763, over 4862.00 frames. ], tot_loss[loss=0.2173, simple_loss=0.2769, pruned_loss=0.07881, over 957626.12 frames. ], batch size: 34, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:21:57,278 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.363e+02 2.030e+02 2.431e+02 2.911e+02 5.485e+02, threshold=4.863e+02, percent-clipped=5.0 +2023-04-26 16:22:08,099 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24741.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:22:27,731 INFO [finetune.py:976] (3/7) Epoch 5, batch 1850, loss[loss=0.2058, simple_loss=0.2808, pruned_loss=0.06534, over 4818.00 frames. ], tot_loss[loss=0.2197, simple_loss=0.2794, pruned_loss=0.08002, over 955922.30 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:22:27,970 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-04-26 16:22:30,287 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:23:29,635 INFO [finetune.py:976] (3/7) Epoch 5, batch 1900, loss[loss=0.219, simple_loss=0.2815, pruned_loss=0.07823, over 4860.00 frames. ], tot_loss[loss=0.22, simple_loss=0.2801, pruned_loss=0.07994, over 955280.12 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:23:44,193 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.736e+02 2.141e+02 2.516e+02 4.890e+02, threshold=4.282e+02, percent-clipped=1.0 +2023-04-26 16:23:50,606 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24826.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:24:16,077 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:24:31,092 INFO [finetune.py:976] (3/7) Epoch 5, batch 1950, loss[loss=0.2269, simple_loss=0.2835, pruned_loss=0.08514, over 4925.00 frames. ], tot_loss[loss=0.2175, simple_loss=0.2779, pruned_loss=0.07855, over 956089.89 frames. ], batch size: 38, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:24:46,907 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6385, 1.1878, 1.3200, 1.3328, 1.8826, 1.4751, 1.1825, 1.2721], + device='cuda:3'), covar=tensor([0.1800, 0.1740, 0.2375, 0.1652, 0.0905, 0.1699, 0.2191, 0.2096], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0334, 0.0351, 0.0309, 0.0345, 0.0345, 0.0309, 0.0352], + device='cuda:3'), out_proj_covar=tensor([6.7407e-05, 7.1537e-05, 7.6067e-05, 6.4719e-05, 7.3084e-05, 7.5060e-05, + 6.7077e-05, 7.5809e-05], device='cuda:3') +2023-04-26 16:24:58,650 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24903.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:25:03,322 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24909.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:25:03,972 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7705, 2.3994, 1.7590, 1.5813, 1.4184, 1.4182, 1.8145, 1.3511], + device='cuda:3'), covar=tensor([0.2079, 0.1750, 0.2078, 0.2431, 0.3102, 0.2447, 0.1472, 0.2485], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0224, 0.0181, 0.0211, 0.0219, 0.0191, 0.0173, 0.0198], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:25:04,923 INFO [finetune.py:976] (3/7) Epoch 5, batch 2000, loss[loss=0.1822, simple_loss=0.2529, pruned_loss=0.05577, over 4792.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2748, pruned_loss=0.07717, over 955829.79 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:25:13,846 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.797e+02 2.078e+02 2.506e+02 3.857e+02, threshold=4.156e+02, percent-clipped=0.0 +2023-04-26 16:25:13,921 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24925.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:25:37,718 INFO [finetune.py:976] (3/7) Epoch 5, batch 2050, loss[loss=0.2186, simple_loss=0.2698, pruned_loss=0.08367, over 4757.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2714, pruned_loss=0.07594, over 955774.36 frames. ], batch size: 26, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:25:46,069 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7424, 1.9121, 0.9616, 1.4555, 2.1671, 1.6523, 1.6054, 1.5412], + device='cuda:3'), covar=tensor([0.0526, 0.0360, 0.0358, 0.0551, 0.0255, 0.0542, 0.0505, 0.0588], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 16:25:56,113 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 16:25:59,892 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24995.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:26:10,620 INFO [finetune.py:976] (3/7) Epoch 5, batch 2100, loss[loss=0.1927, simple_loss=0.2503, pruned_loss=0.06751, over 4926.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2705, pruned_loss=0.07589, over 955364.23 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:26:21,075 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.851e+02 2.091e+02 2.608e+02 4.715e+02, threshold=4.183e+02, percent-clipped=2.0 +2023-04-26 16:26:22,428 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8681, 1.7337, 1.5097, 1.3654, 1.7289, 1.4877, 2.0886, 1.2970], + device='cuda:3'), covar=tensor([0.3541, 0.1464, 0.4256, 0.2786, 0.1576, 0.2186, 0.1557, 0.4228], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0353, 0.0433, 0.0364, 0.0394, 0.0381, 0.0393, 0.0414], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:26:26,470 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 16:26:40,918 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25056.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 16:26:43,818 INFO [finetune.py:976] (3/7) Epoch 5, batch 2150, loss[loss=0.2891, simple_loss=0.3315, pruned_loss=0.1233, over 4825.00 frames. ], tot_loss[loss=0.2136, simple_loss=0.273, pruned_loss=0.07703, over 954760.98 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:26:55,946 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-26 16:26:58,313 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-04-26 16:27:17,137 INFO [finetune.py:976] (3/7) Epoch 5, batch 2200, loss[loss=0.2717, simple_loss=0.3246, pruned_loss=0.1094, over 4858.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2757, pruned_loss=0.0774, over 955354.40 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:27:24,268 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25121.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:27:27,076 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.776e+02 2.171e+02 2.634e+02 4.750e+02, threshold=4.343e+02, percent-clipped=2.0 +2023-04-26 16:27:45,756 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6150, 1.6570, 1.7040, 1.8462, 1.6354, 1.8757, 1.8036, 1.7681], + device='cuda:3'), covar=tensor([0.6384, 1.2256, 1.0463, 0.9460, 1.1156, 1.5736, 1.2400, 1.1364], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0399, 0.0318, 0.0328, 0.0350, 0.0414, 0.0381, 0.0337], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 16:27:47,563 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8738, 1.7362, 1.9764, 2.3403, 2.3227, 2.0095, 1.5112, 1.9469], + device='cuda:3'), covar=tensor([0.1083, 0.1123, 0.0692, 0.0551, 0.0628, 0.0913, 0.0996, 0.0629], + device='cuda:3'), in_proj_covar=tensor([0.0206, 0.0207, 0.0183, 0.0180, 0.0181, 0.0195, 0.0167, 0.0192], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:27:50,269 INFO [finetune.py:976] (3/7) Epoch 5, batch 2250, loss[loss=0.2075, simple_loss=0.2866, pruned_loss=0.06426, over 4848.00 frames. ], tot_loss[loss=0.2164, simple_loss=0.2773, pruned_loss=0.07779, over 956379.96 frames. ], batch size: 49, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:28:06,298 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-26 16:28:26,332 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25203.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:28:26,891 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25204.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:28:32,139 INFO [finetune.py:976] (3/7) Epoch 5, batch 2300, loss[loss=0.195, simple_loss=0.252, pruned_loss=0.06894, over 4787.00 frames. ], tot_loss[loss=0.2176, simple_loss=0.2786, pruned_loss=0.07825, over 954297.36 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:28:52,192 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.823e+02 2.210e+02 2.598e+02 6.851e+02, threshold=4.421e+02, percent-clipped=2.0 +2023-04-26 16:28:52,293 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25225.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:29:26,418 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25251.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:29:38,267 INFO [finetune.py:976] (3/7) Epoch 5, batch 2350, loss[loss=0.1911, simple_loss=0.2509, pruned_loss=0.06568, over 4796.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2751, pruned_loss=0.07664, over 952813.67 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:29:57,348 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25273.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:30:39,274 INFO [finetune.py:976] (3/7) Epoch 5, batch 2400, loss[loss=0.203, simple_loss=0.2669, pruned_loss=0.06952, over 4895.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.272, pruned_loss=0.07607, over 954456.62 frames. ], batch size: 32, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:30:48,375 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.795e+02 2.098e+02 2.522e+02 5.509e+02, threshold=4.195e+02, percent-clipped=3.0 +2023-04-26 16:31:12,378 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25351.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:31:18,437 INFO [finetune.py:976] (3/7) Epoch 5, batch 2450, loss[loss=0.243, simple_loss=0.2922, pruned_loss=0.09689, over 4791.00 frames. ], tot_loss[loss=0.2104, simple_loss=0.2694, pruned_loss=0.07572, over 953354.38 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:31:26,407 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9167, 2.6702, 1.9644, 1.8404, 1.4603, 1.4680, 2.0752, 1.4449], + device='cuda:3'), covar=tensor([0.1880, 0.1843, 0.1801, 0.2323, 0.2899, 0.2163, 0.1304, 0.2362], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0222, 0.0180, 0.0209, 0.0217, 0.0189, 0.0172, 0.0196], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:31:36,980 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7327, 1.8367, 1.8133, 1.4689, 1.9187, 1.5876, 2.6331, 1.5536], + device='cuda:3'), covar=tensor([0.4433, 0.1940, 0.5280, 0.3280, 0.1905, 0.2431, 0.1439, 0.4695], + device='cuda:3'), in_proj_covar=tensor([0.0349, 0.0354, 0.0435, 0.0367, 0.0396, 0.0382, 0.0395, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:32:08,176 INFO [finetune.py:976] (3/7) Epoch 5, batch 2500, loss[loss=0.177, simple_loss=0.2534, pruned_loss=0.0503, over 4906.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2688, pruned_loss=0.07498, over 954003.26 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:32:15,261 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25421.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:32:17,547 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.862e+02 2.343e+02 2.826e+02 5.817e+02, threshold=4.685e+02, percent-clipped=1.0 +2023-04-26 16:32:19,540 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25428.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:32:41,938 INFO [finetune.py:976] (3/7) Epoch 5, batch 2550, loss[loss=0.1823, simple_loss=0.2423, pruned_loss=0.06112, over 4723.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2727, pruned_loss=0.07648, over 953836.02 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:32:47,380 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25469.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:33:01,455 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25489.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 16:33:11,437 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25504.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:33:15,641 INFO [finetune.py:976] (3/7) Epoch 5, batch 2600, loss[loss=0.2428, simple_loss=0.3034, pruned_loss=0.09114, over 4824.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2751, pruned_loss=0.07742, over 953668.50 frames. ], batch size: 38, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:33:25,305 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.303e+02 1.830e+02 2.137e+02 2.517e+02 4.934e+02, threshold=4.274e+02, percent-clipped=1.0 +2023-04-26 16:33:30,801 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3092, 1.0927, 1.5761, 1.4595, 1.2114, 1.0106, 1.2006, 0.7928], + device='cuda:3'), covar=tensor([0.0773, 0.0939, 0.0552, 0.0755, 0.0972, 0.1507, 0.0814, 0.1025], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0077, 0.0075, 0.0069, 0.0081, 0.0096, 0.0083, 0.0079], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:33:43,927 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25552.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:33:48,856 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2952, 1.6076, 1.5973, 2.1258, 1.8092, 2.0296, 1.5432, 4.3646], + device='cuda:3'), covar=tensor([0.0692, 0.0864, 0.0886, 0.1267, 0.0727, 0.0594, 0.0874, 0.0158], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0041, 0.0040, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 16:33:49,359 INFO [finetune.py:976] (3/7) Epoch 5, batch 2650, loss[loss=0.2135, simple_loss=0.2807, pruned_loss=0.07313, over 4867.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.2771, pruned_loss=0.07818, over 954516.25 frames. ], batch size: 34, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:34:28,770 INFO [finetune.py:976] (3/7) Epoch 5, batch 2700, loss[loss=0.227, simple_loss=0.301, pruned_loss=0.07652, over 4813.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2771, pruned_loss=0.07797, over 954835.44 frames. ], batch size: 41, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:34:29,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4337, 3.2713, 0.8242, 1.8469, 1.9531, 2.2797, 1.8761, 1.0191], + device='cuda:3'), covar=tensor([0.1360, 0.0900, 0.2121, 0.1243, 0.0973, 0.1074, 0.1514, 0.1975], + device='cuda:3'), in_proj_covar=tensor([0.0123, 0.0266, 0.0148, 0.0130, 0.0139, 0.0162, 0.0127, 0.0129], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:34:48,615 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.304e+02 1.757e+02 2.118e+02 2.619e+02 4.731e+02, threshold=4.237e+02, percent-clipped=2.0 +2023-04-26 16:35:00,108 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4334, 1.2413, 4.0580, 3.7999, 3.6451, 3.8346, 3.8410, 3.6095], + device='cuda:3'), covar=tensor([0.6814, 0.5586, 0.1104, 0.1597, 0.1055, 0.1489, 0.1584, 0.1492], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0304, 0.0417, 0.0423, 0.0357, 0.0408, 0.0321, 0.0377], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:35:20,266 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25651.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:35:37,126 INFO [finetune.py:976] (3/7) Epoch 5, batch 2750, loss[loss=0.2756, simple_loss=0.3186, pruned_loss=0.1163, over 4906.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2744, pruned_loss=0.0773, over 955081.68 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:35:49,520 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6421, 1.6476, 1.8101, 2.0715, 2.0497, 1.6569, 1.4166, 1.7964], + device='cuda:3'), covar=tensor([0.0968, 0.1277, 0.0734, 0.0690, 0.0635, 0.1024, 0.0909, 0.0663], + device='cuda:3'), in_proj_covar=tensor([0.0208, 0.0210, 0.0185, 0.0182, 0.0182, 0.0198, 0.0169, 0.0194], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:36:24,084 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25699.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:36:27,669 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1854, 2.5474, 1.0773, 1.3183, 2.0559, 1.2677, 3.3046, 1.7296], + device='cuda:3'), covar=tensor([0.0665, 0.0627, 0.0838, 0.1348, 0.0551, 0.1016, 0.0248, 0.0669], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0071, 0.0052, 0.0049, 0.0054, 0.0055, 0.0082, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 16:36:32,332 INFO [finetune.py:976] (3/7) Epoch 5, batch 2800, loss[loss=0.1987, simple_loss=0.2613, pruned_loss=0.06803, over 4820.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2707, pruned_loss=0.07537, over 957876.75 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:36:39,168 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3319, 2.4420, 2.8049, 2.8453, 2.6619, 2.2902, 1.8625, 2.3821], + device='cuda:3'), covar=tensor([0.1150, 0.0980, 0.0511, 0.0711, 0.0759, 0.1134, 0.1114, 0.0765], + device='cuda:3'), in_proj_covar=tensor([0.0208, 0.0210, 0.0185, 0.0182, 0.0182, 0.0198, 0.0169, 0.0194], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:36:46,287 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0774, 1.4459, 1.9534, 2.2669, 1.8256, 1.4600, 1.1183, 1.6296], + device='cuda:3'), covar=tensor([0.4276, 0.4782, 0.2043, 0.3561, 0.3879, 0.3476, 0.6060, 0.3460], + device='cuda:3'), in_proj_covar=tensor([0.0274, 0.0259, 0.0219, 0.0328, 0.0219, 0.0228, 0.0244, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 16:36:47,384 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.821e+02 2.084e+02 2.532e+02 5.696e+02, threshold=4.167e+02, percent-clipped=4.0 +2023-04-26 16:37:23,259 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 16:37:32,910 INFO [finetune.py:976] (3/7) Epoch 5, batch 2850, loss[loss=0.166, simple_loss=0.2312, pruned_loss=0.05044, over 4719.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2701, pruned_loss=0.07525, over 958482.80 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:37:46,999 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25784.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:38:05,601 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0825, 1.4241, 1.9714, 2.3916, 1.8476, 1.4242, 1.2052, 1.7190], + device='cuda:3'), covar=tensor([0.4271, 0.4768, 0.2068, 0.3338, 0.4127, 0.3643, 0.5795, 0.3520], + device='cuda:3'), in_proj_covar=tensor([0.0274, 0.0259, 0.0220, 0.0329, 0.0219, 0.0229, 0.0244, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 16:38:06,695 INFO [finetune.py:976] (3/7) Epoch 5, batch 2900, loss[loss=0.2472, simple_loss=0.3091, pruned_loss=0.0927, over 4915.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2726, pruned_loss=0.07653, over 958714.11 frames. ], batch size: 43, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:38:15,790 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.320e+02 1.753e+02 2.187e+02 2.584e+02 6.163e+02, threshold=4.374e+02, percent-clipped=2.0 +2023-04-26 16:38:27,336 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 16:38:37,783 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25858.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:38:39,948 INFO [finetune.py:976] (3/7) Epoch 5, batch 2950, loss[loss=0.2336, simple_loss=0.2947, pruned_loss=0.08624, over 4802.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2748, pruned_loss=0.07674, over 957503.00 frames. ], batch size: 51, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:39:12,518 INFO [finetune.py:976] (3/7) Epoch 5, batch 3000, loss[loss=0.2041, simple_loss=0.2496, pruned_loss=0.07933, over 4731.00 frames. ], tot_loss[loss=0.2162, simple_loss=0.2769, pruned_loss=0.07775, over 957207.15 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:39:12,518 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 16:39:21,768 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7207, 1.8529, 1.9730, 2.2155, 2.1591, 1.7903, 1.3957, 1.9595], + device='cuda:3'), covar=tensor([0.0963, 0.1098, 0.0698, 0.0602, 0.0619, 0.0972, 0.0949, 0.0585], + device='cuda:3'), in_proj_covar=tensor([0.0208, 0.0210, 0.0185, 0.0182, 0.0182, 0.0198, 0.0169, 0.0194], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:39:29,056 INFO [finetune.py:1010] (3/7) Epoch 5, validation: loss=0.1595, simple_loss=0.233, pruned_loss=0.04303, over 2265189.00 frames. +2023-04-26 16:39:29,057 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 16:39:41,223 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25919.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:39:50,621 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0373, 2.5013, 0.8111, 1.4691, 1.5042, 1.9168, 1.5790, 0.8563], + device='cuda:3'), covar=tensor([0.1528, 0.1105, 0.1853, 0.1364, 0.1118, 0.0911, 0.1573, 0.1752], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0263, 0.0147, 0.0129, 0.0138, 0.0161, 0.0126, 0.0129], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:39:51,733 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.259e+02 1.795e+02 2.151e+02 2.692e+02 4.010e+02, threshold=4.303e+02, percent-clipped=0.0 +2023-04-26 16:40:35,903 INFO [finetune.py:976] (3/7) Epoch 5, batch 3050, loss[loss=0.226, simple_loss=0.285, pruned_loss=0.08352, over 4859.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2775, pruned_loss=0.0772, over 956289.60 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:41:32,850 INFO [finetune.py:976] (3/7) Epoch 5, batch 3100, loss[loss=0.1979, simple_loss=0.2642, pruned_loss=0.06582, over 4828.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2743, pruned_loss=0.07574, over 958327.73 frames. ], batch size: 33, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:41:43,947 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.748e+02 1.976e+02 2.307e+02 6.292e+02, threshold=3.952e+02, percent-clipped=1.0 +2023-04-26 16:42:06,248 INFO [finetune.py:976] (3/7) Epoch 5, batch 3150, loss[loss=0.1802, simple_loss=0.2323, pruned_loss=0.06408, over 4721.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.2705, pruned_loss=0.07463, over 957540.85 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:42:27,117 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26084.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 16:42:46,460 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.2756, 1.3301, 1.3952, 1.0101, 1.3291, 1.1455, 1.6533, 1.2272], + device='cuda:3'), covar=tensor([0.4017, 0.1848, 0.5119, 0.2754, 0.1667, 0.2436, 0.1857, 0.4990], + device='cuda:3'), in_proj_covar=tensor([0.0353, 0.0357, 0.0437, 0.0371, 0.0400, 0.0387, 0.0398, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:42:46,463 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26102.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:42:54,361 INFO [finetune.py:976] (3/7) Epoch 5, batch 3200, loss[loss=0.2177, simple_loss=0.285, pruned_loss=0.07515, over 4903.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2676, pruned_loss=0.07345, over 957516.92 frames. ], batch size: 43, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:43:07,555 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3844, 2.3105, 1.8696, 2.0735, 2.1783, 1.9488, 3.0390, 1.4876], + device='cuda:3'), covar=tensor([0.3616, 0.1568, 0.4412, 0.3276, 0.2098, 0.2496, 0.2034, 0.4722], + device='cuda:3'), in_proj_covar=tensor([0.0353, 0.0357, 0.0437, 0.0371, 0.0400, 0.0386, 0.0397, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:43:09,265 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.183e+02 1.750e+02 2.081e+02 2.590e+02 4.901e+02, threshold=4.161e+02, percent-clipped=2.0 +2023-04-26 16:43:14,416 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26132.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:43:31,951 INFO [finetune.py:976] (3/7) Epoch 5, batch 3250, loss[loss=0.1783, simple_loss=0.2531, pruned_loss=0.05174, over 4897.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2675, pruned_loss=0.07343, over 957485.66 frames. ], batch size: 35, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:43:33,263 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26163.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:43:37,344 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26168.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:44:05,486 INFO [finetune.py:976] (3/7) Epoch 5, batch 3300, loss[loss=0.2133, simple_loss=0.2771, pruned_loss=0.07473, over 4182.00 frames. ], tot_loss[loss=0.2104, simple_loss=0.2706, pruned_loss=0.07505, over 953689.20 frames. ], batch size: 65, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:44:07,397 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26214.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 16:44:10,361 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8793, 1.8270, 2.0467, 2.2870, 2.2912, 1.8469, 1.4411, 1.8251], + device='cuda:3'), covar=tensor([0.1097, 0.1183, 0.0715, 0.0659, 0.0705, 0.1075, 0.1059, 0.0770], + device='cuda:3'), in_proj_covar=tensor([0.0207, 0.0209, 0.0185, 0.0181, 0.0181, 0.0197, 0.0168, 0.0193], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:44:16,104 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.343e+02 1.793e+02 2.255e+02 2.677e+02 5.857e+02, threshold=4.510e+02, percent-clipped=2.0 +2023-04-26 16:44:17,434 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26229.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:44:32,051 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26250.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:44:38,662 INFO [finetune.py:976] (3/7) Epoch 5, batch 3350, loss[loss=0.1891, simple_loss=0.2615, pruned_loss=0.05836, over 4775.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2718, pruned_loss=0.07527, over 951985.60 frames. ], batch size: 54, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:45:31,118 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-26 16:45:39,811 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 16:45:50,830 INFO [finetune.py:976] (3/7) Epoch 5, batch 3400, loss[loss=0.2114, simple_loss=0.2704, pruned_loss=0.07622, over 4882.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2739, pruned_loss=0.07589, over 951276.54 frames. ], batch size: 32, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:45:50,976 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26311.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:46:13,228 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.738e+02 2.082e+02 2.437e+02 3.720e+02, threshold=4.164e+02, percent-clipped=0.0 +2023-04-26 16:46:56,945 INFO [finetune.py:976] (3/7) Epoch 5, batch 3450, loss[loss=0.2319, simple_loss=0.2925, pruned_loss=0.08559, over 4815.00 frames. ], tot_loss[loss=0.2135, simple_loss=0.2747, pruned_loss=0.07614, over 950892.84 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:47:18,761 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 16:47:52,441 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2667, 2.6328, 0.9870, 1.4017, 2.0876, 1.3203, 3.5865, 1.9091], + device='cuda:3'), covar=tensor([0.0642, 0.0608, 0.0849, 0.1328, 0.0558, 0.1018, 0.0217, 0.0621], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0071, 0.0053, 0.0049, 0.0054, 0.0055, 0.0083, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 16:48:03,178 INFO [finetune.py:976] (3/7) Epoch 5, batch 3500, loss[loss=0.1931, simple_loss=0.2583, pruned_loss=0.06392, over 4824.00 frames. ], tot_loss[loss=0.2116, simple_loss=0.2726, pruned_loss=0.07535, over 952061.37 frames. ], batch size: 39, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:48:05,084 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26414.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:48:25,047 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.307e+02 1.713e+02 2.095e+02 2.561e+02 6.592e+02, threshold=4.191e+02, percent-clipped=2.0 +2023-04-26 16:48:28,705 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5187, 3.5389, 1.1128, 1.8262, 2.1096, 2.3977, 2.0538, 0.9827], + device='cuda:3'), covar=tensor([0.1358, 0.1051, 0.1832, 0.1405, 0.0957, 0.1102, 0.1501, 0.2112], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0261, 0.0146, 0.0128, 0.0137, 0.0159, 0.0125, 0.0128], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 16:49:08,071 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-26 16:49:08,576 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:49:10,351 INFO [finetune.py:976] (3/7) Epoch 5, batch 3550, loss[loss=0.1746, simple_loss=0.2402, pruned_loss=0.05457, over 4765.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2692, pruned_loss=0.07438, over 953786.47 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:49:24,634 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26475.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:49:49,838 INFO [finetune.py:976] (3/7) Epoch 5, batch 3600, loss[loss=0.204, simple_loss=0.2601, pruned_loss=0.07395, over 4766.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2666, pruned_loss=0.07336, over 955679.72 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:49:51,765 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26514.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:49:57,838 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26524.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:49:59,593 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.194e+02 1.718e+02 2.043e+02 2.547e+02 4.174e+02, threshold=4.086e+02, percent-clipped=0.0 +2023-04-26 16:50:05,098 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.91 vs. limit=5.0 +2023-04-26 16:50:21,006 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 16:50:23,342 INFO [finetune.py:976] (3/7) Epoch 5, batch 3650, loss[loss=0.2519, simple_loss=0.3171, pruned_loss=0.09335, over 4754.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2689, pruned_loss=0.07469, over 953919.18 frames. ], batch size: 54, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:50:24,021 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26562.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:50:53,158 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26606.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:50:57,093 INFO [finetune.py:976] (3/7) Epoch 5, batch 3700, loss[loss=0.1956, simple_loss=0.2669, pruned_loss=0.06221, over 4860.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.2738, pruned_loss=0.07641, over 954867.39 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:51:06,780 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.827e+02 2.264e+02 2.659e+02 6.887e+02, threshold=4.529e+02, percent-clipped=2.0 +2023-04-26 16:51:29,842 INFO [finetune.py:976] (3/7) Epoch 5, batch 3750, loss[loss=0.1934, simple_loss=0.2486, pruned_loss=0.06909, over 4707.00 frames. ], tot_loss[loss=0.2147, simple_loss=0.2754, pruned_loss=0.077, over 955331.31 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:51:47,309 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 16:52:13,189 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-04-26 16:52:20,492 INFO [finetune.py:976] (3/7) Epoch 5, batch 3800, loss[loss=0.2196, simple_loss=0.2845, pruned_loss=0.07735, over 4799.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2773, pruned_loss=0.07779, over 954128.78 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:52:31,669 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.252e+02 1.856e+02 2.180e+02 2.581e+02 5.516e+02, threshold=4.361e+02, percent-clipped=1.0 +2023-04-26 16:52:52,597 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26758.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:52:54,307 INFO [finetune.py:976] (3/7) Epoch 5, batch 3850, loss[loss=0.2185, simple_loss=0.2833, pruned_loss=0.07681, over 4812.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2749, pruned_loss=0.07625, over 955426.82 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:53:00,825 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26770.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:53:40,974 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26806.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:53:44,435 INFO [finetune.py:976] (3/7) Epoch 5, batch 3900, loss[loss=0.1963, simple_loss=0.2494, pruned_loss=0.0716, over 4800.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2706, pruned_loss=0.07466, over 955144.47 frames. ], batch size: 45, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:54:03,742 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26824.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:54:05,469 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.758e+02 2.108e+02 2.612e+02 6.373e+02, threshold=4.215e+02, percent-clipped=3.0 +2023-04-26 16:54:50,051 INFO [finetune.py:976] (3/7) Epoch 5, batch 3950, loss[loss=0.2405, simple_loss=0.3021, pruned_loss=0.0894, over 4929.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2674, pruned_loss=0.07331, over 954662.54 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:55:08,488 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26872.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:55:47,037 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26904.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:55:48,226 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2054, 1.5532, 1.4603, 1.7720, 1.6479, 1.9155, 1.4212, 3.1527], + device='cuda:3'), covar=tensor([0.0715, 0.0818, 0.0832, 0.1257, 0.0700, 0.0504, 0.0763, 0.0184], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0041, 0.0039, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 16:55:48,232 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26906.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:55:51,740 INFO [finetune.py:976] (3/7) Epoch 5, batch 4000, loss[loss=0.2299, simple_loss=0.2849, pruned_loss=0.08745, over 4265.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.267, pruned_loss=0.07366, over 952522.43 frames. ], batch size: 65, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:56:02,921 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.248e+02 1.675e+02 2.014e+02 2.454e+02 6.687e+02, threshold=4.029e+02, percent-clipped=1.0 +2023-04-26 16:56:17,505 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26950.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:56:19,909 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26954.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:56:25,065 INFO [finetune.py:976] (3/7) Epoch 5, batch 4050, loss[loss=0.1841, simple_loss=0.2338, pruned_loss=0.06714, over 4087.00 frames. ], tot_loss[loss=0.2102, simple_loss=0.2704, pruned_loss=0.075, over 953260.92 frames. ], batch size: 17, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:56:25,783 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9895, 1.8376, 2.0517, 2.3333, 2.2657, 1.7424, 1.4683, 1.9333], + device='cuda:3'), covar=tensor([0.0906, 0.1060, 0.0566, 0.0601, 0.0627, 0.0987, 0.0981, 0.0647], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0205, 0.0181, 0.0177, 0.0176, 0.0193, 0.0164, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 16:56:28,122 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26965.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 16:56:58,980 INFO [finetune.py:976] (3/7) Epoch 5, batch 4100, loss[loss=0.2522, simple_loss=0.3138, pruned_loss=0.09534, over 4812.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2746, pruned_loss=0.07657, over 952092.91 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:56:59,106 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27011.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:57:20,555 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.859e+02 2.211e+02 2.694e+02 5.784e+02, threshold=4.421e+02, percent-clipped=3.0 +2023-04-26 16:58:04,488 INFO [finetune.py:976] (3/7) Epoch 5, batch 4150, loss[loss=0.2544, simple_loss=0.3251, pruned_loss=0.09182, over 4817.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2757, pruned_loss=0.07705, over 952968.66 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:58:16,079 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27070.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:58:47,721 INFO [finetune.py:976] (3/7) Epoch 5, batch 4200, loss[loss=0.2069, simple_loss=0.2705, pruned_loss=0.07161, over 4789.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.275, pruned_loss=0.0758, over 953234.18 frames. ], batch size: 25, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:58:53,014 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27118.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:58:58,900 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.181e+01 1.721e+02 2.038e+02 2.503e+02 4.432e+02, threshold=4.077e+02, percent-clipped=1.0 +2023-04-26 16:59:06,580 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7583, 1.1043, 1.2400, 1.3786, 1.3802, 1.5752, 1.3532, 1.3005], + device='cuda:3'), covar=tensor([0.5819, 0.8624, 0.7293, 0.7245, 0.9128, 1.3688, 0.8764, 0.8368], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0396, 0.0318, 0.0327, 0.0348, 0.0412, 0.0379, 0.0336], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 16:59:13,622 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27148.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:59:21,419 INFO [finetune.py:976] (3/7) Epoch 5, batch 4250, loss[loss=0.1892, simple_loss=0.2522, pruned_loss=0.06307, over 4906.00 frames. ], tot_loss[loss=0.2116, simple_loss=0.2728, pruned_loss=0.07519, over 951381.56 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:59:54,266 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27209.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:59:55,345 INFO [finetune.py:976] (3/7) Epoch 5, batch 4300, loss[loss=0.1808, simple_loss=0.242, pruned_loss=0.05977, over 4910.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2693, pruned_loss=0.07408, over 951343.50 frames. ], batch size: 46, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:00:17,882 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.798e+02 2.267e+02 2.763e+02 5.468e+02, threshold=4.535e+02, percent-clipped=5.0 +2023-04-26 17:01:04,154 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27260.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:01:04,678 INFO [finetune.py:976] (3/7) Epoch 5, batch 4350, loss[loss=0.2127, simple_loss=0.2616, pruned_loss=0.08195, over 4821.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2658, pruned_loss=0.0728, over 952191.21 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:01:46,637 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27306.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:01:49,558 INFO [finetune.py:976] (3/7) Epoch 5, batch 4400, loss[loss=0.1965, simple_loss=0.2707, pruned_loss=0.0611, over 4803.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2676, pruned_loss=0.07398, over 953562.06 frames. ], batch size: 25, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:02:00,136 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.703e+02 2.157e+02 2.468e+02 4.465e+02, threshold=4.314e+02, percent-clipped=0.0 +2023-04-26 17:02:04,588 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 17:02:09,677 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2402, 1.4528, 1.4505, 2.0598, 2.3541, 1.8579, 1.7322, 1.5249], + device='cuda:3'), covar=tensor([0.2043, 0.2290, 0.2621, 0.1567, 0.1258, 0.2422, 0.3243, 0.2356], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0332, 0.0348, 0.0306, 0.0341, 0.0338, 0.0306, 0.0349], + device='cuda:3'), out_proj_covar=tensor([6.6590e-05, 7.1028e-05, 7.5400e-05, 6.3800e-05, 7.2068e-05, 7.3393e-05, + 6.6450e-05, 7.5099e-05], device='cuda:3') +2023-04-26 17:02:23,093 INFO [finetune.py:976] (3/7) Epoch 5, batch 4450, loss[loss=0.2439, simple_loss=0.3154, pruned_loss=0.08621, over 4856.00 frames. ], tot_loss[loss=0.2124, simple_loss=0.2728, pruned_loss=0.07602, over 953565.90 frames. ], batch size: 44, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:02:33,440 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1533, 1.3776, 1.4572, 1.6568, 1.5508, 1.6937, 1.5353, 1.5681], + device='cuda:3'), covar=tensor([0.6787, 1.0499, 0.8537, 0.7628, 0.9285, 1.3467, 1.0848, 0.9113], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0398, 0.0318, 0.0327, 0.0348, 0.0413, 0.0380, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:02:56,797 INFO [finetune.py:976] (3/7) Epoch 5, batch 4500, loss[loss=0.258, simple_loss=0.3064, pruned_loss=0.1048, over 4230.00 frames. ], tot_loss[loss=0.215, simple_loss=0.2753, pruned_loss=0.07737, over 952737.47 frames. ], batch size: 65, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:03:17,002 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5001, 1.7833, 1.3217, 0.9915, 1.2322, 1.1963, 1.3208, 1.1535], + device='cuda:3'), covar=tensor([0.1815, 0.1498, 0.1904, 0.2032, 0.2690, 0.2239, 0.1296, 0.2266], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0220, 0.0178, 0.0208, 0.0215, 0.0188, 0.0171, 0.0195], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 17:03:17,443 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.259e+02 1.778e+02 2.173e+02 2.607e+02 4.915e+02, threshold=4.346e+02, percent-clipped=1.0 +2023-04-26 17:03:48,457 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6829, 1.2978, 1.2317, 1.3493, 1.9345, 1.5547, 1.2324, 1.1979], + device='cuda:3'), covar=tensor([0.1848, 0.1612, 0.2164, 0.1537, 0.0898, 0.1523, 0.2320, 0.2088], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0335, 0.0351, 0.0309, 0.0343, 0.0340, 0.0308, 0.0351], + device='cuda:3'), out_proj_covar=tensor([6.6983e-05, 7.1670e-05, 7.5915e-05, 6.4480e-05, 7.2689e-05, 7.3784e-05, + 6.6826e-05, 7.5677e-05], device='cuda:3') +2023-04-26 17:04:02,849 INFO [finetune.py:976] (3/7) Epoch 5, batch 4550, loss[loss=0.2307, simple_loss=0.2923, pruned_loss=0.08461, over 4907.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.2765, pruned_loss=0.07722, over 955440.92 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:04:12,349 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-04-26 17:04:31,478 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27504.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:04:36,155 INFO [finetune.py:976] (3/7) Epoch 5, batch 4600, loss[loss=0.1752, simple_loss=0.2468, pruned_loss=0.05178, over 4788.00 frames. ], tot_loss[loss=0.2141, simple_loss=0.2748, pruned_loss=0.07667, over 952935.57 frames. ], batch size: 29, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:04:46,279 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.803e+02 2.105e+02 2.477e+02 5.679e+02, threshold=4.210e+02, percent-clipped=3.0 +2023-04-26 17:05:08,931 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27560.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:05:09,450 INFO [finetune.py:976] (3/7) Epoch 5, batch 4650, loss[loss=0.1835, simple_loss=0.2486, pruned_loss=0.05922, over 4773.00 frames. ], tot_loss[loss=0.2113, simple_loss=0.2717, pruned_loss=0.07546, over 953735.23 frames. ], batch size: 26, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:05:29,222 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6184, 1.8488, 1.7527, 1.9535, 1.8206, 2.0242, 1.8279, 1.8325], + device='cuda:3'), covar=tensor([0.7585, 1.1848, 1.0433, 0.8777, 1.0255, 1.4521, 1.2699, 1.1179], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0401, 0.0320, 0.0329, 0.0351, 0.0415, 0.0382, 0.0337], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:05:40,075 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27606.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:05:41,303 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27608.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:05:43,535 INFO [finetune.py:976] (3/7) Epoch 5, batch 4700, loss[loss=0.1441, simple_loss=0.2027, pruned_loss=0.04275, over 4028.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2686, pruned_loss=0.07439, over 953675.32 frames. ], batch size: 17, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:05:46,698 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0631, 2.4730, 2.0272, 2.4327, 1.7368, 2.1948, 2.2715, 1.7634], + device='cuda:3'), covar=tensor([0.2086, 0.1567, 0.1113, 0.1386, 0.3378, 0.1523, 0.1869, 0.2805], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0325, 0.0236, 0.0300, 0.0319, 0.0279, 0.0268, 0.0292], + device='cuda:3'), out_proj_covar=tensor([1.2495e-04, 1.3194e-04, 9.5916e-05, 1.2050e-04, 1.3120e-04, 1.1298e-04, + 1.1021e-04, 1.1793e-04], device='cuda:3') +2023-04-26 17:05:54,153 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.191e+02 1.680e+02 2.058e+02 2.502e+02 5.893e+02, threshold=4.117e+02, percent-clipped=4.0 +2023-04-26 17:06:01,005 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2076, 3.1250, 2.4771, 2.7150, 2.2869, 2.6141, 2.6650, 2.1344], + device='cuda:3'), covar=tensor([0.2688, 0.1314, 0.0963, 0.1654, 0.3059, 0.1456, 0.2254, 0.3259], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0324, 0.0236, 0.0300, 0.0319, 0.0279, 0.0268, 0.0292], + device='cuda:3'), out_proj_covar=tensor([1.2499e-04, 1.3178e-04, 9.5917e-05, 1.2045e-04, 1.3123e-04, 1.1281e-04, + 1.1006e-04, 1.1786e-04], device='cuda:3') +2023-04-26 17:06:23,332 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:06:34,207 INFO [finetune.py:976] (3/7) Epoch 5, batch 4750, loss[loss=0.1638, simple_loss=0.2181, pruned_loss=0.0548, over 4279.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.266, pruned_loss=0.07323, over 955240.77 frames. ], batch size: 18, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:06:35,412 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9433, 1.8062, 2.0373, 2.4297, 2.3446, 1.9236, 1.6160, 2.0674], + device='cuda:3'), covar=tensor([0.1113, 0.1130, 0.0696, 0.0654, 0.0680, 0.1092, 0.0967, 0.0681], + device='cuda:3'), in_proj_covar=tensor([0.0205, 0.0209, 0.0183, 0.0180, 0.0179, 0.0196, 0.0167, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:06:54,205 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4707, 0.8955, 1.3184, 1.8963, 1.6293, 1.4072, 1.3789, 1.4641], + device='cuda:3'), covar=tensor([0.9338, 1.2321, 1.3398, 1.3571, 1.0385, 1.5055, 1.4595, 1.2342], + device='cuda:3'), in_proj_covar=tensor([0.0418, 0.0452, 0.0537, 0.0554, 0.0447, 0.0469, 0.0483, 0.0483], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:07:40,899 INFO [finetune.py:976] (3/7) Epoch 5, batch 4800, loss[loss=0.2073, simple_loss=0.2933, pruned_loss=0.06063, over 4826.00 frames. ], tot_loss[loss=0.2095, simple_loss=0.2696, pruned_loss=0.07469, over 955029.13 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:07:48,584 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27714.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:08:02,771 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.360e+02 1.880e+02 2.179e+02 2.503e+02 4.628e+02, threshold=4.358e+02, percent-clipped=2.0 +2023-04-26 17:08:33,008 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5927, 1.7017, 0.7752, 1.3414, 1.8471, 1.4651, 1.3829, 1.4298], + device='cuda:3'), covar=tensor([0.0549, 0.0393, 0.0425, 0.0579, 0.0298, 0.0555, 0.0539, 0.0628], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0030, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 17:08:41,285 INFO [finetune.py:976] (3/7) Epoch 5, batch 4850, loss[loss=0.2679, simple_loss=0.328, pruned_loss=0.1039, over 4750.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.274, pruned_loss=0.0758, over 955388.04 frames. ], batch size: 59, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:08:50,920 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27775.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:09:14,394 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27804.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:09:18,625 INFO [finetune.py:976] (3/7) Epoch 5, batch 4900, loss[loss=0.2719, simple_loss=0.3292, pruned_loss=0.1073, over 4879.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.276, pruned_loss=0.07659, over 953790.80 frames. ], batch size: 35, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:09:18,723 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0078, 2.4303, 0.8882, 1.3528, 1.6927, 1.2882, 2.9693, 1.6390], + device='cuda:3'), covar=tensor([0.0676, 0.0618, 0.0795, 0.1201, 0.0537, 0.0971, 0.0260, 0.0619], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0054, 0.0054, 0.0082, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 17:09:30,227 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.253e+02 1.856e+02 2.309e+02 2.750e+02 8.138e+02, threshold=4.618e+02, percent-clipped=6.0 +2023-04-26 17:09:46,106 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27852.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:09:52,541 INFO [finetune.py:976] (3/7) Epoch 5, batch 4950, loss[loss=0.2218, simple_loss=0.2935, pruned_loss=0.0751, over 4777.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2767, pruned_loss=0.07659, over 953565.46 frames. ], batch size: 28, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:09:55,471 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 17:10:25,997 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8379, 1.6543, 2.1096, 2.2145, 2.0040, 1.7416, 1.9214, 1.9669], + device='cuda:3'), covar=tensor([1.0704, 1.4151, 1.4921, 1.5872, 1.1598, 1.9216, 1.8409, 1.4673], + device='cuda:3'), in_proj_covar=tensor([0.0420, 0.0454, 0.0539, 0.0559, 0.0449, 0.0472, 0.0484, 0.0485], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:10:31,676 INFO [finetune.py:976] (3/7) Epoch 5, batch 5000, loss[loss=0.2493, simple_loss=0.3213, pruned_loss=0.08865, over 4810.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2742, pruned_loss=0.07562, over 953971.98 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:10:42,356 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.656e+02 1.985e+02 2.433e+02 4.076e+02, threshold=3.970e+02, percent-clipped=0.0 +2023-04-26 17:10:52,561 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 17:10:55,502 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2589, 1.5067, 1.5600, 1.7569, 1.6064, 1.8149, 1.6633, 1.6187], + device='cuda:3'), covar=tensor([0.7490, 1.0584, 0.8856, 0.7584, 0.9560, 1.2725, 1.0878, 0.9676], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0397, 0.0319, 0.0328, 0.0349, 0.0414, 0.0379, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:11:04,343 INFO [finetune.py:976] (3/7) Epoch 5, batch 5050, loss[loss=0.217, simple_loss=0.2735, pruned_loss=0.08026, over 4892.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2714, pruned_loss=0.07511, over 955297.03 frames. ], batch size: 35, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:12:05,263 INFO [finetune.py:976] (3/7) Epoch 5, batch 5100, loss[loss=0.1731, simple_loss=0.2397, pruned_loss=0.05321, over 4797.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2681, pruned_loss=0.07348, over 956954.14 frames. ], batch size: 45, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:12:27,654 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.777e+02 2.079e+02 2.348e+02 6.496e+02, threshold=4.158e+02, percent-clipped=5.0 +2023-04-26 17:12:48,105 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4094, 2.9876, 1.0220, 1.4751, 2.0470, 1.3958, 4.0477, 1.9381], + device='cuda:3'), covar=tensor([0.0643, 0.0812, 0.0904, 0.1313, 0.0578, 0.1037, 0.0219, 0.0640], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0054, 0.0054, 0.0082, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 17:13:12,933 INFO [finetune.py:976] (3/7) Epoch 5, batch 5150, loss[loss=0.183, simple_loss=0.25, pruned_loss=0.058, over 4749.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2684, pruned_loss=0.07399, over 956412.45 frames. ], batch size: 27, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:13:31,486 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28070.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:14:09,050 INFO [finetune.py:976] (3/7) Epoch 5, batch 5200, loss[loss=0.2246, simple_loss=0.2848, pruned_loss=0.08221, over 4916.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2709, pruned_loss=0.07457, over 956898.73 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:14:19,886 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.281e+02 1.900e+02 2.150e+02 2.501e+02 5.102e+02, threshold=4.301e+02, percent-clipped=1.0 +2023-04-26 17:14:20,016 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8752, 2.4347, 2.0660, 2.3855, 1.6197, 2.1489, 2.0944, 1.7689], + device='cuda:3'), covar=tensor([0.2272, 0.1254, 0.0950, 0.1254, 0.3434, 0.1153, 0.1973, 0.2536], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0324, 0.0236, 0.0299, 0.0318, 0.0279, 0.0268, 0.0290], + device='cuda:3'), out_proj_covar=tensor([1.2423e-04, 1.3155e-04, 9.5911e-05, 1.1991e-04, 1.3103e-04, 1.1270e-04, + 1.1032e-04, 1.1710e-04], device='cuda:3') +2023-04-26 17:14:38,267 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7691, 1.9843, 1.9349, 1.5510, 2.0743, 1.7172, 2.6493, 1.5972], + device='cuda:3'), covar=tensor([0.3953, 0.1731, 0.4520, 0.3153, 0.1576, 0.2475, 0.1274, 0.4587], + device='cuda:3'), in_proj_covar=tensor([0.0352, 0.0355, 0.0439, 0.0369, 0.0398, 0.0387, 0.0394, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:14:42,338 INFO [finetune.py:976] (3/7) Epoch 5, batch 5250, loss[loss=0.2142, simple_loss=0.269, pruned_loss=0.07966, over 4862.00 frames. ], tot_loss[loss=0.2123, simple_loss=0.2729, pruned_loss=0.07589, over 954912.98 frames. ], batch size: 31, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:15:16,323 INFO [finetune.py:976] (3/7) Epoch 5, batch 5300, loss[loss=0.225, simple_loss=0.2905, pruned_loss=0.07971, over 4920.00 frames. ], tot_loss[loss=0.2136, simple_loss=0.2747, pruned_loss=0.0763, over 955591.87 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:15:27,005 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.419e+02 1.781e+02 2.085e+02 2.383e+02 5.799e+02, threshold=4.171e+02, percent-clipped=2.0 +2023-04-26 17:15:28,326 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28229.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:15:46,193 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28255.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:15:49,729 INFO [finetune.py:976] (3/7) Epoch 5, batch 5350, loss[loss=0.2641, simple_loss=0.3029, pruned_loss=0.1126, over 4061.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2743, pruned_loss=0.07539, over 951136.57 frames. ], batch size: 65, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:16:07,723 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28287.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:16:10,016 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28290.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:16:23,680 INFO [finetune.py:976] (3/7) Epoch 5, batch 5400, loss[loss=0.1511, simple_loss=0.2236, pruned_loss=0.03933, over 4888.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.271, pruned_loss=0.07379, over 951699.87 frames. ], batch size: 32, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:16:27,327 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28316.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 17:16:34,348 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.191e+02 1.689e+02 2.005e+02 2.420e+02 4.760e+02, threshold=4.009e+02, percent-clipped=1.0 +2023-04-26 17:16:49,196 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28348.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:16:56,322 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1493, 3.3038, 0.7941, 1.5805, 1.6781, 2.2147, 1.8381, 0.9854], + device='cuda:3'), covar=tensor([0.2032, 0.2026, 0.2735, 0.2164, 0.1699, 0.1668, 0.1946, 0.2500], + device='cuda:3'), in_proj_covar=tensor([0.0124, 0.0267, 0.0148, 0.0130, 0.0141, 0.0164, 0.0126, 0.0130], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 17:16:57,452 INFO [finetune.py:976] (3/7) Epoch 5, batch 5450, loss[loss=0.1678, simple_loss=0.2342, pruned_loss=0.05072, over 4826.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2684, pruned_loss=0.07296, over 952364.26 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:17:03,579 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28370.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:17:33,974 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1015, 2.6154, 1.0161, 1.4380, 1.8049, 1.2487, 3.5645, 1.7405], + device='cuda:3'), covar=tensor([0.0702, 0.0764, 0.0855, 0.1304, 0.0582, 0.0987, 0.0234, 0.0680], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0054, 0.0054, 0.0082, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 17:17:37,439 INFO [finetune.py:976] (3/7) Epoch 5, batch 5500, loss[loss=0.2017, simple_loss=0.2589, pruned_loss=0.07225, over 4905.00 frames. ], tot_loss[loss=0.205, simple_loss=0.2659, pruned_loss=0.072, over 953095.25 frames. ], batch size: 35, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:17:47,036 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28418.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:17:59,135 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.711e+02 2.116e+02 2.529e+02 5.388e+02, threshold=4.231e+02, percent-clipped=4.0 +2023-04-26 17:18:20,484 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7833, 1.6031, 1.9994, 2.0781, 2.0064, 1.6372, 1.7812, 1.8357], + device='cuda:3'), covar=tensor([1.0722, 1.4552, 1.5506, 1.5729, 1.1241, 1.8634, 1.8114, 1.5205], + device='cuda:3'), in_proj_covar=tensor([0.0421, 0.0453, 0.0538, 0.0559, 0.0449, 0.0472, 0.0484, 0.0485], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:18:49,580 INFO [finetune.py:976] (3/7) Epoch 5, batch 5550, loss[loss=0.2524, simple_loss=0.3032, pruned_loss=0.1008, over 4819.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2673, pruned_loss=0.0728, over 953817.07 frames. ], batch size: 40, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:19:04,273 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5545, 3.4513, 2.8274, 4.0898, 3.4022, 3.5324, 1.8006, 3.5300], + device='cuda:3'), covar=tensor([0.1720, 0.1207, 0.3769, 0.1390, 0.2835, 0.1849, 0.4925, 0.2176], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0221, 0.0257, 0.0314, 0.0305, 0.0257, 0.0278, 0.0279], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:19:09,807 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28486.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:19:10,870 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3032, 3.0218, 0.9624, 1.4815, 1.9745, 1.4230, 3.9790, 1.8959], + device='cuda:3'), covar=tensor([0.0674, 0.0751, 0.0964, 0.1432, 0.0616, 0.0997, 0.0181, 0.0654], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0054, 0.0054, 0.0082, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 17:19:14,433 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28492.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:19:25,623 INFO [finetune.py:976] (3/7) Epoch 5, batch 5600, loss[loss=0.2647, simple_loss=0.3179, pruned_loss=0.1057, over 4919.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2726, pruned_loss=0.07455, over 954538.57 frames. ], batch size: 37, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:19:34,906 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.757e+02 2.198e+02 2.682e+02 4.335e+02, threshold=4.395e+02, percent-clipped=1.0 +2023-04-26 17:19:47,089 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28547.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:19:51,033 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28553.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:19:55,612 INFO [finetune.py:976] (3/7) Epoch 5, batch 5650, loss[loss=0.2393, simple_loss=0.2888, pruned_loss=0.09489, over 4799.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.275, pruned_loss=0.07526, over 954764.52 frames. ], batch size: 25, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:20:09,689 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28585.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:20:25,139 INFO [finetune.py:976] (3/7) Epoch 5, batch 5700, loss[loss=0.1622, simple_loss=0.2166, pruned_loss=0.05387, over 3574.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2707, pruned_loss=0.07421, over 941652.74 frames. ], batch size: 15, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:20:25,177 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28611.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:20:34,600 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.668e+02 2.021e+02 2.528e+02 3.624e+02, threshold=4.042e+02, percent-clipped=0.0 +2023-04-26 17:20:58,934 INFO [finetune.py:976] (3/7) Epoch 6, batch 0, loss[loss=0.2168, simple_loss=0.2848, pruned_loss=0.07437, over 4819.00 frames. ], tot_loss[loss=0.2168, simple_loss=0.2848, pruned_loss=0.07437, over 4819.00 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:20:58,935 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 17:21:02,300 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1782, 1.6620, 1.4051, 1.8697, 1.6071, 1.9177, 1.4166, 3.0591], + device='cuda:3'), covar=tensor([0.0682, 0.0760, 0.0752, 0.1159, 0.0631, 0.0447, 0.0742, 0.0187], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 17:21:13,392 INFO [finetune.py:1010] (3/7) Epoch 6, validation: loss=0.1605, simple_loss=0.2337, pruned_loss=0.04366, over 2265189.00 frames. +2023-04-26 17:21:13,392 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 17:21:19,202 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28643.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:21:19,250 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0666, 1.9725, 2.2558, 2.4816, 2.4766, 2.0081, 1.6070, 2.1452], + device='cuda:3'), covar=tensor([0.0957, 0.1052, 0.0617, 0.0631, 0.0601, 0.0964, 0.0989, 0.0636], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0204, 0.0181, 0.0177, 0.0177, 0.0192, 0.0165, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:21:21,107 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0237, 1.8052, 2.1155, 2.3703, 2.3892, 1.9033, 1.4653, 2.0359], + device='cuda:3'), covar=tensor([0.0966, 0.1110, 0.0600, 0.0621, 0.0598, 0.1037, 0.1093, 0.0661], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0204, 0.0181, 0.0177, 0.0177, 0.0192, 0.0165, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:21:29,878 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6509, 1.1176, 1.3231, 1.2829, 1.8793, 1.4978, 1.2086, 1.2136], + device='cuda:3'), covar=tensor([0.1743, 0.1721, 0.2192, 0.1570, 0.0898, 0.1458, 0.2244, 0.2124], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0331, 0.0349, 0.0305, 0.0341, 0.0336, 0.0305, 0.0351], + device='cuda:3'), out_proj_covar=tensor([6.6652e-05, 7.0850e-05, 7.5417e-05, 6.3753e-05, 7.2213e-05, 7.2886e-05, + 6.6233e-05, 7.5610e-05], device='cuda:3') +2023-04-26 17:21:59,786 INFO [finetune.py:976] (3/7) Epoch 6, batch 50, loss[loss=0.1798, simple_loss=0.2444, pruned_loss=0.05761, over 4064.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2728, pruned_loss=0.07357, over 215623.29 frames. ], batch size: 66, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:22:14,739 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4292, 1.2187, 1.6740, 1.5695, 1.3306, 1.1616, 1.3365, 0.8368], + device='cuda:3'), covar=tensor([0.0696, 0.0859, 0.0428, 0.0709, 0.0937, 0.1379, 0.0585, 0.0932], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0075, 0.0074, 0.0068, 0.0079, 0.0096, 0.0082, 0.0077], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 17:22:24,839 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.869e+02 2.241e+02 2.636e+02 5.025e+02, threshold=4.483e+02, percent-clipped=7.0 +2023-04-26 17:22:33,641 INFO [finetune.py:976] (3/7) Epoch 6, batch 100, loss[loss=0.1998, simple_loss=0.2613, pruned_loss=0.06915, over 4901.00 frames. ], tot_loss[loss=0.2064, simple_loss=0.2687, pruned_loss=0.07204, over 381163.90 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:22:43,767 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0987, 2.6495, 1.4485, 1.8364, 2.5894, 2.0119, 1.9365, 2.1041], + device='cuda:3'), covar=tensor([0.0518, 0.0329, 0.0316, 0.0557, 0.0223, 0.0531, 0.0513, 0.0560], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0049], + device='cuda:3') +2023-04-26 17:23:06,410 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-26 17:23:06,833 INFO [finetune.py:976] (3/7) Epoch 6, batch 150, loss[loss=0.1923, simple_loss=0.2553, pruned_loss=0.06469, over 4820.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2669, pruned_loss=0.07311, over 506252.93 frames. ], batch size: 41, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:23:10,939 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6137, 1.3460, 1.9088, 1.8022, 1.4207, 1.1485, 1.5158, 0.8873], + device='cuda:3'), covar=tensor([0.0798, 0.0929, 0.0483, 0.0822, 0.1057, 0.1542, 0.0785, 0.1070], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0076, 0.0074, 0.0069, 0.0079, 0.0096, 0.0082, 0.0077], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 17:23:36,887 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.735e+02 2.162e+02 2.514e+02 4.413e+02, threshold=4.325e+02, percent-clipped=0.0 +2023-04-26 17:23:57,268 INFO [finetune.py:976] (3/7) Epoch 6, batch 200, loss[loss=0.2012, simple_loss=0.2726, pruned_loss=0.0649, over 4822.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2668, pruned_loss=0.07415, over 606600.15 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:23:59,202 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28842.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:24:08,154 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:24:22,305 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28861.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:24:55,334 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28885.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:25:03,860 INFO [finetune.py:976] (3/7) Epoch 6, batch 250, loss[loss=0.2217, simple_loss=0.2924, pruned_loss=0.07546, over 4901.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.2692, pruned_loss=0.07414, over 683473.39 frames. ], batch size: 35, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:25:15,815 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9900, 2.6135, 2.0573, 2.4904, 1.8698, 2.1481, 2.1871, 1.8530], + device='cuda:3'), covar=tensor([0.2273, 0.1191, 0.0959, 0.1353, 0.2901, 0.1304, 0.1831, 0.2959], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0323, 0.0235, 0.0298, 0.0317, 0.0278, 0.0266, 0.0290], + device='cuda:3'), out_proj_covar=tensor([1.2401e-04, 1.3126e-04, 9.5598e-05, 1.1960e-04, 1.3056e-04, 1.1257e-04, + 1.0937e-04, 1.1682e-04], device='cuda:3') +2023-04-26 17:25:19,139 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 17:25:29,050 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28911.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 17:25:40,829 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28922.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:25:46,465 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.358e+02 1.791e+02 2.229e+02 2.957e+02 8.157e+02, threshold=4.458e+02, percent-clipped=7.0 +2023-04-26 17:25:54,904 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28933.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:25:59,502 INFO [finetune.py:976] (3/7) Epoch 6, batch 300, loss[loss=0.1991, simple_loss=0.2725, pruned_loss=0.06283, over 4889.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2707, pruned_loss=0.07426, over 744887.84 frames. ], batch size: 35, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:26:08,045 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28943.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:26:28,687 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28959.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:27:05,111 INFO [finetune.py:976] (3/7) Epoch 6, batch 350, loss[loss=0.2466, simple_loss=0.3076, pruned_loss=0.09279, over 4814.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2738, pruned_loss=0.07596, over 790777.98 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:27:11,878 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28991.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:27:24,953 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29002.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:27:53,128 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.828e+02 2.192e+02 2.525e+02 3.973e+02, threshold=4.385e+02, percent-clipped=0.0 +2023-04-26 17:27:59,802 INFO [finetune.py:976] (3/7) Epoch 6, batch 400, loss[loss=0.2275, simple_loss=0.2982, pruned_loss=0.07842, over 4813.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2737, pruned_loss=0.07525, over 826320.32 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:28:01,734 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-26 17:28:16,907 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29063.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:28:33,169 INFO [finetune.py:976] (3/7) Epoch 6, batch 450, loss[loss=0.1837, simple_loss=0.2502, pruned_loss=0.05865, over 4820.00 frames. ], tot_loss[loss=0.2118, simple_loss=0.2736, pruned_loss=0.07497, over 856468.64 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:28:36,866 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1137, 1.3182, 1.5106, 1.6774, 1.6228, 1.7698, 1.5285, 1.5914], + device='cuda:3'), covar=tensor([0.6644, 0.9899, 0.9106, 0.7746, 0.9548, 1.3120, 1.0502, 0.8736], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0398, 0.0318, 0.0328, 0.0348, 0.0414, 0.0378, 0.0334], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:28:59,677 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.667e+02 1.998e+02 2.325e+02 3.966e+02, threshold=3.996e+02, percent-clipped=0.0 +2023-04-26 17:29:06,385 INFO [finetune.py:976] (3/7) Epoch 6, batch 500, loss[loss=0.2008, simple_loss=0.2586, pruned_loss=0.07151, over 4705.00 frames. ], tot_loss[loss=0.2089, simple_loss=0.2705, pruned_loss=0.07362, over 877746.64 frames. ], batch size: 23, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:29:08,295 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29142.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:13,390 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29148.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:39,684 INFO [finetune.py:976] (3/7) Epoch 6, batch 550, loss[loss=0.2753, simple_loss=0.3015, pruned_loss=0.1246, over 4690.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2671, pruned_loss=0.0723, over 896361.14 frames. ], batch size: 23, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:29:40,347 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29190.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:44,458 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29196.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:59,117 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29217.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:30:05,732 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5716, 1.5813, 3.6655, 3.4315, 3.3080, 3.3374, 3.3304, 3.3020], + device='cuda:3'), covar=tensor([0.6363, 0.4573, 0.1165, 0.1621, 0.1058, 0.1858, 0.3836, 0.1357], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0309, 0.0421, 0.0425, 0.0358, 0.0414, 0.0322, 0.0379], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:30:06,230 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.252e+02 1.833e+02 2.160e+02 2.601e+02 5.201e+02, threshold=4.320e+02, percent-clipped=1.0 +2023-04-26 17:30:11,168 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:30:12,890 INFO [finetune.py:976] (3/7) Epoch 6, batch 600, loss[loss=0.1461, simple_loss=0.2123, pruned_loss=0.03994, over 4743.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2676, pruned_loss=0.07319, over 909889.24 frames. ], batch size: 27, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:30:47,030 INFO [finetune.py:976] (3/7) Epoch 6, batch 650, loss[loss=0.241, simple_loss=0.3111, pruned_loss=0.08551, over 4816.00 frames. ], tot_loss[loss=0.211, simple_loss=0.272, pruned_loss=0.07497, over 922452.94 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:30:57,020 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5249, 1.6046, 0.8566, 1.2784, 1.7050, 1.4374, 1.3286, 1.3608], + device='cuda:3'), covar=tensor([0.0565, 0.0431, 0.0426, 0.0588, 0.0321, 0.0593, 0.0592, 0.0660], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:3') +2023-04-26 17:30:57,642 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29297.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:31:17,383 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7412, 2.3239, 1.9270, 2.2727, 1.6182, 1.8526, 2.0881, 1.6377], + device='cuda:3'), covar=tensor([0.2487, 0.1468, 0.0907, 0.1337, 0.3162, 0.1326, 0.1976, 0.2762], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0326, 0.0236, 0.0301, 0.0319, 0.0281, 0.0268, 0.0292], + device='cuda:3'), out_proj_covar=tensor([1.2491e-04, 1.3243e-04, 9.6161e-05, 1.2093e-04, 1.3140e-04, 1.1365e-04, + 1.1034e-04, 1.1792e-04], device='cuda:3') +2023-04-26 17:31:42,432 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.237e+02 1.829e+02 2.148e+02 2.649e+02 3.979e+02, threshold=4.296e+02, percent-clipped=0.0 +2023-04-26 17:32:01,094 INFO [finetune.py:976] (3/7) Epoch 6, batch 700, loss[loss=0.2658, simple_loss=0.3167, pruned_loss=0.1075, over 4069.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2724, pruned_loss=0.07492, over 929263.79 frames. ], batch size: 65, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:32:25,506 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29358.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:33:07,833 INFO [finetune.py:976] (3/7) Epoch 6, batch 750, loss[loss=0.1907, simple_loss=0.2486, pruned_loss=0.06644, over 4740.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2742, pruned_loss=0.07541, over 937112.59 frames. ], batch size: 23, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:33:23,034 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4474, 0.9591, 1.1707, 1.0830, 1.6330, 1.2740, 1.0133, 1.1298], + device='cuda:3'), covar=tensor([0.1974, 0.1567, 0.2204, 0.1553, 0.0899, 0.1671, 0.2223, 0.2149], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0332, 0.0351, 0.0308, 0.0342, 0.0337, 0.0307, 0.0352], + device='cuda:3'), out_proj_covar=tensor([6.6792e-05, 7.0915e-05, 7.6051e-05, 6.4234e-05, 7.2273e-05, 7.2948e-05, + 6.6471e-05, 7.5794e-05], device='cuda:3') +2023-04-26 17:33:42,159 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2743, 3.3768, 2.7683, 3.8803, 3.2795, 3.3285, 1.7168, 3.3743], + device='cuda:3'), covar=tensor([0.2030, 0.1207, 0.3769, 0.1794, 0.2659, 0.1817, 0.4838, 0.2325], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0218, 0.0254, 0.0312, 0.0303, 0.0257, 0.0275, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:34:03,539 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.229e+02 1.738e+02 2.111e+02 2.599e+02 4.246e+02, threshold=4.221e+02, percent-clipped=0.0 +2023-04-26 17:34:05,514 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.3242, 1.3386, 1.3882, 0.9406, 1.3694, 1.1219, 1.7519, 1.2607], + device='cuda:3'), covar=tensor([0.3892, 0.1809, 0.5634, 0.2824, 0.1782, 0.2253, 0.1595, 0.5035], + device='cuda:3'), in_proj_covar=tensor([0.0351, 0.0359, 0.0439, 0.0370, 0.0398, 0.0387, 0.0393, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:34:14,520 INFO [finetune.py:976] (3/7) Epoch 6, batch 800, loss[loss=0.1613, simple_loss=0.2358, pruned_loss=0.0434, over 4808.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2735, pruned_loss=0.07444, over 941993.92 frames. ], batch size: 41, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:34:32,123 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-26 17:34:48,533 INFO [finetune.py:976] (3/7) Epoch 6, batch 850, loss[loss=0.2327, simple_loss=0.2815, pruned_loss=0.09191, over 4745.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2708, pruned_loss=0.07373, over 945661.81 frames. ], batch size: 27, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:35:07,988 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29517.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:35:14,515 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.686e+02 1.967e+02 2.412e+02 4.596e+02, threshold=3.934e+02, percent-clipped=4.0 +2023-04-26 17:35:22,137 INFO [finetune.py:976] (3/7) Epoch 6, batch 900, loss[loss=0.2198, simple_loss=0.2914, pruned_loss=0.07408, over 4828.00 frames. ], tot_loss[loss=0.208, simple_loss=0.2687, pruned_loss=0.07362, over 947766.34 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:35:25,903 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6546, 2.0008, 1.5730, 1.2789, 1.2842, 1.2563, 1.5601, 1.2440], + device='cuda:3'), covar=tensor([0.1878, 0.1671, 0.1722, 0.2229, 0.2740, 0.2249, 0.1339, 0.2309], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0220, 0.0176, 0.0206, 0.0212, 0.0186, 0.0169, 0.0192], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 17:35:30,042 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:35:38,987 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:35:42,776 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.35 vs. limit=5.0 +2023-04-26 17:35:45,778 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.43 vs. limit=5.0 +2023-04-26 17:35:56,021 INFO [finetune.py:976] (3/7) Epoch 6, batch 950, loss[loss=0.2035, simple_loss=0.2566, pruned_loss=0.0752, over 4870.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2661, pruned_loss=0.0725, over 949336.90 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:35:57,877 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29592.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:36:10,613 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29612.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:36:23,634 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9666, 2.3951, 1.1413, 1.2592, 1.9292, 1.1432, 2.9445, 1.5000], + device='cuda:3'), covar=tensor([0.0668, 0.0596, 0.0755, 0.1203, 0.0460, 0.0992, 0.0244, 0.0662], + device='cuda:3'), in_proj_covar=tensor([0.0053, 0.0070, 0.0052, 0.0049, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 17:36:27,225 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.239e+02 1.864e+02 2.273e+02 2.659e+02 4.416e+02, threshold=4.547e+02, percent-clipped=3.0 +2023-04-26 17:36:39,506 INFO [finetune.py:976] (3/7) Epoch 6, batch 1000, loss[loss=0.2562, simple_loss=0.3057, pruned_loss=0.1033, over 4930.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2692, pruned_loss=0.07372, over 950312.45 frames. ], batch size: 33, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:36:48,246 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-26 17:36:56,968 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-26 17:37:09,274 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29658.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:37:11,721 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29662.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:37:52,136 INFO [finetune.py:976] (3/7) Epoch 6, batch 1050, loss[loss=0.1931, simple_loss=0.2672, pruned_loss=0.05945, over 4863.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2708, pruned_loss=0.07365, over 950641.74 frames. ], batch size: 34, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:37:56,557 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1815, 1.5983, 1.3254, 1.8134, 1.5667, 1.9528, 1.3601, 3.5861], + device='cuda:3'), covar=tensor([0.0705, 0.0838, 0.0874, 0.1250, 0.0697, 0.0591, 0.0845, 0.0149], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 17:38:13,519 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29706.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:38:14,497 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-04-26 17:38:17,206 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5707, 1.9040, 1.6512, 1.8493, 1.4672, 1.4899, 1.6382, 1.3014], + device='cuda:3'), covar=tensor([0.2090, 0.1298, 0.0987, 0.1259, 0.3482, 0.1464, 0.2102, 0.2674], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0326, 0.0237, 0.0300, 0.0318, 0.0281, 0.0268, 0.0293], + device='cuda:3'), out_proj_covar=tensor([1.2478e-04, 1.3209e-04, 9.6338e-05, 1.2045e-04, 1.3101e-04, 1.1374e-04, + 1.0992e-04, 1.1800e-04], device='cuda:3') +2023-04-26 17:38:34,966 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29723.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 17:38:39,284 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.744e+02 2.054e+02 2.475e+02 7.840e+02, threshold=4.108e+02, percent-clipped=2.0 +2023-04-26 17:38:57,681 INFO [finetune.py:976] (3/7) Epoch 6, batch 1100, loss[loss=0.2183, simple_loss=0.2773, pruned_loss=0.07966, over 4902.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2743, pruned_loss=0.07559, over 952209.28 frames. ], batch size: 37, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:39:51,670 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-26 17:39:58,701 INFO [finetune.py:976] (3/7) Epoch 6, batch 1150, loss[loss=0.1809, simple_loss=0.2618, pruned_loss=0.05, over 4948.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2744, pruned_loss=0.07526, over 952997.59 frames. ], batch size: 29, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:40:07,756 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29802.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:40:22,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7471, 2.2796, 1.9314, 2.2291, 1.6195, 1.9636, 2.0233, 1.4733], + device='cuda:3'), covar=tensor([0.2465, 0.1465, 0.0887, 0.1395, 0.3167, 0.1390, 0.2024, 0.3177], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0325, 0.0237, 0.0300, 0.0318, 0.0281, 0.0267, 0.0293], + device='cuda:3'), out_proj_covar=tensor([1.2492e-04, 1.3191e-04, 9.6184e-05, 1.2029e-04, 1.3072e-04, 1.1375e-04, + 1.0972e-04, 1.1803e-04], device='cuda:3') +2023-04-26 17:40:24,363 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.784e+02 2.105e+02 2.513e+02 8.942e+02, threshold=4.210e+02, percent-clipped=4.0 +2023-04-26 17:40:31,966 INFO [finetune.py:976] (3/7) Epoch 6, batch 1200, loss[loss=0.2558, simple_loss=0.2837, pruned_loss=0.1139, over 4189.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2725, pruned_loss=0.0749, over 950579.80 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:40:48,154 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29863.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:41:05,314 INFO [finetune.py:976] (3/7) Epoch 6, batch 1250, loss[loss=0.2625, simple_loss=0.2974, pruned_loss=0.1138, over 4198.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2696, pruned_loss=0.07352, over 950805.32 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:41:07,207 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29892.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:41:17,261 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29907.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:41:30,522 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.701e+02 1.937e+02 2.233e+02 5.324e+02, threshold=3.875e+02, percent-clipped=2.0 +2023-04-26 17:41:38,730 INFO [finetune.py:976] (3/7) Epoch 6, batch 1300, loss[loss=0.1791, simple_loss=0.2331, pruned_loss=0.06255, over 4837.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2661, pruned_loss=0.07184, over 953582.48 frames. ], batch size: 25, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:41:39,346 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:42:12,107 INFO [finetune.py:976] (3/7) Epoch 6, batch 1350, loss[loss=0.2399, simple_loss=0.3082, pruned_loss=0.08586, over 4800.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2677, pruned_loss=0.07298, over 953712.88 frames. ], batch size: 45, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:42:18,307 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 17:42:32,672 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30018.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:42:39,145 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.331e+02 1.856e+02 2.103e+02 2.497e+02 6.080e+02, threshold=4.207e+02, percent-clipped=3.0 +2023-04-26 17:42:40,831 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 17:42:42,318 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2166, 1.9535, 2.3008, 2.6093, 2.6095, 2.1703, 1.8196, 2.0811], + device='cuda:3'), covar=tensor([0.0795, 0.1036, 0.0498, 0.0525, 0.0519, 0.0774, 0.0981, 0.0638], + device='cuda:3'), in_proj_covar=tensor([0.0203, 0.0206, 0.0182, 0.0179, 0.0179, 0.0194, 0.0166, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:42:52,097 INFO [finetune.py:976] (3/7) Epoch 6, batch 1400, loss[loss=0.2421, simple_loss=0.2923, pruned_loss=0.09596, over 4774.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.271, pruned_loss=0.07361, over 953965.45 frames. ], batch size: 28, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:43:56,757 INFO [finetune.py:976] (3/7) Epoch 6, batch 1450, loss[loss=0.2872, simple_loss=0.336, pruned_loss=0.1192, over 4250.00 frames. ], tot_loss[loss=0.2102, simple_loss=0.273, pruned_loss=0.07373, over 954626.83 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:44:42,773 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30125.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:44:44,462 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.421e+02 1.906e+02 2.234e+02 2.642e+02 4.905e+02, threshold=4.468e+02, percent-clipped=1.0 +2023-04-26 17:44:54,921 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 17:44:57,145 INFO [finetune.py:976] (3/7) Epoch 6, batch 1500, loss[loss=0.1969, simple_loss=0.2736, pruned_loss=0.06006, over 4822.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2754, pruned_loss=0.075, over 954804.80 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:44:57,470 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-26 17:45:25,797 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30158.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:45:59,237 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30186.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:46:00,949 INFO [finetune.py:976] (3/7) Epoch 6, batch 1550, loss[loss=0.1919, simple_loss=0.256, pruned_loss=0.06393, over 4862.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.274, pruned_loss=0.0741, over 954880.91 frames. ], batch size: 34, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:46:30,716 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30207.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:46:31,956 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30209.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:46:56,315 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.842e+02 2.143e+02 2.537e+02 6.404e+02, threshold=4.287e+02, percent-clipped=2.0 +2023-04-26 17:47:09,505 INFO [finetune.py:976] (3/7) Epoch 6, batch 1600, loss[loss=0.2203, simple_loss=0.2833, pruned_loss=0.07866, over 4755.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2704, pruned_loss=0.07292, over 953999.79 frames. ], batch size: 59, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:47:31,236 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30255.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:47:38,063 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 17:47:41,431 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30270.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:47:46,954 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8032, 1.7548, 1.9989, 2.3268, 2.2756, 1.8116, 1.5141, 1.9227], + device='cuda:3'), covar=tensor([0.0960, 0.1117, 0.0657, 0.0583, 0.0621, 0.0930, 0.0909, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0205, 0.0181, 0.0178, 0.0179, 0.0194, 0.0165, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:47:53,335 INFO [finetune.py:976] (3/7) Epoch 6, batch 1650, loss[loss=0.1466, simple_loss=0.2165, pruned_loss=0.03836, over 4765.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2678, pruned_loss=0.07213, over 954332.04 frames. ], batch size: 28, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:48:13,368 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30318.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:48:16,464 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2093, 1.4043, 1.5641, 1.7474, 1.6343, 1.7804, 1.6591, 1.6911], + device='cuda:3'), covar=tensor([0.6596, 1.0135, 0.8232, 0.7440, 0.9057, 1.2834, 0.9559, 0.8389], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0396, 0.0318, 0.0328, 0.0347, 0.0413, 0.0377, 0.0334], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:48:19,318 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.688e+02 2.054e+02 2.477e+02 3.898e+02, threshold=4.108e+02, percent-clipped=0.0 +2023-04-26 17:48:26,039 INFO [finetune.py:976] (3/7) Epoch 6, batch 1700, loss[loss=0.2185, simple_loss=0.277, pruned_loss=0.08, over 4936.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2659, pruned_loss=0.07172, over 956891.28 frames. ], batch size: 33, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:48:26,263 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.04 vs. limit=2.0 +2023-04-26 17:48:51,276 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30360.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:49:00,654 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30366.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:49:26,051 INFO [finetune.py:976] (3/7) Epoch 6, batch 1750, loss[loss=0.2521, simple_loss=0.3246, pruned_loss=0.08979, over 4865.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2669, pruned_loss=0.0719, over 955526.52 frames. ], batch size: 44, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:50:03,364 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30421.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:50:08,047 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.357e+02 1.827e+02 2.242e+02 2.631e+02 5.410e+02, threshold=4.483e+02, percent-clipped=4.0 +2023-04-26 17:50:20,187 INFO [finetune.py:976] (3/7) Epoch 6, batch 1800, loss[loss=0.2046, simple_loss=0.2809, pruned_loss=0.06415, over 4755.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2695, pruned_loss=0.07216, over 955233.24 frames. ], batch size: 54, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:50:30,064 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 17:50:43,032 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:51:09,062 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30481.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:51:13,897 INFO [finetune.py:976] (3/7) Epoch 6, batch 1850, loss[loss=0.197, simple_loss=0.2678, pruned_loss=0.06309, over 4732.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.2714, pruned_loss=0.07309, over 955576.07 frames. ], batch size: 59, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:51:24,867 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:51:40,051 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.288e+02 1.791e+02 2.106e+02 2.463e+02 4.581e+02, threshold=4.213e+02, percent-clipped=1.0 +2023-04-26 17:51:47,205 INFO [finetune.py:976] (3/7) Epoch 6, batch 1900, loss[loss=0.2071, simple_loss=0.2708, pruned_loss=0.07174, over 4826.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2726, pruned_loss=0.07331, over 956275.85 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:52:17,583 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:52:34,610 INFO [finetune.py:976] (3/7) Epoch 6, batch 1950, loss[loss=0.2211, simple_loss=0.2835, pruned_loss=0.07934, over 4285.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2694, pruned_loss=0.07225, over 954940.65 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:53:00,332 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.660e+02 1.977e+02 2.493e+02 4.247e+02, threshold=3.954e+02, percent-clipped=1.0 +2023-04-26 17:53:08,042 INFO [finetune.py:976] (3/7) Epoch 6, batch 2000, loss[loss=0.2117, simple_loss=0.2639, pruned_loss=0.07974, over 4819.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2677, pruned_loss=0.07234, over 954007.77 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:53:08,698 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0578, 1.4202, 4.8632, 4.5145, 4.2282, 4.5553, 4.3419, 4.2746], + device='cuda:3'), covar=tensor([0.6149, 0.5463, 0.0981, 0.1825, 0.0916, 0.1107, 0.1448, 0.1531], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0303, 0.0414, 0.0417, 0.0353, 0.0407, 0.0316, 0.0372], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:53:17,876 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30655.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:53:20,216 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8284, 4.3337, 0.8419, 2.2770, 2.4826, 2.8980, 2.5490, 0.9352], + device='cuda:3'), covar=tensor([0.1392, 0.0905, 0.2338, 0.1243, 0.1005, 0.1042, 0.1373, 0.2209], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0260, 0.0146, 0.0127, 0.0137, 0.0159, 0.0123, 0.0126], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 17:53:41,352 INFO [finetune.py:976] (3/7) Epoch 6, batch 2050, loss[loss=0.1549, simple_loss=0.225, pruned_loss=0.04238, over 4755.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2645, pruned_loss=0.0709, over 955281.02 frames. ], batch size: 27, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:53:57,178 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4522, 1.7604, 1.3423, 1.1425, 1.2514, 1.1943, 1.3551, 1.1470], + device='cuda:3'), covar=tensor([0.1552, 0.1435, 0.1673, 0.1998, 0.2479, 0.1961, 0.1168, 0.2061], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0219, 0.0175, 0.0205, 0.0211, 0.0185, 0.0167, 0.0192], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 17:53:58,934 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30716.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:53:58,985 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30716.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 17:54:12,373 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.673e+02 1.948e+02 2.190e+02 5.524e+02, threshold=3.896e+02, percent-clipped=3.0 +2023-04-26 17:54:14,797 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.0444, 4.0050, 2.8436, 4.7506, 4.1788, 4.0870, 1.9584, 4.0440], + device='cuda:3'), covar=tensor([0.1512, 0.0994, 0.3031, 0.1397, 0.3099, 0.1833, 0.5216, 0.2204], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0220, 0.0255, 0.0313, 0.0304, 0.0256, 0.0276, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:54:25,816 INFO [finetune.py:976] (3/7) Epoch 6, batch 2100, loss[loss=0.2709, simple_loss=0.3196, pruned_loss=0.1111, over 4056.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2648, pruned_loss=0.0711, over 955185.86 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:54:40,396 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5972, 1.3020, 1.7539, 1.9610, 1.7764, 1.5523, 1.5786, 1.6719], + device='cuda:3'), covar=tensor([0.8790, 1.1369, 1.1458, 1.2843, 0.9869, 1.4391, 1.3925, 1.2174], + device='cuda:3'), in_proj_covar=tensor([0.0416, 0.0444, 0.0530, 0.0550, 0.0445, 0.0466, 0.0479, 0.0478], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 17:54:54,253 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30781.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:54:59,003 INFO [finetune.py:976] (3/7) Epoch 6, batch 2150, loss[loss=0.2239, simple_loss=0.2926, pruned_loss=0.07759, over 4830.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2707, pruned_loss=0.07368, over 956083.56 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:55:10,758 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.21 vs. limit=5.0 +2023-04-26 17:55:17,736 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3502, 2.7700, 1.0885, 1.4904, 2.1051, 1.3999, 3.8217, 1.9028], + device='cuda:3'), covar=tensor([0.0594, 0.0657, 0.0805, 0.1284, 0.0506, 0.0970, 0.0200, 0.0588], + device='cuda:3'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 17:55:40,198 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.783e+02 2.191e+02 2.615e+02 4.573e+02, threshold=4.381e+02, percent-clipped=3.0 +2023-04-26 17:55:40,855 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30829.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:55:54,384 INFO [finetune.py:976] (3/7) Epoch 6, batch 2200, loss[loss=0.1958, simple_loss=0.2597, pruned_loss=0.0659, over 4858.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2727, pruned_loss=0.07472, over 956617.61 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:56:23,798 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30865.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:56:50,922 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1868, 1.4281, 1.5015, 1.6321, 1.5062, 1.7074, 1.6006, 1.5771], + device='cuda:3'), covar=tensor([0.6578, 0.9808, 0.8429, 0.7831, 0.9340, 1.3304, 1.0022, 0.8803], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0394, 0.0316, 0.0326, 0.0345, 0.0410, 0.0374, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:56:51,897 INFO [finetune.py:976] (3/7) Epoch 6, batch 2250, loss[loss=0.2272, simple_loss=0.2938, pruned_loss=0.08025, over 4861.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2748, pruned_loss=0.07557, over 954799.36 frames. ], batch size: 34, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:57:11,899 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30903.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:57:22,864 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30913.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:57:40,822 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30925.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:57:42,535 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.673e+02 1.940e+02 2.404e+02 3.842e+02, threshold=3.880e+02, percent-clipped=0.0 +2023-04-26 17:57:46,818 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30934.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:57:50,253 INFO [finetune.py:976] (3/7) Epoch 6, batch 2300, loss[loss=0.173, simple_loss=0.2503, pruned_loss=0.04786, over 4907.00 frames. ], tot_loss[loss=0.2131, simple_loss=0.2754, pruned_loss=0.07542, over 955998.01 frames. ], batch size: 37, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:58:07,559 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30964.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:08,773 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5563, 1.7070, 1.6902, 2.3124, 2.5324, 2.1372, 1.9930, 1.8273], + device='cuda:3'), covar=tensor([0.1650, 0.1889, 0.2271, 0.1577, 0.1144, 0.2003, 0.2559, 0.2086], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0329, 0.0348, 0.0304, 0.0339, 0.0334, 0.0307, 0.0349], + device='cuda:3'), out_proj_covar=tensor([6.6197e-05, 7.0328e-05, 7.5343e-05, 6.3378e-05, 7.1561e-05, 7.2253e-05, + 6.6603e-05, 7.5253e-05], device='cuda:3') +2023-04-26 17:58:21,388 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30986.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:23,530 INFO [finetune.py:976] (3/7) Epoch 6, batch 2350, loss[loss=0.2371, simple_loss=0.299, pruned_loss=0.08764, over 4912.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2726, pruned_loss=0.07457, over 953613.69 frames. ], batch size: 46, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:58:28,702 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30995.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:35,902 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8873, 2.8585, 2.1684, 3.3243, 2.8539, 2.8696, 1.0657, 2.7923], + device='cuda:3'), covar=tensor([0.2004, 0.1584, 0.3396, 0.2602, 0.3088, 0.2227, 0.5813, 0.3093], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0220, 0.0255, 0.0314, 0.0303, 0.0255, 0.0275, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 17:58:39,475 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31011.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:58:39,503 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31011.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:58:42,542 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31016.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:49,762 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.679e+02 2.129e+02 2.680e+02 4.388e+02, threshold=4.258e+02, percent-clipped=1.0 +2023-04-26 17:58:56,924 INFO [finetune.py:976] (3/7) Epoch 6, batch 2400, loss[loss=0.1871, simple_loss=0.2625, pruned_loss=0.05586, over 4822.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2704, pruned_loss=0.07378, over 954181.80 frames. ], batch size: 25, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:59:03,511 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-04-26 17:59:14,972 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31064.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:59:19,941 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31072.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 17:59:30,874 INFO [finetune.py:976] (3/7) Epoch 6, batch 2450, loss[loss=0.2039, simple_loss=0.2568, pruned_loss=0.07548, over 4827.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2667, pruned_loss=0.07222, over 954116.01 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:59:31,652 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-26 17:59:57,657 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.941e+02 2.240e+02 2.729e+02 4.178e+02, threshold=4.480e+02, percent-clipped=0.0 +2023-04-26 18:00:04,430 INFO [finetune.py:976] (3/7) Epoch 6, batch 2500, loss[loss=0.1835, simple_loss=0.2427, pruned_loss=0.06214, over 4770.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2676, pruned_loss=0.07294, over 953284.78 frames. ], batch size: 28, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:00:37,627 INFO [finetune.py:976] (3/7) Epoch 6, batch 2550, loss[loss=0.1937, simple_loss=0.266, pruned_loss=0.06067, over 4890.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.271, pruned_loss=0.0742, over 953072.20 frames. ], batch size: 32, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:01:09,724 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.313e+02 1.772e+02 2.143e+02 2.617e+02 5.206e+02, threshold=4.285e+02, percent-clipped=3.0 +2023-04-26 18:01:18,259 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9132, 2.5407, 0.9312, 1.1560, 1.6834, 1.1413, 3.3600, 1.4937], + device='cuda:3'), covar=tensor([0.0906, 0.0824, 0.1065, 0.1825, 0.0753, 0.1451, 0.0399, 0.1058], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0070, 0.0051, 0.0049, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 18:01:22,510 INFO [finetune.py:976] (3/7) Epoch 6, batch 2600, loss[loss=0.2115, simple_loss=0.2808, pruned_loss=0.07116, over 4919.00 frames. ], tot_loss[loss=0.2095, simple_loss=0.2715, pruned_loss=0.07371, over 951783.64 frames. ], batch size: 38, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:01:51,863 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31259.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:02:10,460 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7198, 1.8691, 0.9290, 1.4539, 2.1316, 1.6420, 1.5913, 1.5520], + device='cuda:3'), covar=tensor([0.0521, 0.0377, 0.0360, 0.0556, 0.0243, 0.0522, 0.0506, 0.0597], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0027, 0.0024, 0.0030, 0.0021, 0.0030, 0.0030, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 18:02:14,148 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31281.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:02:24,319 INFO [finetune.py:976] (3/7) Epoch 6, batch 2650, loss[loss=0.2027, simple_loss=0.2737, pruned_loss=0.06582, over 4890.00 frames. ], tot_loss[loss=0.212, simple_loss=0.2742, pruned_loss=0.07494, over 952187.72 frames. ], batch size: 43, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:02:24,977 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31290.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:02:55,887 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31311.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:03:18,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.837e+02 2.164e+02 2.605e+02 4.419e+02, threshold=4.327e+02, percent-clipped=1.0 +2023-04-26 18:03:30,927 INFO [finetune.py:976] (3/7) Epoch 6, batch 2700, loss[loss=0.2012, simple_loss=0.2641, pruned_loss=0.06919, over 4819.00 frames. ], tot_loss[loss=0.2101, simple_loss=0.2726, pruned_loss=0.07382, over 952672.34 frames. ], batch size: 33, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:03:42,717 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-26 18:04:01,836 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31359.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:04:14,067 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31367.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 18:04:44,998 INFO [finetune.py:976] (3/7) Epoch 6, batch 2750, loss[loss=0.2296, simple_loss=0.267, pruned_loss=0.09609, over 4239.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.27, pruned_loss=0.0731, over 955101.77 frames. ], batch size: 65, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:05:10,770 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31409.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:05:28,047 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-26 18:05:28,795 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.757e+02 2.143e+02 2.485e+02 4.889e+02, threshold=4.286e+02, percent-clipped=1.0 +2023-04-26 18:05:35,955 INFO [finetune.py:976] (3/7) Epoch 6, batch 2800, loss[loss=0.2256, simple_loss=0.285, pruned_loss=0.08312, over 4851.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2655, pruned_loss=0.07152, over 957139.31 frames. ], batch size: 44, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:05:37,319 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2750, 2.8537, 2.2243, 2.0735, 1.5796, 1.5830, 2.3102, 1.5331], + device='cuda:3'), covar=tensor([0.1702, 0.1656, 0.1600, 0.2037, 0.2798, 0.2188, 0.1152, 0.2274], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0220, 0.0176, 0.0206, 0.0212, 0.0186, 0.0168, 0.0192], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:05:57,414 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31470.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:06:09,971 INFO [finetune.py:976] (3/7) Epoch 6, batch 2850, loss[loss=0.2391, simple_loss=0.2936, pruned_loss=0.09229, over 4014.00 frames. ], tot_loss[loss=0.2024, simple_loss=0.2636, pruned_loss=0.07066, over 956058.74 frames. ], batch size: 65, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:06:10,665 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9035, 2.8709, 2.2539, 3.2900, 2.8819, 2.8774, 1.2676, 2.8264], + device='cuda:3'), covar=tensor([0.2003, 0.1682, 0.3159, 0.2752, 0.3604, 0.2189, 0.5376, 0.2619], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0221, 0.0255, 0.0315, 0.0304, 0.0256, 0.0277, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:06:11,305 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31491.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:06:36,454 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.825e+02 2.131e+02 2.409e+02 4.261e+02, threshold=4.261e+02, percent-clipped=0.0 +2023-04-26 18:06:37,217 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6764, 2.3207, 1.7390, 1.6580, 1.2328, 1.2604, 1.8888, 1.2444], + device='cuda:3'), covar=tensor([0.1680, 0.1547, 0.1561, 0.1867, 0.2548, 0.2001, 0.1098, 0.2099], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0219, 0.0176, 0.0205, 0.0211, 0.0186, 0.0167, 0.0192], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:06:40,367 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 18:06:43,782 INFO [finetune.py:976] (3/7) Epoch 6, batch 2900, loss[loss=0.2028, simple_loss=0.2765, pruned_loss=0.06451, over 4827.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2666, pruned_loss=0.07164, over 956116.63 frames. ], batch size: 40, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:06:51,779 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31552.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 18:06:52,354 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0718, 2.6140, 1.2541, 1.4488, 2.1467, 1.4404, 3.1917, 1.6624], + device='cuda:3'), covar=tensor([0.0642, 0.0839, 0.0788, 0.1128, 0.0423, 0.0891, 0.0199, 0.0608], + device='cuda:3'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0053, 0.0054, 0.0081, 0.0053], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 18:06:56,007 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31559.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:04,858 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 18:07:11,690 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31581.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:14,178 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3825, 2.3758, 2.0297, 2.1991, 2.5505, 2.1334, 3.4061, 1.8226], + device='cuda:3'), covar=tensor([0.4516, 0.2495, 0.5242, 0.3990, 0.2107, 0.2782, 0.1980, 0.4753], + device='cuda:3'), in_proj_covar=tensor([0.0350, 0.0357, 0.0439, 0.0369, 0.0394, 0.0385, 0.0389, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:07:16,959 INFO [finetune.py:976] (3/7) Epoch 6, batch 2950, loss[loss=0.1987, simple_loss=0.2497, pruned_loss=0.07383, over 4865.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2688, pruned_loss=0.07218, over 955059.46 frames. ], batch size: 31, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:07:17,672 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31590.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:28,066 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31607.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:42,551 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.797e+02 2.161e+02 2.728e+02 5.785e+02, threshold=4.321e+02, percent-clipped=2.0 +2023-04-26 18:07:43,218 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31629.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:49,139 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31638.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:49,694 INFO [finetune.py:976] (3/7) Epoch 6, batch 3000, loss[loss=0.1957, simple_loss=0.2732, pruned_loss=0.05908, over 4738.00 frames. ], tot_loss[loss=0.2095, simple_loss=0.2715, pruned_loss=0.07377, over 954412.22 frames. ], batch size: 54, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:07:49,695 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 18:07:55,994 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4972, 3.4059, 2.4925, 3.8808, 3.5418, 3.5432, 1.5440, 3.4203], + device='cuda:3'), covar=tensor([0.1495, 0.1208, 0.3186, 0.1923, 0.2714, 0.1764, 0.5571, 0.2379], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0221, 0.0256, 0.0315, 0.0304, 0.0257, 0.0278, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:08:00,217 INFO [finetune.py:1010] (3/7) Epoch 6, validation: loss=0.1565, simple_loss=0.2301, pruned_loss=0.04144, over 2265189.00 frames. +2023-04-26 18:08:00,218 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 18:08:17,856 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31667.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:08:48,664 INFO [finetune.py:976] (3/7) Epoch 6, batch 3050, loss[loss=0.1802, simple_loss=0.2518, pruned_loss=0.05431, over 4893.00 frames. ], tot_loss[loss=0.2113, simple_loss=0.2736, pruned_loss=0.07455, over 954748.80 frames. ], batch size: 32, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:09:09,619 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6013, 2.0744, 1.5511, 1.1963, 1.1956, 1.2211, 1.5750, 1.1798], + device='cuda:3'), covar=tensor([0.2058, 0.1585, 0.2092, 0.2500, 0.3084, 0.2362, 0.1379, 0.2577], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0220, 0.0176, 0.0206, 0.0211, 0.0187, 0.0168, 0.0193], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:09:10,806 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31704.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:09:20,318 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2329, 1.5574, 2.0699, 2.4524, 1.9679, 1.5288, 1.3762, 1.7699], + device='cuda:3'), covar=tensor([0.3382, 0.3902, 0.1784, 0.2937, 0.3328, 0.3145, 0.5145, 0.2980], + device='cuda:3'), in_proj_covar=tensor([0.0277, 0.0256, 0.0220, 0.0327, 0.0216, 0.0230, 0.0242, 0.0192], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:09:23,336 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31715.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:09:40,856 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.692e+02 2.096e+02 2.514e+02 4.783e+02, threshold=4.192e+02, percent-clipped=1.0 +2023-04-26 18:09:53,207 INFO [finetune.py:976] (3/7) Epoch 6, batch 3100, loss[loss=0.2035, simple_loss=0.2742, pruned_loss=0.06639, over 4909.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2714, pruned_loss=0.07365, over 955165.61 frames. ], batch size: 37, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:10:25,847 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:10:25,894 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:10:56,966 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31788.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:10:57,967 INFO [finetune.py:976] (3/7) Epoch 6, batch 3150, loss[loss=0.2132, simple_loss=0.2644, pruned_loss=0.08105, over 4861.00 frames. ], tot_loss[loss=0.2064, simple_loss=0.2683, pruned_loss=0.07226, over 956453.33 frames. ], batch size: 49, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:11:21,862 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3896, 3.4214, 2.5712, 3.9250, 3.4662, 3.4434, 1.4766, 3.3576], + device='cuda:3'), covar=tensor([0.1855, 0.1390, 0.3019, 0.2061, 0.3590, 0.1972, 0.5992, 0.2420], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0220, 0.0254, 0.0314, 0.0303, 0.0255, 0.0277, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:11:38,715 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-26 18:11:42,879 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31823.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:11:51,413 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.949e+01 1.610e+02 1.947e+02 2.415e+02 6.715e+02, threshold=3.894e+02, percent-clipped=1.0 +2023-04-26 18:12:02,485 INFO [finetune.py:976] (3/7) Epoch 6, batch 3200, loss[loss=0.2032, simple_loss=0.265, pruned_loss=0.07068, over 4848.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2646, pruned_loss=0.07056, over 956307.62 frames. ], batch size: 44, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:12:14,984 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31847.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 18:12:15,733 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-04-26 18:12:22,008 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:12:36,044 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-26 18:12:57,064 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31878.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:13:06,251 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31884.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:13:09,138 INFO [finetune.py:976] (3/7) Epoch 6, batch 3250, loss[loss=0.2696, simple_loss=0.3222, pruned_loss=0.1085, over 4742.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2668, pruned_loss=0.07195, over 952939.35 frames. ], batch size: 54, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:13:39,525 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31915.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:13:57,725 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.762e+02 2.242e+02 2.827e+02 6.116e+02, threshold=4.483e+02, percent-clipped=8.0 +2023-04-26 18:14:10,188 INFO [finetune.py:976] (3/7) Epoch 6, batch 3300, loss[loss=0.2407, simple_loss=0.3086, pruned_loss=0.08636, over 4789.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2704, pruned_loss=0.07287, over 951781.60 frames. ], batch size: 51, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:14:10,323 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31939.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:14:18,830 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31950.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:14:35,929 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31976.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:14:43,708 INFO [finetune.py:976] (3/7) Epoch 6, batch 3350, loss[loss=0.227, simple_loss=0.298, pruned_loss=0.07803, over 4727.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.272, pruned_loss=0.07344, over 953567.97 frames. ], batch size: 59, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:15:00,880 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=32011.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:15:12,029 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.734e+02 2.033e+02 2.508e+02 4.432e+02, threshold=4.066e+02, percent-clipped=0.0 +2023-04-26 18:15:24,408 INFO [finetune.py:976] (3/7) Epoch 6, batch 3400, loss[loss=0.1586, simple_loss=0.2367, pruned_loss=0.0402, over 4750.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2711, pruned_loss=0.0728, over 951356.58 frames. ], batch size: 27, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:15:43,778 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.93 vs. limit=5.0 +2023-04-26 18:15:55,467 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32060.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:15:55,490 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4433, 3.2615, 0.8912, 1.7526, 1.8903, 2.2406, 1.9618, 0.9119], + device='cuda:3'), covar=tensor([0.1437, 0.0971, 0.2059, 0.1403, 0.1068, 0.1159, 0.1554, 0.2205], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0261, 0.0146, 0.0128, 0.0138, 0.0161, 0.0123, 0.0127], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:16:03,474 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9434, 1.1936, 3.2683, 3.0139, 2.9760, 3.2335, 3.2262, 2.9123], + device='cuda:3'), covar=tensor([0.6815, 0.5099, 0.1364, 0.2139, 0.1343, 0.1673, 0.1325, 0.1631], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0304, 0.0411, 0.0415, 0.0350, 0.0407, 0.0314, 0.0371], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:16:04,716 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32065.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:16:20,374 INFO [finetune.py:976] (3/7) Epoch 6, batch 3450, loss[loss=0.2329, simple_loss=0.2855, pruned_loss=0.09017, over 4859.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2702, pruned_loss=0.07203, over 951809.49 frames. ], batch size: 31, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:16:37,417 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32113.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:16:47,433 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.699e+02 2.063e+02 2.455e+02 6.391e+02, threshold=4.126e+02, percent-clipped=3.0 +2023-04-26 18:16:54,107 INFO [finetune.py:976] (3/7) Epoch 6, batch 3500, loss[loss=0.2152, simple_loss=0.2678, pruned_loss=0.08134, over 4892.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2692, pruned_loss=0.07218, over 955137.71 frames. ], batch size: 32, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:16:55,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1530, 1.3184, 1.4479, 1.6232, 1.5451, 1.1882, 0.9282, 1.3725], + device='cuda:3'), covar=tensor([0.0961, 0.1169, 0.0805, 0.0584, 0.0682, 0.0906, 0.0931, 0.0668], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0206, 0.0181, 0.0179, 0.0180, 0.0195, 0.0165, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:16:57,307 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32144.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:16:59,156 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32147.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 18:17:05,081 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5023, 1.2877, 4.2297, 3.9622, 3.6924, 3.9898, 3.8756, 3.7344], + device='cuda:3'), covar=tensor([0.6638, 0.5892, 0.0888, 0.1496, 0.1028, 0.1382, 0.1837, 0.1343], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0305, 0.0412, 0.0416, 0.0352, 0.0408, 0.0315, 0.0371], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:17:05,985 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-26 18:17:21,486 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32179.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:17:33,407 INFO [finetune.py:976] (3/7) Epoch 6, batch 3550, loss[loss=0.1855, simple_loss=0.2498, pruned_loss=0.06054, over 4901.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2659, pruned_loss=0.07103, over 954762.98 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:17:41,659 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7884, 2.3582, 1.8181, 1.4616, 1.2852, 1.3500, 1.9835, 1.2134], + device='cuda:3'), covar=tensor([0.1830, 0.1666, 0.1743, 0.2295, 0.2720, 0.2162, 0.1194, 0.2377], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0220, 0.0177, 0.0207, 0.0212, 0.0187, 0.0168, 0.0194], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:17:43,432 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32195.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 18:17:45,964 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0844, 1.4949, 1.2944, 1.7928, 1.5588, 1.8282, 1.3762, 3.0373], + device='cuda:3'), covar=tensor([0.0705, 0.0772, 0.0806, 0.1193, 0.0640, 0.0518, 0.0750, 0.0192], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:18:29,258 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.692e+02 2.038e+02 2.540e+02 1.815e+03, threshold=4.076e+02, percent-clipped=3.0 +2023-04-26 18:18:39,219 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32234.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:18:47,838 INFO [finetune.py:976] (3/7) Epoch 6, batch 3600, loss[loss=0.2573, simple_loss=0.298, pruned_loss=0.1083, over 4896.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2624, pruned_loss=0.06949, over 955287.46 frames. ], batch size: 32, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:19:25,146 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32271.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:19:55,460 INFO [finetune.py:976] (3/7) Epoch 6, batch 3650, loss[loss=0.2214, simple_loss=0.2898, pruned_loss=0.07648, over 4856.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2671, pruned_loss=0.07194, over 954982.23 frames. ], batch size: 44, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:20:16,834 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32306.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:20:43,392 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.808e+02 2.117e+02 2.519e+02 3.732e+02, threshold=4.235e+02, percent-clipped=0.0 +2023-04-26 18:21:03,305 INFO [finetune.py:976] (3/7) Epoch 6, batch 3700, loss[loss=0.1683, simple_loss=0.2428, pruned_loss=0.04694, over 4799.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2695, pruned_loss=0.07234, over 955726.71 frames. ], batch size: 29, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:21:26,772 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32360.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:22:04,004 INFO [finetune.py:976] (3/7) Epoch 6, batch 3750, loss[loss=0.1774, simple_loss=0.2266, pruned_loss=0.06409, over 4085.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2703, pruned_loss=0.07249, over 954541.08 frames. ], batch size: 17, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:22:04,836 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-26 18:22:21,627 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32408.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:22:23,515 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2270, 1.5420, 1.3405, 1.4339, 1.3429, 1.2493, 1.3149, 1.1133], + device='cuda:3'), covar=tensor([0.1935, 0.1583, 0.1032, 0.1241, 0.3539, 0.1522, 0.1869, 0.2365], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0327, 0.0237, 0.0299, 0.0321, 0.0280, 0.0267, 0.0291], + device='cuda:3'), out_proj_covar=tensor([1.2426e-04, 1.3260e-04, 9.5994e-05, 1.2007e-04, 1.3219e-04, 1.1341e-04, + 1.0959e-04, 1.1697e-04], device='cuda:3') +2023-04-26 18:22:35,218 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.756e+02 2.047e+02 2.329e+02 4.527e+02, threshold=4.095e+02, percent-clipped=1.0 +2023-04-26 18:22:40,990 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 18:22:43,808 INFO [finetune.py:976] (3/7) Epoch 6, batch 3800, loss[loss=0.2254, simple_loss=0.2814, pruned_loss=0.08472, over 4913.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2729, pruned_loss=0.07385, over 955929.89 frames. ], batch size: 37, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:22:53,227 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32444.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:23:08,424 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-26 18:23:21,570 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32479.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:23:33,783 INFO [finetune.py:976] (3/7) Epoch 6, batch 3850, loss[loss=0.1789, simple_loss=0.2447, pruned_loss=0.05655, over 4891.00 frames. ], tot_loss[loss=0.2085, simple_loss=0.2713, pruned_loss=0.07289, over 956315.15 frames. ], batch size: 35, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:23:41,607 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32492.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:23:41,657 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2514, 1.7414, 1.5395, 1.9753, 1.7767, 2.1531, 1.5137, 3.8647], + device='cuda:3'), covar=tensor([0.0657, 0.0712, 0.0791, 0.1132, 0.0607, 0.0527, 0.0725, 0.0136], + device='cuda:3'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0040, 0.0040, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:24:04,918 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5961, 1.6633, 1.8123, 2.0688, 2.0220, 1.5808, 1.2424, 1.7574], + device='cuda:3'), covar=tensor([0.0920, 0.1087, 0.0726, 0.0622, 0.0621, 0.0961, 0.0983, 0.0671], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0207, 0.0181, 0.0179, 0.0180, 0.0195, 0.0166, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:24:18,831 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32527.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:24:19,366 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.258e+02 1.697e+02 2.014e+02 2.514e+02 4.116e+02, threshold=4.029e+02, percent-clipped=1.0 +2023-04-26 18:24:24,615 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32534.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:24:26,364 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4441, 1.2665, 1.7359, 1.7188, 1.3314, 1.1246, 1.4174, 0.8861], + device='cuda:3'), covar=tensor([0.0759, 0.0984, 0.0563, 0.0827, 0.1069, 0.1466, 0.0835, 0.0997], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0074, 0.0073, 0.0067, 0.0077, 0.0094, 0.0080, 0.0076], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:24:28,028 INFO [finetune.py:976] (3/7) Epoch 6, batch 3900, loss[loss=0.2224, simple_loss=0.2771, pruned_loss=0.08383, over 4940.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2677, pruned_loss=0.07154, over 956723.55 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:25:00,094 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 18:25:09,830 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32571.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:25:22,277 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32582.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:25:33,445 INFO [finetune.py:976] (3/7) Epoch 6, batch 3950, loss[loss=0.2252, simple_loss=0.2734, pruned_loss=0.08848, over 4339.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.2642, pruned_loss=0.07007, over 956772.07 frames. ], batch size: 19, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:25:50,983 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32606.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:25:58,796 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32619.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:26:10,336 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.669e+02 1.908e+02 2.385e+02 4.552e+02, threshold=3.816e+02, percent-clipped=3.0 +2023-04-26 18:26:10,967 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7306, 2.0713, 1.7639, 1.9814, 1.6161, 1.6722, 1.5996, 1.5264], + device='cuda:3'), covar=tensor([0.2046, 0.1423, 0.0986, 0.1167, 0.3248, 0.1373, 0.2052, 0.2547], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0328, 0.0238, 0.0300, 0.0322, 0.0281, 0.0268, 0.0292], + device='cuda:3'), out_proj_covar=tensor([1.2423e-04, 1.3327e-04, 9.6473e-05, 1.2069e-04, 1.3270e-04, 1.1382e-04, + 1.1015e-04, 1.1736e-04], device='cuda:3') +2023-04-26 18:26:22,976 INFO [finetune.py:976] (3/7) Epoch 6, batch 4000, loss[loss=0.1767, simple_loss=0.2429, pruned_loss=0.05528, over 4903.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.2636, pruned_loss=0.07033, over 956701.79 frames. ], batch size: 43, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:26:33,665 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.55 vs. limit=5.0 +2023-04-26 18:26:51,940 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:27:19,404 INFO [finetune.py:976] (3/7) Epoch 6, batch 4050, loss[loss=0.2244, simple_loss=0.2976, pruned_loss=0.07563, over 4916.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2684, pruned_loss=0.07297, over 954380.80 frames. ], batch size: 42, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:27:30,496 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2238, 2.5797, 0.8253, 1.5343, 1.5845, 1.8730, 1.6778, 0.8208], + device='cuda:3'), covar=tensor([0.1519, 0.1145, 0.1897, 0.1393, 0.1212, 0.1058, 0.1651, 0.1740], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0260, 0.0146, 0.0128, 0.0138, 0.0160, 0.0123, 0.0127], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:27:46,364 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.872e+02 2.266e+02 2.687e+02 3.966e+02, threshold=4.532e+02, percent-clipped=2.0 +2023-04-26 18:27:52,930 INFO [finetune.py:976] (3/7) Epoch 6, batch 4100, loss[loss=0.2482, simple_loss=0.3124, pruned_loss=0.09203, over 4837.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.2704, pruned_loss=0.07365, over 952581.61 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:28:13,071 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1651, 2.3115, 2.6623, 2.8789, 2.5240, 2.2176, 1.7997, 2.2959], + device='cuda:3'), covar=tensor([0.1045, 0.0939, 0.0554, 0.0657, 0.0749, 0.1117, 0.1059, 0.0697], + device='cuda:3'), in_proj_covar=tensor([0.0201, 0.0205, 0.0180, 0.0178, 0.0179, 0.0194, 0.0165, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:28:26,745 INFO [finetune.py:976] (3/7) Epoch 6, batch 4150, loss[loss=0.1517, simple_loss=0.2173, pruned_loss=0.04303, over 4728.00 frames. ], tot_loss[loss=0.2113, simple_loss=0.2728, pruned_loss=0.07493, over 951657.62 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:29:05,515 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.362e+02 1.802e+02 2.275e+02 2.602e+02 5.008e+02, threshold=4.549e+02, percent-clipped=2.0 +2023-04-26 18:29:12,122 INFO [finetune.py:976] (3/7) Epoch 6, batch 4200, loss[loss=0.1923, simple_loss=0.2562, pruned_loss=0.06421, over 4894.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2728, pruned_loss=0.07455, over 951386.54 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:29:33,421 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-26 18:29:45,429 INFO [finetune.py:976] (3/7) Epoch 6, batch 4250, loss[loss=0.249, simple_loss=0.2945, pruned_loss=0.1017, over 4823.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.27, pruned_loss=0.07334, over 949219.30 frames. ], batch size: 40, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:29:50,264 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 18:30:13,034 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.159e+01 1.639e+02 1.929e+02 2.275e+02 6.095e+02, threshold=3.858e+02, percent-clipped=1.0 +2023-04-26 18:30:19,138 INFO [finetune.py:976] (3/7) Epoch 6, batch 4300, loss[loss=0.1875, simple_loss=0.2501, pruned_loss=0.0625, over 4904.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2667, pruned_loss=0.07175, over 952382.68 frames. ], batch size: 46, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:30:27,886 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-04-26 18:30:47,972 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=32966.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:31:19,601 INFO [finetune.py:976] (3/7) Epoch 6, batch 4350, loss[loss=0.1325, simple_loss=0.2017, pruned_loss=0.03159, over 4834.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2625, pruned_loss=0.06998, over 954820.83 frames. ], batch size: 25, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:32:12,348 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9290, 1.8498, 2.2015, 2.3830, 1.7489, 1.5125, 1.9432, 1.1051], + device='cuda:3'), covar=tensor([0.0868, 0.1037, 0.0716, 0.1094, 0.1056, 0.1384, 0.1018, 0.1230], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0076, 0.0074, 0.0068, 0.0079, 0.0096, 0.0082, 0.0078], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:32:12,355 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33027.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:32:13,431 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.910e+01 1.721e+02 1.970e+02 2.465e+02 4.001e+02, threshold=3.941e+02, percent-clipped=3.0 +2023-04-26 18:32:13,997 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-26 18:32:24,805 INFO [finetune.py:976] (3/7) Epoch 6, batch 4400, loss[loss=0.2436, simple_loss=0.3025, pruned_loss=0.0923, over 4179.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2633, pruned_loss=0.07032, over 954600.89 frames. ], batch size: 65, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:32:27,394 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 18:32:45,510 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9263, 2.3274, 2.0224, 2.2200, 1.7361, 1.9828, 2.0875, 1.7572], + device='cuda:3'), covar=tensor([0.1899, 0.1193, 0.0794, 0.1141, 0.2749, 0.1039, 0.1843, 0.1959], + device='cuda:3'), in_proj_covar=tensor([0.0302, 0.0322, 0.0234, 0.0296, 0.0317, 0.0276, 0.0264, 0.0287], + device='cuda:3'), out_proj_covar=tensor([1.2259e-04, 1.3057e-04, 9.4818e-05, 1.1896e-04, 1.3055e-04, 1.1138e-04, + 1.0834e-04, 1.1563e-04], device='cuda:3') +2023-04-26 18:33:06,922 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33069.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:33:11,030 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4754, 3.4640, 0.9493, 1.9116, 1.8885, 2.4916, 1.9637, 0.9202], + device='cuda:3'), covar=tensor([0.1407, 0.0928, 0.1909, 0.1292, 0.1076, 0.0980, 0.1602, 0.2286], + device='cuda:3'), in_proj_covar=tensor([0.0122, 0.0260, 0.0145, 0.0127, 0.0138, 0.0160, 0.0123, 0.0126], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:33:17,894 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2547, 1.6710, 1.5560, 2.1522, 1.7901, 2.0340, 1.5902, 4.5174], + device='cuda:3'), covar=tensor([0.0663, 0.0820, 0.0906, 0.1225, 0.0704, 0.0603, 0.0825, 0.0178], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:33:20,948 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9880, 1.3166, 1.1579, 1.6047, 1.3978, 1.5559, 1.2643, 2.5003], + device='cuda:3'), covar=tensor([0.0680, 0.0839, 0.0942, 0.1249, 0.0724, 0.0582, 0.0841, 0.0261], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:33:29,651 INFO [finetune.py:976] (3/7) Epoch 6, batch 4450, loss[loss=0.2033, simple_loss=0.271, pruned_loss=0.06779, over 4802.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2672, pruned_loss=0.07188, over 954549.77 frames. ], batch size: 45, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:33:43,450 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1512, 1.5385, 1.4221, 1.9545, 1.6587, 1.8462, 1.4837, 3.1520], + device='cuda:3'), covar=tensor([0.0705, 0.0817, 0.0854, 0.1125, 0.0677, 0.0519, 0.0774, 0.0189], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:33:54,498 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0577, 2.6535, 2.0380, 1.9095, 1.5194, 1.4816, 2.1278, 1.4367], + device='cuda:3'), covar=tensor([0.1984, 0.1758, 0.1740, 0.2356, 0.2790, 0.2187, 0.1244, 0.2509], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0218, 0.0174, 0.0204, 0.0209, 0.0185, 0.0166, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 18:34:08,076 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3152, 3.3199, 1.0186, 1.8748, 1.7758, 2.3872, 1.8990, 1.0334], + device='cuda:3'), covar=tensor([0.1448, 0.0951, 0.1927, 0.1345, 0.1179, 0.1080, 0.1542, 0.2156], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0259, 0.0144, 0.0127, 0.0138, 0.0159, 0.0123, 0.0126], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:34:12,568 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.742e+02 2.062e+02 2.633e+02 5.031e+02, threshold=4.124e+02, percent-clipped=5.0 +2023-04-26 18:34:13,337 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:34:18,686 INFO [finetune.py:976] (3/7) Epoch 6, batch 4500, loss[loss=0.1671, simple_loss=0.2244, pruned_loss=0.05489, over 4694.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2695, pruned_loss=0.07269, over 957056.98 frames. ], batch size: 23, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:34:23,814 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 18:34:29,024 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33155.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:34:49,147 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 18:34:52,378 INFO [finetune.py:976] (3/7) Epoch 6, batch 4550, loss[loss=0.2615, simple_loss=0.3055, pruned_loss=0.1087, over 4146.00 frames. ], tot_loss[loss=0.2089, simple_loss=0.2712, pruned_loss=0.0733, over 954997.14 frames. ], batch size: 66, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:34:54,335 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9678, 2.4365, 2.0365, 2.2601, 1.8024, 1.9590, 1.9738, 1.7129], + device='cuda:3'), covar=tensor([0.1882, 0.1334, 0.0850, 0.1068, 0.3024, 0.1259, 0.1850, 0.2295], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0323, 0.0234, 0.0297, 0.0318, 0.0276, 0.0264, 0.0288], + device='cuda:3'), out_proj_covar=tensor([1.2322e-04, 1.3087e-04, 9.4978e-05, 1.1919e-04, 1.3076e-04, 1.1165e-04, + 1.0853e-04, 1.1593e-04], device='cuda:3') +2023-04-26 18:34:56,768 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2844, 1.5238, 1.2470, 1.4251, 1.3331, 1.1309, 1.3454, 1.1272], + device='cuda:3'), covar=tensor([0.1958, 0.1452, 0.1166, 0.1349, 0.3879, 0.1513, 0.1981, 0.2467], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0323, 0.0234, 0.0297, 0.0318, 0.0276, 0.0264, 0.0288], + device='cuda:3'), out_proj_covar=tensor([1.2321e-04, 1.3086e-04, 9.4997e-05, 1.1919e-04, 1.3079e-04, 1.1165e-04, + 1.0854e-04, 1.1593e-04], device='cuda:3') +2023-04-26 18:35:09,504 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33216.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:35:18,794 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.680e+02 1.969e+02 2.447e+02 4.378e+02, threshold=3.937e+02, percent-clipped=2.0 +2023-04-26 18:35:25,831 INFO [finetune.py:976] (3/7) Epoch 6, batch 4600, loss[loss=0.1677, simple_loss=0.2331, pruned_loss=0.05116, over 4780.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2703, pruned_loss=0.07237, over 955861.95 frames. ], batch size: 29, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:35:36,856 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3860, 2.3870, 2.0214, 2.1746, 2.5059, 1.9979, 3.3575, 1.8163], + device='cuda:3'), covar=tensor([0.4845, 0.2538, 0.5195, 0.4007, 0.2223, 0.3089, 0.2057, 0.4803], + device='cuda:3'), in_proj_covar=tensor([0.0352, 0.0358, 0.0441, 0.0368, 0.0397, 0.0387, 0.0390, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:35:59,783 INFO [finetune.py:976] (3/7) Epoch 6, batch 4650, loss[loss=0.2198, simple_loss=0.2878, pruned_loss=0.07593, over 4891.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2676, pruned_loss=0.07158, over 956562.44 frames. ], batch size: 35, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:36:01,691 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6149, 3.5085, 2.6865, 4.1563, 3.6463, 3.6280, 1.6325, 3.5465], + device='cuda:3'), covar=tensor([0.1864, 0.1334, 0.3478, 0.2012, 0.3656, 0.1844, 0.5811, 0.2675], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0221, 0.0256, 0.0314, 0.0305, 0.0257, 0.0277, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:36:06,634 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4327, 2.2051, 2.5134, 2.7192, 2.4563, 2.1720, 2.3358, 2.3554], + device='cuda:3'), covar=tensor([0.7909, 1.0593, 1.1845, 1.1792, 0.9469, 1.4833, 1.4452, 1.1043], + device='cuda:3'), in_proj_covar=tensor([0.0414, 0.0440, 0.0525, 0.0545, 0.0443, 0.0462, 0.0474, 0.0473], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:36:20,281 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33322.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:36:22,851 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-04-26 18:36:25,474 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.206e+02 1.668e+02 2.007e+02 2.366e+02 4.187e+02, threshold=4.014e+02, percent-clipped=2.0 +2023-04-26 18:36:32,593 INFO [finetune.py:976] (3/7) Epoch 6, batch 4700, loss[loss=0.1573, simple_loss=0.2319, pruned_loss=0.04139, over 4753.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2643, pruned_loss=0.06989, over 956205.11 frames. ], batch size: 27, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:37:20,897 INFO [finetune.py:976] (3/7) Epoch 6, batch 4750, loss[loss=0.1698, simple_loss=0.2366, pruned_loss=0.05147, over 4894.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2612, pruned_loss=0.06891, over 956833.71 frames. ], batch size: 32, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:37:41,120 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1397, 0.5132, 0.8955, 0.6735, 1.2225, 0.8882, 0.7161, 0.9512], + device='cuda:3'), covar=tensor([0.1771, 0.1972, 0.2422, 0.1802, 0.1118, 0.1933, 0.1977, 0.2244], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0332, 0.0351, 0.0304, 0.0341, 0.0332, 0.0309, 0.0353], + device='cuda:3'), out_proj_covar=tensor([6.6268e-05, 7.0753e-05, 7.6057e-05, 6.3176e-05, 7.1836e-05, 7.1814e-05, + 6.6924e-05, 7.5977e-05], device='cuda:3') +2023-04-26 18:37:51,438 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 18:38:06,058 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33425.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:38:14,209 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.777e+02 2.013e+02 2.553e+02 6.681e+02, threshold=4.026e+02, percent-clipped=2.0 +2023-04-26 18:38:27,280 INFO [finetune.py:976] (3/7) Epoch 6, batch 4800, loss[loss=0.1837, simple_loss=0.2549, pruned_loss=0.0562, over 4776.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2648, pruned_loss=0.07067, over 952942.95 frames. ], batch size: 28, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:39:01,136 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33472.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:39:12,310 INFO [finetune.py:976] (3/7) Epoch 6, batch 4850, loss[loss=0.2003, simple_loss=0.2388, pruned_loss=0.08086, over 3967.00 frames. ], tot_loss[loss=0.2063, simple_loss=0.2687, pruned_loss=0.07197, over 952130.36 frames. ], batch size: 17, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:39:13,542 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1486, 1.4883, 2.0335, 2.3928, 1.9048, 1.5017, 1.1738, 1.7776], + device='cuda:3'), covar=tensor([0.4129, 0.4477, 0.1825, 0.3737, 0.3839, 0.3388, 0.5672, 0.3237], + device='cuda:3'), in_proj_covar=tensor([0.0278, 0.0255, 0.0219, 0.0326, 0.0215, 0.0228, 0.0240, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:39:19,957 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4469, 2.5685, 1.1659, 1.4502, 2.0290, 1.4982, 3.3035, 1.8068], + device='cuda:3'), covar=tensor([0.0557, 0.0747, 0.0857, 0.1213, 0.0483, 0.0913, 0.0355, 0.0599], + device='cuda:3'), in_proj_covar=tensor([0.0053, 0.0069, 0.0052, 0.0048, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 18:39:27,145 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33511.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:39:38,470 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.796e+02 2.126e+02 2.490e+02 5.374e+02, threshold=4.252e+02, percent-clipped=2.0 +2023-04-26 18:39:41,041 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33533.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:39:44,984 INFO [finetune.py:976] (3/7) Epoch 6, batch 4900, loss[loss=0.2401, simple_loss=0.3106, pruned_loss=0.08481, over 4805.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.27, pruned_loss=0.07258, over 951078.10 frames. ], batch size: 38, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:40:11,225 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4016, 3.2956, 1.0428, 1.8379, 1.8387, 2.3251, 1.9006, 0.9641], + device='cuda:3'), covar=tensor([0.1502, 0.1134, 0.2018, 0.1358, 0.1166, 0.1112, 0.1694, 0.2002], + device='cuda:3'), in_proj_covar=tensor([0.0121, 0.0259, 0.0144, 0.0126, 0.0137, 0.0159, 0.0122, 0.0125], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:40:18,322 INFO [finetune.py:976] (3/7) Epoch 6, batch 4950, loss[loss=0.1907, simple_loss=0.2618, pruned_loss=0.05978, over 4892.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2707, pruned_loss=0.07199, over 954144.48 frames. ], batch size: 43, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:40:27,252 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-26 18:40:27,387 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8435, 3.7989, 2.8240, 4.4176, 3.8229, 3.8023, 1.8159, 3.7395], + device='cuda:3'), covar=tensor([0.1542, 0.1108, 0.3246, 0.1454, 0.2508, 0.1712, 0.5138, 0.2222], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0218, 0.0253, 0.0311, 0.0301, 0.0254, 0.0274, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:40:40,489 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33622.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:40:44,643 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.256e+02 1.697e+02 1.943e+02 2.276e+02 3.620e+02, threshold=3.887e+02, percent-clipped=0.0 +2023-04-26 18:40:51,673 INFO [finetune.py:976] (3/7) Epoch 6, batch 5000, loss[loss=0.1602, simple_loss=0.2279, pruned_loss=0.04628, over 4747.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2686, pruned_loss=0.07099, over 955499.13 frames. ], batch size: 28, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:41:24,493 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33670.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:41:28,777 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33677.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:41:32,990 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 18:41:36,417 INFO [finetune.py:976] (3/7) Epoch 6, batch 5050, loss[loss=0.2171, simple_loss=0.2728, pruned_loss=0.08068, over 4867.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2674, pruned_loss=0.07113, over 955044.37 frames. ], batch size: 31, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:42:01,201 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33725.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:42:03,504 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.184e+02 1.719e+02 2.034e+02 2.333e+02 5.936e+02, threshold=4.068e+02, percent-clipped=3.0 +2023-04-26 18:42:14,950 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33738.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:42:15,448 INFO [finetune.py:976] (3/7) Epoch 6, batch 5100, loss[loss=0.2485, simple_loss=0.3027, pruned_loss=0.09716, over 4837.00 frames. ], tot_loss[loss=0.2017, simple_loss=0.2637, pruned_loss=0.06982, over 956752.28 frames. ], batch size: 47, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:42:59,652 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-26 18:43:00,673 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33773.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:43:21,824 INFO [finetune.py:976] (3/7) Epoch 6, batch 5150, loss[loss=0.2092, simple_loss=0.2679, pruned_loss=0.07529, over 4836.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2635, pruned_loss=0.06967, over 955652.04 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:43:33,824 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0233, 0.9134, 1.2424, 1.1277, 0.9767, 0.8378, 1.0071, 0.4968], + device='cuda:3'), covar=tensor([0.0637, 0.0785, 0.0656, 0.0743, 0.0796, 0.1486, 0.0594, 0.1087], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0075, 0.0074, 0.0068, 0.0079, 0.0096, 0.0081, 0.0078], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:43:40,817 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 18:43:42,911 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33811.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:43:44,071 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5828, 0.9617, 1.5529, 1.9606, 1.6615, 1.5062, 1.5271, 1.5987], + device='cuda:3'), covar=tensor([0.7743, 1.0683, 1.0827, 1.1865, 0.9787, 1.2796, 1.2273, 1.0733], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0437, 0.0521, 0.0541, 0.0439, 0.0459, 0.0473, 0.0471], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:43:54,702 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33828.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:43:55,209 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.192e+02 1.912e+02 2.137e+02 2.570e+02 6.415e+02, threshold=4.274e+02, percent-clipped=3.0 +2023-04-26 18:44:07,683 INFO [finetune.py:976] (3/7) Epoch 6, batch 5200, loss[loss=0.2228, simple_loss=0.2798, pruned_loss=0.08288, over 4896.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2669, pruned_loss=0.0711, over 954160.15 frames. ], batch size: 35, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:44:38,045 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:45:14,301 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0913, 2.6255, 2.2382, 2.5076, 1.9797, 2.2411, 2.4534, 1.9209], + device='cuda:3'), covar=tensor([0.2030, 0.0971, 0.0739, 0.0943, 0.2634, 0.1106, 0.1634, 0.2299], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0324, 0.0237, 0.0297, 0.0318, 0.0279, 0.0265, 0.0290], + device='cuda:3'), out_proj_covar=tensor([1.2448e-04, 1.3154e-04, 9.6164e-05, 1.1910e-04, 1.3106e-04, 1.1244e-04, + 1.0890e-04, 1.1695e-04], device='cuda:3') +2023-04-26 18:45:14,783 INFO [finetune.py:976] (3/7) Epoch 6, batch 5250, loss[loss=0.2103, simple_loss=0.2706, pruned_loss=0.07497, over 4100.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2693, pruned_loss=0.07202, over 952485.85 frames. ], batch size: 65, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:45:25,945 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1035, 1.2757, 1.5487, 1.7248, 1.6181, 1.7788, 1.5844, 1.6274], + device='cuda:3'), covar=tensor([0.5257, 0.8602, 0.7394, 0.7058, 0.8323, 1.1997, 0.8894, 0.7639], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0392, 0.0318, 0.0327, 0.0344, 0.0410, 0.0373, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:45:33,582 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 18:45:58,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.252e+02 1.714e+02 2.064e+02 2.470e+02 4.995e+02, threshold=4.129e+02, percent-clipped=2.0 +2023-04-26 18:46:01,885 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-26 18:46:04,411 INFO [finetune.py:976] (3/7) Epoch 6, batch 5300, loss[loss=0.217, simple_loss=0.2826, pruned_loss=0.07572, over 4823.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2711, pruned_loss=0.07271, over 953621.83 frames. ], batch size: 39, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:46:04,528 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33939.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:46:38,099 INFO [finetune.py:976] (3/7) Epoch 6, batch 5350, loss[loss=0.1797, simple_loss=0.2479, pruned_loss=0.05579, over 4918.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2705, pruned_loss=0.07199, over 954409.92 frames. ], batch size: 38, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:46:43,528 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3035, 2.1254, 2.5655, 2.9351, 2.7267, 2.3013, 1.8096, 2.2988], + device='cuda:3'), covar=tensor([0.0854, 0.0988, 0.0532, 0.0590, 0.0583, 0.0862, 0.0929, 0.0720], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0207, 0.0179, 0.0179, 0.0180, 0.0195, 0.0165, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:46:46,589 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34000.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:47:06,989 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.964e+01 1.645e+02 1.895e+02 2.308e+02 4.781e+02, threshold=3.789e+02, percent-clipped=2.0 +2023-04-26 18:47:09,501 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34033.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:47:13,058 INFO [finetune.py:976] (3/7) Epoch 6, batch 5400, loss[loss=0.1653, simple_loss=0.2288, pruned_loss=0.05086, over 4455.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2685, pruned_loss=0.07161, over 956552.37 frames. ], batch size: 19, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:47:16,180 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34044.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:47:26,170 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1122, 1.6159, 1.3983, 1.8124, 1.6301, 1.8495, 1.4652, 3.4251], + device='cuda:3'), covar=tensor([0.0674, 0.0766, 0.0790, 0.1157, 0.0627, 0.0565, 0.0697, 0.0164], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:47:33,754 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-26 18:47:34,034 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 18:47:46,754 INFO [finetune.py:976] (3/7) Epoch 6, batch 5450, loss[loss=0.1336, simple_loss=0.2114, pruned_loss=0.02791, over 4770.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2658, pruned_loss=0.07021, over 957718.08 frames. ], batch size: 28, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:47:57,074 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34105.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:48:12,467 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34128.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:48:12,947 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.622e+02 2.010e+02 2.382e+02 4.007e+02, threshold=4.020e+02, percent-clipped=1.0 +2023-04-26 18:48:25,400 INFO [finetune.py:976] (3/7) Epoch 6, batch 5500, loss[loss=0.2119, simple_loss=0.2671, pruned_loss=0.07832, over 4768.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2624, pruned_loss=0.06935, over 957162.39 frames. ], batch size: 23, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:48:29,206 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34145.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:48:35,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0602, 1.6396, 1.4491, 1.8736, 1.6484, 2.2027, 1.4660, 3.7277], + device='cuda:3'), covar=tensor([0.0690, 0.0789, 0.0835, 0.1209, 0.0660, 0.0488, 0.0768, 0.0151], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0041, 0.0040, 0.0039, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:49:10,833 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34176.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:49:30,902 INFO [finetune.py:976] (3/7) Epoch 6, batch 5550, loss[loss=0.2512, simple_loss=0.3114, pruned_loss=0.09549, over 4812.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2636, pruned_loss=0.06951, over 955181.18 frames. ], batch size: 45, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:49:52,205 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34206.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:50:24,792 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.272e+02 1.890e+02 2.190e+02 2.677e+02 6.143e+02, threshold=4.380e+02, percent-clipped=3.0 +2023-04-26 18:50:36,986 INFO [finetune.py:976] (3/7) Epoch 6, batch 5600, loss[loss=0.2128, simple_loss=0.2697, pruned_loss=0.07795, over 4844.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2683, pruned_loss=0.07121, over 953135.85 frames. ], batch size: 47, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:50:40,051 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5715, 2.0002, 1.7317, 1.9620, 1.5415, 1.6912, 1.6740, 1.4828], + device='cuda:3'), covar=tensor([0.2136, 0.1429, 0.0915, 0.1107, 0.3124, 0.1248, 0.1888, 0.2456], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0320, 0.0234, 0.0294, 0.0315, 0.0276, 0.0263, 0.0286], + device='cuda:3'), out_proj_covar=tensor([1.2319e-04, 1.2982e-04, 9.4952e-05, 1.1816e-04, 1.2966e-04, 1.1142e-04, + 1.0782e-04, 1.1493e-04], device='cuda:3') +2023-04-26 18:51:25,429 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7882, 1.4115, 1.3547, 1.5744, 1.9920, 1.6456, 1.4312, 1.2896], + device='cuda:3'), covar=tensor([0.1836, 0.1664, 0.2272, 0.1524, 0.1022, 0.1860, 0.2533, 0.2108], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0331, 0.0351, 0.0303, 0.0341, 0.0330, 0.0309, 0.0353], + device='cuda:3'), out_proj_covar=tensor([6.6163e-05, 7.0642e-05, 7.6127e-05, 6.2998e-05, 7.1786e-05, 7.1280e-05, + 6.6834e-05, 7.6126e-05], device='cuda:3') +2023-04-26 18:51:28,270 INFO [finetune.py:976] (3/7) Epoch 6, batch 5650, loss[loss=0.1919, simple_loss=0.265, pruned_loss=0.05935, over 4898.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.2727, pruned_loss=0.0724, over 952169.08 frames. ], batch size: 35, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:51:31,899 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34295.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:51:38,386 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7593, 1.6383, 0.7831, 1.4417, 1.5923, 1.6501, 1.5256, 1.5520], + device='cuda:3'), covar=tensor([0.0505, 0.0364, 0.0385, 0.0538, 0.0269, 0.0492, 0.0498, 0.0544], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 18:52:00,816 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-26 18:52:03,001 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.702e+02 2.093e+02 2.545e+02 5.557e+02, threshold=4.186e+02, percent-clipped=1.0 +2023-04-26 18:52:05,491 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34333.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:52:09,024 INFO [finetune.py:976] (3/7) Epoch 6, batch 5700, loss[loss=0.1596, simple_loss=0.2171, pruned_loss=0.05104, over 4258.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.268, pruned_loss=0.07208, over 932751.36 frames. ], batch size: 18, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:52:19,835 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6219, 1.3140, 1.7057, 1.7473, 1.3914, 1.2637, 1.3897, 0.9340], + device='cuda:3'), covar=tensor([0.0496, 0.1098, 0.0755, 0.0788, 0.0960, 0.1200, 0.0760, 0.0967], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0075, 0.0073, 0.0068, 0.0078, 0.0096, 0.0081, 0.0077], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:52:22,708 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1513, 1.2537, 1.2974, 1.4724, 1.4297, 1.1819, 0.9731, 1.2846], + device='cuda:3'), covar=tensor([0.0977, 0.1374, 0.0829, 0.0732, 0.0766, 0.0937, 0.1005, 0.0723], + device='cuda:3'), in_proj_covar=tensor([0.0205, 0.0209, 0.0183, 0.0181, 0.0181, 0.0197, 0.0167, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:52:39,656 INFO [finetune.py:976] (3/7) Epoch 7, batch 0, loss[loss=0.2551, simple_loss=0.3142, pruned_loss=0.09795, over 4813.00 frames. ], tot_loss[loss=0.2551, simple_loss=0.3142, pruned_loss=0.09795, over 4813.00 frames. ], batch size: 55, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:52:39,656 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 18:52:49,036 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6268, 2.1931, 1.8228, 2.0918, 1.7658, 1.7748, 1.7522, 1.4935], + device='cuda:3'), covar=tensor([0.2460, 0.1390, 0.1040, 0.1363, 0.3561, 0.1531, 0.2225, 0.2735], + device='cuda:3'), in_proj_covar=tensor([0.0304, 0.0321, 0.0235, 0.0295, 0.0315, 0.0277, 0.0263, 0.0286], + device='cuda:3'), out_proj_covar=tensor([1.2337e-04, 1.3009e-04, 9.5201e-05, 1.1844e-04, 1.2973e-04, 1.1183e-04, + 1.0785e-04, 1.1517e-04], device='cuda:3') +2023-04-26 18:52:50,219 INFO [finetune.py:1010] (3/7) Epoch 7, validation: loss=0.1579, simple_loss=0.2317, pruned_loss=0.04207, over 2265189.00 frames. +2023-04-26 18:52:50,220 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 18:52:56,989 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9524, 1.8212, 2.2641, 2.4132, 1.7865, 1.4938, 1.9333, 1.0311], + device='cuda:3'), covar=tensor([0.0711, 0.0986, 0.0577, 0.0737, 0.0949, 0.1461, 0.0906, 0.1196], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0075, 0.0073, 0.0068, 0.0078, 0.0096, 0.0081, 0.0077], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:52:58,162 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5847, 1.3966, 0.6383, 1.2951, 1.4272, 1.4521, 1.3449, 1.4160], + device='cuda:3'), covar=tensor([0.0581, 0.0441, 0.0453, 0.0618, 0.0329, 0.0609, 0.0574, 0.0664], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 18:52:59,338 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34381.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:53:01,228 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2476, 1.5001, 1.5871, 1.7509, 1.6006, 1.7511, 1.7411, 1.7223], + device='cuda:3'), covar=tensor([0.5870, 0.9265, 0.7908, 0.7190, 0.8533, 1.2134, 0.9114, 0.8128], + device='cuda:3'), in_proj_covar=tensor([0.0319, 0.0392, 0.0317, 0.0327, 0.0343, 0.0409, 0.0372, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:53:09,190 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 18:53:11,518 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34400.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:53:23,170 INFO [finetune.py:976] (3/7) Epoch 7, batch 50, loss[loss=0.2258, simple_loss=0.2889, pruned_loss=0.08139, over 4812.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2703, pruned_loss=0.0719, over 216838.17 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:53:32,047 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.800e+02 2.094e+02 2.566e+02 4.468e+02, threshold=4.189e+02, percent-clipped=1.0 +2023-04-26 18:53:56,439 INFO [finetune.py:976] (3/7) Epoch 7, batch 100, loss[loss=0.1775, simple_loss=0.2467, pruned_loss=0.05418, over 4759.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.263, pruned_loss=0.06885, over 380804.42 frames. ], batch size: 28, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:54:19,341 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34501.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:54:29,757 INFO [finetune.py:976] (3/7) Epoch 7, batch 150, loss[loss=0.2003, simple_loss=0.2507, pruned_loss=0.07489, over 4693.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.26, pruned_loss=0.06932, over 507690.66 frames. ], batch size: 23, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:54:39,165 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.717e+02 2.069e+02 2.417e+02 7.374e+02, threshold=4.138e+02, percent-clipped=4.0 +2023-04-26 18:54:40,565 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1681, 1.6304, 2.1240, 2.5981, 2.0527, 1.5587, 1.3244, 1.9556], + device='cuda:3'), covar=tensor([0.4004, 0.4153, 0.1941, 0.3145, 0.3605, 0.3193, 0.5448, 0.3050], + device='cuda:3'), in_proj_covar=tensor([0.0275, 0.0252, 0.0216, 0.0320, 0.0213, 0.0227, 0.0237, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 18:55:02,021 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5676, 1.3516, 1.7892, 2.0367, 1.7480, 1.5736, 1.6667, 1.6394], + device='cuda:3'), covar=tensor([0.8075, 1.0002, 1.0300, 1.1068, 0.9341, 1.2093, 1.3323, 1.1922], + device='cuda:3'), in_proj_covar=tensor([0.0412, 0.0435, 0.0523, 0.0541, 0.0441, 0.0459, 0.0473, 0.0472], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:55:03,720 INFO [finetune.py:976] (3/7) Epoch 7, batch 200, loss[loss=0.181, simple_loss=0.2428, pruned_loss=0.05957, over 4761.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.262, pruned_loss=0.07153, over 608242.10 frames. ], batch size: 27, lr: 3.89e-03, grad_scale: 32.0 +2023-04-26 18:55:27,242 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34593.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:55:33,661 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34595.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:55:46,313 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4814, 2.0014, 1.4327, 1.2266, 1.1445, 1.1569, 1.4773, 1.0766], + device='cuda:3'), covar=tensor([0.1948, 0.1533, 0.1954, 0.2242, 0.2794, 0.2304, 0.1345, 0.2389], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0220, 0.0175, 0.0206, 0.0210, 0.0186, 0.0167, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:56:04,710 INFO [finetune.py:976] (3/7) Epoch 7, batch 250, loss[loss=0.174, simple_loss=0.2401, pruned_loss=0.05398, over 4773.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2662, pruned_loss=0.073, over 683505.01 frames. ], batch size: 28, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:56:09,042 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34622.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:56:17,340 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34626.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:56:19,023 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.883e+02 2.183e+02 2.769e+02 4.729e+02, threshold=4.366e+02, percent-clipped=3.0 +2023-04-26 18:56:39,245 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34643.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:56:41,143 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9715, 1.8431, 2.1624, 2.5509, 2.4885, 1.9367, 1.6658, 2.1748], + device='cuda:3'), covar=tensor([0.0954, 0.1004, 0.0575, 0.0584, 0.0518, 0.1003, 0.0830, 0.0560], + device='cuda:3'), in_proj_covar=tensor([0.0205, 0.0209, 0.0183, 0.0180, 0.0182, 0.0197, 0.0167, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:56:51,406 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:57:11,711 INFO [finetune.py:976] (3/7) Epoch 7, batch 300, loss[loss=0.2647, simple_loss=0.3222, pruned_loss=0.1036, over 4745.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2699, pruned_loss=0.07312, over 742865.39 frames. ], batch size: 59, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:57:14,790 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3737, 1.2138, 1.6997, 1.5997, 1.1997, 1.1200, 1.3595, 0.9528], + device='cuda:3'), covar=tensor([0.0739, 0.0891, 0.0566, 0.0767, 0.1129, 0.1435, 0.0744, 0.0924], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0075, 0.0073, 0.0068, 0.0078, 0.0095, 0.0081, 0.0077], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 18:57:35,525 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34683.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:57:37,960 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34687.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:57:42,155 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.5190, 4.4436, 3.0247, 5.1893, 4.4767, 4.5385, 2.0561, 4.5689], + device='cuda:3'), covar=tensor([0.1407, 0.1054, 0.3031, 0.0879, 0.2547, 0.1766, 0.5254, 0.1682], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0218, 0.0253, 0.0310, 0.0301, 0.0253, 0.0274, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 18:57:51,400 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34700.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:57:52,051 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2977, 1.6818, 2.0265, 2.7464, 2.1197, 1.5955, 1.5360, 2.0890], + device='cuda:3'), covar=tensor([0.4145, 0.4388, 0.2157, 0.3272, 0.4025, 0.3583, 0.5431, 0.3182], + device='cuda:3'), in_proj_covar=tensor([0.0277, 0.0254, 0.0218, 0.0323, 0.0214, 0.0228, 0.0239, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 18:57:52,625 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0165, 1.5677, 1.3083, 1.8188, 1.5422, 1.8595, 1.3473, 3.3474], + device='cuda:3'), covar=tensor([0.0675, 0.0778, 0.0824, 0.1162, 0.0670, 0.0531, 0.0791, 0.0166], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 18:57:53,203 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5827, 1.1569, 1.2928, 1.2903, 1.8263, 1.4157, 1.1582, 1.1779], + device='cuda:3'), covar=tensor([0.1602, 0.1428, 0.1770, 0.1531, 0.0725, 0.1580, 0.2324, 0.1994], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0332, 0.0352, 0.0305, 0.0341, 0.0332, 0.0310, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.6279e-05, 7.0770e-05, 7.6223e-05, 6.3548e-05, 7.1802e-05, 7.1618e-05, + 6.6991e-05, 7.6313e-05], device='cuda:3') +2023-04-26 18:58:03,745 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34710.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 18:58:12,744 INFO [finetune.py:976] (3/7) Epoch 7, batch 350, loss[loss=0.2248, simple_loss=0.2928, pruned_loss=0.07835, over 4813.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2712, pruned_loss=0.07367, over 790103.53 frames. ], batch size: 40, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:58:28,031 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.287e+02 1.776e+02 2.211e+02 2.674e+02 4.769e+02, threshold=4.422e+02, percent-clipped=1.0 +2023-04-26 18:58:35,796 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-26 18:58:50,331 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34748.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:58:52,205 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8591, 1.7358, 2.0701, 2.3533, 2.3217, 1.8298, 1.4783, 2.0473], + device='cuda:3'), covar=tensor([0.0941, 0.1129, 0.0582, 0.0671, 0.0619, 0.0975, 0.0951, 0.0576], + device='cuda:3'), in_proj_covar=tensor([0.0204, 0.0209, 0.0183, 0.0180, 0.0182, 0.0197, 0.0167, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:58:57,110 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7469, 1.4581, 1.3899, 1.5266, 2.0490, 1.6631, 1.3785, 1.3400], + device='cuda:3'), covar=tensor([0.1627, 0.1449, 0.1828, 0.1338, 0.0728, 0.1433, 0.2155, 0.1841], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0331, 0.0351, 0.0304, 0.0341, 0.0330, 0.0309, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.6056e-05, 7.0673e-05, 7.6148e-05, 6.3361e-05, 7.1727e-05, 7.1362e-05, + 6.6749e-05, 7.6158e-05], device='cuda:3') +2023-04-26 18:59:00,538 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3510, 2.4201, 1.8717, 2.1279, 2.4771, 1.8244, 3.3126, 1.5647], + device='cuda:3'), covar=tensor([0.4470, 0.2362, 0.5150, 0.4127, 0.2122, 0.3199, 0.1902, 0.5261], + device='cuda:3'), in_proj_covar=tensor([0.0350, 0.0357, 0.0440, 0.0367, 0.0395, 0.0387, 0.0390, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 18:59:02,247 INFO [finetune.py:976] (3/7) Epoch 7, batch 400, loss[loss=0.2136, simple_loss=0.2762, pruned_loss=0.0755, over 4909.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2725, pruned_loss=0.0736, over 828258.49 frames. ], batch size: 38, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:59:05,938 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34771.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 18:59:15,278 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4567, 0.9423, 0.3602, 1.1628, 1.0816, 1.3524, 1.2339, 1.2172], + device='cuda:3'), covar=tensor([0.0576, 0.0438, 0.0478, 0.0606, 0.0334, 0.0555, 0.0589, 0.0616], + device='cuda:3'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 18:59:26,209 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34801.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:59:36,186 INFO [finetune.py:976] (3/7) Epoch 7, batch 450, loss[loss=0.1918, simple_loss=0.2581, pruned_loss=0.06272, over 4908.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2694, pruned_loss=0.07177, over 856738.00 frames. ], batch size: 36, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:59:43,946 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1462, 2.5733, 1.1140, 1.3362, 2.1167, 1.3353, 3.3913, 1.6499], + device='cuda:3'), covar=tensor([0.0663, 0.0651, 0.0885, 0.1350, 0.0517, 0.1024, 0.0334, 0.0724], + device='cuda:3'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 18:59:45,537 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.735e+02 2.065e+02 2.383e+02 5.194e+02, threshold=4.130e+02, percent-clipped=2.0 +2023-04-26 18:59:58,678 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:00:09,433 INFO [finetune.py:976] (3/7) Epoch 7, batch 500, loss[loss=0.1861, simple_loss=0.2413, pruned_loss=0.06545, over 4800.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2664, pruned_loss=0.07124, over 880773.63 frames. ], batch size: 25, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:00:21,973 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-26 19:00:42,230 INFO [finetune.py:976] (3/7) Epoch 7, batch 550, loss[loss=0.1862, simple_loss=0.2549, pruned_loss=0.05872, over 4925.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2635, pruned_loss=0.07019, over 896324.13 frames. ], batch size: 36, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:00:51,085 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.669e+02 2.030e+02 2.418e+02 4.759e+02, threshold=4.059e+02, percent-clipped=2.0 +2023-04-26 19:01:10,916 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34949.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:01:32,183 INFO [finetune.py:976] (3/7) Epoch 7, batch 600, loss[loss=0.1786, simple_loss=0.2482, pruned_loss=0.0545, over 4758.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.266, pruned_loss=0.07139, over 909352.18 frames. ], batch size: 27, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:01:32,921 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2594, 1.2740, 1.4251, 1.6429, 1.5327, 1.2629, 0.9895, 1.4328], + device='cuda:3'), covar=tensor([0.0938, 0.1341, 0.0793, 0.0638, 0.0732, 0.0992, 0.1030, 0.0649], + device='cuda:3'), in_proj_covar=tensor([0.0204, 0.0209, 0.0183, 0.0180, 0.0182, 0.0197, 0.0167, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:01:45,331 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34978.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:01:54,263 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34982.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:02:05,672 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6900, 2.3198, 1.7138, 1.6005, 1.2149, 1.2850, 1.7877, 1.1329], + device='cuda:3'), covar=tensor([0.1961, 0.1674, 0.1670, 0.2099, 0.2844, 0.2211, 0.1286, 0.2350], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0219, 0.0174, 0.0206, 0.0210, 0.0186, 0.0166, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 19:02:23,587 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 19:02:27,947 INFO [finetune.py:976] (3/7) Epoch 7, batch 650, loss[loss=0.169, simple_loss=0.2269, pruned_loss=0.05551, over 4061.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.267, pruned_loss=0.07121, over 917502.16 frames. ], batch size: 17, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:02:42,343 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.760e+02 2.002e+02 2.405e+02 4.301e+02, threshold=4.004e+02, percent-clipped=1.0 +2023-04-26 19:03:19,053 INFO [finetune.py:976] (3/7) Epoch 7, batch 700, loss[loss=0.2272, simple_loss=0.2923, pruned_loss=0.08111, over 4841.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2681, pruned_loss=0.07065, over 925568.73 frames. ], batch size: 47, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:03:19,119 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35066.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 19:03:47,928 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35092.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:03:58,685 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8216, 2.3127, 2.0303, 2.2116, 1.6525, 1.8494, 1.9072, 1.6629], + device='cuda:3'), covar=tensor([0.2159, 0.1296, 0.0892, 0.1255, 0.3551, 0.1434, 0.1969, 0.2484], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0326, 0.0238, 0.0299, 0.0318, 0.0279, 0.0266, 0.0289], + device='cuda:3'), out_proj_covar=tensor([1.2523e-04, 1.3196e-04, 9.6434e-05, 1.2005e-04, 1.3112e-04, 1.1273e-04, + 1.0928e-04, 1.1652e-04], device='cuda:3') +2023-04-26 19:04:19,897 INFO [finetune.py:976] (3/7) Epoch 7, batch 750, loss[loss=0.1603, simple_loss=0.2222, pruned_loss=0.04917, over 4753.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2687, pruned_loss=0.07041, over 931842.27 frames. ], batch size: 27, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:04:33,722 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.243e+02 1.705e+02 2.087e+02 2.408e+02 7.580e+02, threshold=4.175e+02, percent-clipped=5.0 +2023-04-26 19:05:02,890 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 19:05:03,860 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7892, 2.2419, 1.9694, 2.0899, 1.5478, 1.8296, 1.8280, 1.6029], + device='cuda:3'), covar=tensor([0.2188, 0.1287, 0.0858, 0.1278, 0.3407, 0.1280, 0.1873, 0.2551], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0327, 0.0239, 0.0300, 0.0321, 0.0280, 0.0268, 0.0291], + device='cuda:3'), out_proj_covar=tensor([1.2586e-04, 1.3250e-04, 9.6879e-05, 1.2067e-04, 1.3200e-04, 1.1312e-04, + 1.0977e-04, 1.1721e-04], device='cuda:3') +2023-04-26 19:05:05,766 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35153.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:05:25,185 INFO [finetune.py:976] (3/7) Epoch 7, batch 800, loss[loss=0.2053, simple_loss=0.274, pruned_loss=0.06831, over 4898.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2679, pruned_loss=0.07038, over 937640.23 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:06:20,321 INFO [finetune.py:976] (3/7) Epoch 7, batch 850, loss[loss=0.1909, simple_loss=0.2547, pruned_loss=0.0636, over 4757.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.2655, pruned_loss=0.06934, over 943792.61 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:06:32,702 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.732e+02 2.031e+02 2.393e+02 4.995e+02, threshold=4.061e+02, percent-clipped=2.0 +2023-04-26 19:06:58,299 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35249.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:07:20,080 INFO [finetune.py:976] (3/7) Epoch 7, batch 900, loss[loss=0.1913, simple_loss=0.2533, pruned_loss=0.06468, over 4869.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2619, pruned_loss=0.06811, over 948023.84 frames. ], batch size: 31, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:07:38,492 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35278.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:07:39,390 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-26 19:07:40,974 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35282.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:08:02,085 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35297.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:08:08,193 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-04-26 19:08:14,479 INFO [finetune.py:976] (3/7) Epoch 7, batch 950, loss[loss=0.1905, simple_loss=0.2636, pruned_loss=0.05871, over 4758.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2601, pruned_loss=0.06774, over 947478.27 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:08:27,255 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35326.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:08:29,018 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.623e+02 1.972e+02 2.345e+02 3.927e+02, threshold=3.944e+02, percent-clipped=0.0 +2023-04-26 19:08:29,702 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35330.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:08:50,410 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35345.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:09:14,004 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2091, 2.6396, 1.0650, 1.2941, 2.0546, 1.2979, 3.5318, 1.7953], + device='cuda:3'), covar=tensor([0.0648, 0.0650, 0.0835, 0.1388, 0.0495, 0.1037, 0.0270, 0.0680], + device='cuda:3'), in_proj_covar=tensor([0.0053, 0.0068, 0.0051, 0.0048, 0.0052, 0.0053, 0.0080, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:3') +2023-04-26 19:09:20,441 INFO [finetune.py:976] (3/7) Epoch 7, batch 1000, loss[loss=0.2074, simple_loss=0.2615, pruned_loss=0.07663, over 4912.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2622, pruned_loss=0.06852, over 949690.64 frames. ], batch size: 32, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:09:20,571 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35366.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 19:09:29,986 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35374.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:10:05,116 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 19:10:06,715 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35406.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 19:10:18,214 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35414.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:10:25,351 INFO [finetune.py:976] (3/7) Epoch 7, batch 1050, loss[loss=0.1919, simple_loss=0.271, pruned_loss=0.05635, over 4811.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2675, pruned_loss=0.07067, over 948556.59 frames. ], batch size: 39, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:10:39,107 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.246e+02 1.810e+02 2.204e+02 2.753e+02 5.526e+02, threshold=4.408e+02, percent-clipped=1.0 +2023-04-26 19:10:48,432 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35435.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:11:01,949 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1786, 1.1470, 1.2579, 1.4687, 1.4390, 1.1858, 0.9110, 1.3148], + device='cuda:3'), covar=tensor([0.0789, 0.1074, 0.0606, 0.0491, 0.0608, 0.0823, 0.0919, 0.0534], + device='cuda:3'), in_proj_covar=tensor([0.0204, 0.0209, 0.0183, 0.0179, 0.0182, 0.0197, 0.0166, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:11:02,532 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35448.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:11:31,122 INFO [finetune.py:976] (3/7) Epoch 7, batch 1100, loss[loss=0.1864, simple_loss=0.25, pruned_loss=0.06144, over 4756.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2704, pruned_loss=0.07174, over 949389.36 frames. ], batch size: 28, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:12:09,916 INFO [finetune.py:976] (3/7) Epoch 7, batch 1150, loss[loss=0.1666, simple_loss=0.2311, pruned_loss=0.05109, over 4779.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2711, pruned_loss=0.07189, over 950794.60 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:12:18,335 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.291e+02 1.835e+02 2.101e+02 2.448e+02 6.183e+02, threshold=4.202e+02, percent-clipped=2.0 +2023-04-26 19:12:19,084 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35530.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:12:26,266 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5369, 1.3415, 4.1994, 3.8992, 3.7478, 3.9686, 4.0080, 3.7014], + device='cuda:3'), covar=tensor([0.6865, 0.5882, 0.1065, 0.1841, 0.1178, 0.1899, 0.1154, 0.1770], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0306, 0.0410, 0.0414, 0.0351, 0.0408, 0.0316, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:12:40,131 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-26 19:12:43,336 INFO [finetune.py:976] (3/7) Epoch 7, batch 1200, loss[loss=0.1951, simple_loss=0.2677, pruned_loss=0.06125, over 4805.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2693, pruned_loss=0.07121, over 950593.20 frames. ], batch size: 41, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:12:44,028 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6362, 1.2352, 4.5488, 4.2061, 4.0297, 4.2875, 4.2594, 3.9609], + device='cuda:3'), covar=tensor([0.6927, 0.6093, 0.0919, 0.1743, 0.1050, 0.1686, 0.1237, 0.1855], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0306, 0.0410, 0.0414, 0.0351, 0.0408, 0.0316, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:13:00,090 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35591.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:13:12,228 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6831, 1.9524, 1.9538, 2.0630, 1.9572, 2.1495, 2.1304, 2.0511], + device='cuda:3'), covar=tensor([0.5182, 0.7734, 0.7013, 0.6267, 0.7478, 1.0569, 0.7820, 0.7109], + device='cuda:3'), in_proj_covar=tensor([0.0319, 0.0390, 0.0316, 0.0326, 0.0342, 0.0407, 0.0370, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:13:17,330 INFO [finetune.py:976] (3/7) Epoch 7, batch 1250, loss[loss=0.2505, simple_loss=0.2956, pruned_loss=0.1026, over 4273.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2656, pruned_loss=0.07005, over 950418.84 frames. ], batch size: 65, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:13:19,218 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3910, 3.1332, 2.5269, 3.8895, 3.2820, 3.3673, 1.6470, 3.2874], + device='cuda:3'), covar=tensor([0.1796, 0.1506, 0.2959, 0.2120, 0.2615, 0.2044, 0.4908, 0.2557], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0218, 0.0252, 0.0308, 0.0301, 0.0253, 0.0273, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:13:26,237 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.763e+02 2.030e+02 2.428e+02 4.549e+02, threshold=4.060e+02, percent-clipped=1.0 +2023-04-26 19:13:43,539 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 19:13:51,308 INFO [finetune.py:976] (3/7) Epoch 7, batch 1300, loss[loss=0.2375, simple_loss=0.2942, pruned_loss=0.09034, over 4211.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2626, pruned_loss=0.06885, over 951897.99 frames. ], batch size: 65, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:13:58,521 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6702, 1.1153, 1.3327, 1.4600, 1.9009, 1.4787, 1.2260, 1.2874], + device='cuda:3'), covar=tensor([0.1916, 0.1666, 0.1998, 0.1382, 0.0870, 0.1695, 0.2558, 0.2149], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0330, 0.0352, 0.0303, 0.0339, 0.0329, 0.0308, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.6128e-05, 7.0355e-05, 7.6299e-05, 6.2978e-05, 7.1257e-05, 7.0873e-05, + 6.6526e-05, 7.6153e-05], device='cuda:3') +2023-04-26 19:14:13,534 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2697, 1.6809, 1.5141, 2.1944, 1.7684, 2.1894, 1.4802, 4.3845], + device='cuda:3'), covar=tensor([0.0597, 0.0713, 0.0795, 0.1073, 0.0648, 0.0518, 0.0761, 0.0141], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-26 19:14:14,101 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35701.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:14:35,193 INFO [finetune.py:976] (3/7) Epoch 7, batch 1350, loss[loss=0.167, simple_loss=0.2318, pruned_loss=0.05113, over 4771.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.263, pruned_loss=0.06931, over 954212.99 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:14:45,760 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9392, 1.6535, 1.9705, 2.3587, 2.3006, 1.8736, 1.5584, 1.9404], + device='cuda:3'), covar=tensor([0.0943, 0.1222, 0.0634, 0.0615, 0.0689, 0.0994, 0.0957, 0.0720], + device='cuda:3'), in_proj_covar=tensor([0.0204, 0.0209, 0.0184, 0.0180, 0.0183, 0.0198, 0.0166, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:14:53,892 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-26 19:14:54,290 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.639e+02 1.977e+02 2.381e+02 4.004e+02, threshold=3.953e+02, percent-clipped=0.0 +2023-04-26 19:14:54,381 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.5148, 4.2731, 3.1546, 5.1383, 4.4463, 4.4464, 2.0111, 4.4164], + device='cuda:3'), covar=tensor([0.1371, 0.1124, 0.3508, 0.1011, 0.3071, 0.1605, 0.5514, 0.2071], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0220, 0.0254, 0.0311, 0.0304, 0.0255, 0.0276, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:14:54,989 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35730.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:15:17,743 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35748.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:15:40,838 INFO [finetune.py:976] (3/7) Epoch 7, batch 1400, loss[loss=0.2425, simple_loss=0.2905, pruned_loss=0.09728, over 4746.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2686, pruned_loss=0.07148, over 955233.66 frames. ], batch size: 27, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:15:44,446 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35771.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:16:06,189 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35796.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:16:26,233 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35811.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:16:30,135 INFO [finetune.py:976] (3/7) Epoch 7, batch 1450, loss[loss=0.1972, simple_loss=0.2655, pruned_loss=0.06446, over 4744.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2701, pruned_loss=0.07111, over 956641.52 frames. ], batch size: 54, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:16:47,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9515, 2.0637, 1.6502, 1.6784, 2.0339, 1.5264, 2.6588, 1.3675], + device='cuda:3'), covar=tensor([0.4078, 0.1942, 0.5758, 0.3251, 0.1944, 0.2809, 0.1413, 0.5277], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0350, 0.0430, 0.0361, 0.0386, 0.0380, 0.0383, 0.0415], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:16:50,309 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.278e+02 1.770e+02 2.190e+02 2.562e+02 4.335e+02, threshold=4.381e+02, percent-clipped=3.0 +2023-04-26 19:16:58,474 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35832.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:17:41,153 INFO [finetune.py:976] (3/7) Epoch 7, batch 1500, loss[loss=0.1778, simple_loss=0.2579, pruned_loss=0.0488, over 4725.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2718, pruned_loss=0.07225, over 956436.48 frames. ], batch size: 54, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:17:46,036 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35872.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:17:55,872 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35886.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:18:30,742 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5611, 1.0544, 1.2907, 1.1946, 1.6846, 1.3505, 1.1155, 1.2974], + device='cuda:3'), covar=tensor([0.1521, 0.1432, 0.2149, 0.1553, 0.0819, 0.1500, 0.1985, 0.2061], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0333, 0.0356, 0.0306, 0.0341, 0.0333, 0.0311, 0.0357], + device='cuda:3'), out_proj_covar=tensor([6.6652e-05, 7.0897e-05, 7.7162e-05, 6.3510e-05, 7.1894e-05, 7.1791e-05, + 6.7121e-05, 7.6947e-05], device='cuda:3') +2023-04-26 19:18:31,248 INFO [finetune.py:976] (3/7) Epoch 7, batch 1550, loss[loss=0.2035, simple_loss=0.262, pruned_loss=0.07247, over 4857.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2707, pruned_loss=0.07188, over 955848.29 frames. ], batch size: 31, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:18:51,454 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.799e+02 2.113e+02 2.516e+02 6.961e+02, threshold=4.225e+02, percent-clipped=2.0 +2023-04-26 19:19:38,739 INFO [finetune.py:976] (3/7) Epoch 7, batch 1600, loss[loss=0.1916, simple_loss=0.25, pruned_loss=0.06654, over 4708.00 frames. ], tot_loss[loss=0.205, simple_loss=0.2685, pruned_loss=0.0708, over 957812.53 frames. ], batch size: 23, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:20:00,168 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7683, 2.4221, 2.0936, 2.2451, 1.7603, 1.9208, 1.9681, 1.5433], + device='cuda:3'), covar=tensor([0.2389, 0.1392, 0.0963, 0.1370, 0.3097, 0.1392, 0.2078, 0.3014], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0324, 0.0237, 0.0300, 0.0318, 0.0279, 0.0266, 0.0289], + device='cuda:3'), out_proj_covar=tensor([1.2488e-04, 1.3125e-04, 9.5924e-05, 1.2035e-04, 1.3094e-04, 1.1257e-04, + 1.0898e-04, 1.1637e-04], device='cuda:3') +2023-04-26 19:20:08,960 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8850, 1.4534, 1.4545, 1.5991, 2.0838, 1.6717, 1.4169, 1.4416], + device='cuda:3'), covar=tensor([0.1549, 0.1769, 0.2043, 0.1457, 0.0921, 0.2155, 0.2332, 0.2303], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0333, 0.0356, 0.0305, 0.0341, 0.0332, 0.0311, 0.0357], + device='cuda:3'), out_proj_covar=tensor([6.6552e-05, 7.0904e-05, 7.7059e-05, 6.3429e-05, 7.1837e-05, 7.1629e-05, + 6.7152e-05, 7.6954e-05], device='cuda:3') +2023-04-26 19:20:30,759 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36001.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 19:20:45,229 INFO [finetune.py:976] (3/7) Epoch 7, batch 1650, loss[loss=0.2095, simple_loss=0.2787, pruned_loss=0.07014, over 4764.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2651, pruned_loss=0.06932, over 958059.01 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:20:59,791 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.778e+02 2.122e+02 2.492e+02 4.114e+02, threshold=4.243e+02, percent-clipped=0.0 +2023-04-26 19:21:00,972 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36030.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:21:26,272 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36049.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:21:29,435 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3755, 0.9000, 1.2172, 1.0578, 1.5057, 1.2667, 1.0650, 1.1649], + device='cuda:3'), covar=tensor([0.1708, 0.1686, 0.1815, 0.1712, 0.1086, 0.1285, 0.2321, 0.1941], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0331, 0.0354, 0.0303, 0.0340, 0.0330, 0.0309, 0.0356], + device='cuda:3'), out_proj_covar=tensor([6.6109e-05, 7.0521e-05, 7.6698e-05, 6.2955e-05, 7.1561e-05, 7.1291e-05, + 6.6817e-05, 7.6571e-05], device='cuda:3') +2023-04-26 19:21:48,194 INFO [finetune.py:976] (3/7) Epoch 7, batch 1700, loss[loss=0.1695, simple_loss=0.2458, pruned_loss=0.04665, over 4902.00 frames. ], tot_loss[loss=0.2017, simple_loss=0.2638, pruned_loss=0.06983, over 956186.59 frames. ], batch size: 37, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:21:56,083 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36078.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:22:05,517 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 19:22:21,093 INFO [finetune.py:976] (3/7) Epoch 7, batch 1750, loss[loss=0.2303, simple_loss=0.3054, pruned_loss=0.0776, over 4815.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2654, pruned_loss=0.07036, over 955102.94 frames. ], batch size: 45, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:22:28,798 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36127.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:22:29,945 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.331e+02 1.866e+02 2.133e+02 2.694e+02 4.644e+02, threshold=4.265e+02, percent-clipped=2.0 +2023-04-26 19:22:52,965 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9128, 1.3980, 1.4567, 1.6414, 2.1277, 1.7159, 1.4281, 1.3954], + device='cuda:3'), covar=tensor([0.1796, 0.1824, 0.2639, 0.1554, 0.0883, 0.1619, 0.2216, 0.2226], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0332, 0.0355, 0.0304, 0.0340, 0.0330, 0.0310, 0.0356], + device='cuda:3'), out_proj_covar=tensor([6.6026e-05, 7.0702e-05, 7.6906e-05, 6.3159e-05, 7.1647e-05, 7.1338e-05, + 6.7065e-05, 7.6681e-05], device='cuda:3') +2023-04-26 19:22:54,679 INFO [finetune.py:976] (3/7) Epoch 7, batch 1800, loss[loss=0.1914, simple_loss=0.2587, pruned_loss=0.06209, over 4871.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2686, pruned_loss=0.0711, over 956020.11 frames. ], batch size: 34, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:22:55,364 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36167.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:23:05,433 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36181.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:23:08,481 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36186.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:23:44,253 INFO [finetune.py:976] (3/7) Epoch 7, batch 1850, loss[loss=0.1999, simple_loss=0.2707, pruned_loss=0.06456, over 4909.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2704, pruned_loss=0.07158, over 957178.98 frames. ], batch size: 36, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:24:03,617 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.720e+02 2.075e+02 2.567e+02 3.756e+02, threshold=4.150e+02, percent-clipped=0.0 +2023-04-26 19:24:07,671 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36234.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:24:19,053 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36242.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:24:45,929 INFO [finetune.py:976] (3/7) Epoch 7, batch 1900, loss[loss=0.1971, simple_loss=0.2503, pruned_loss=0.07195, over 4891.00 frames. ], tot_loss[loss=0.2079, simple_loss=0.2718, pruned_loss=0.07198, over 955508.92 frames. ], batch size: 32, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:25:19,583 INFO [finetune.py:976] (3/7) Epoch 7, batch 1950, loss[loss=0.259, simple_loss=0.3027, pruned_loss=0.1077, over 4753.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2703, pruned_loss=0.07149, over 954845.25 frames. ], batch size: 59, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:25:27,369 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.259e+02 1.677e+02 1.937e+02 2.378e+02 4.833e+02, threshold=3.873e+02, percent-clipped=2.0 +2023-04-26 19:25:32,074 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6687, 3.9998, 0.7502, 2.2362, 2.3383, 2.7668, 2.3872, 0.9691], + device='cuda:3'), covar=tensor([0.1418, 0.0979, 0.2275, 0.1235, 0.1058, 0.1074, 0.1331, 0.2225], + device='cuda:3'), in_proj_covar=tensor([0.0120, 0.0258, 0.0145, 0.0127, 0.0137, 0.0158, 0.0123, 0.0125], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 19:26:09,641 INFO [finetune.py:976] (3/7) Epoch 7, batch 2000, loss[loss=0.1922, simple_loss=0.2532, pruned_loss=0.06564, over 4784.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2661, pruned_loss=0.06987, over 954924.73 frames. ], batch size: 29, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:26:51,958 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8966, 1.2493, 3.2885, 3.0467, 2.9619, 3.1828, 3.1975, 2.9003], + device='cuda:3'), covar=tensor([0.7323, 0.5130, 0.1349, 0.1967, 0.1289, 0.2056, 0.1628, 0.1566], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0302, 0.0407, 0.0411, 0.0349, 0.0406, 0.0315, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:27:00,244 INFO [finetune.py:976] (3/7) Epoch 7, batch 2050, loss[loss=0.2338, simple_loss=0.2766, pruned_loss=0.09557, over 4850.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2628, pruned_loss=0.06887, over 954272.89 frames. ], batch size: 47, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:27:05,286 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8012, 1.6636, 1.9921, 2.2177, 2.1063, 1.7597, 1.3633, 1.7819], + device='cuda:3'), covar=tensor([0.1030, 0.1186, 0.0669, 0.0672, 0.0746, 0.1076, 0.1027, 0.0729], + device='cuda:3'), in_proj_covar=tensor([0.0200, 0.0206, 0.0181, 0.0177, 0.0180, 0.0194, 0.0162, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:27:07,132 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36427.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:27:08,256 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.550e+02 1.908e+02 2.415e+02 5.206e+02, threshold=3.816e+02, percent-clipped=3.0 +2023-04-26 19:27:43,883 INFO [finetune.py:976] (3/7) Epoch 7, batch 2100, loss[loss=0.2126, simple_loss=0.2906, pruned_loss=0.06733, over 4818.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2631, pruned_loss=0.06883, over 954405.72 frames. ], batch size: 39, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:27:45,076 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:27:49,907 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36475.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:27:56,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2971, 1.5136, 1.6186, 1.8023, 1.6350, 1.7640, 1.7951, 1.7440], + device='cuda:3'), covar=tensor([0.5811, 0.8028, 0.6863, 0.6411, 0.7663, 1.0922, 0.8035, 0.7392], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0394, 0.0319, 0.0330, 0.0346, 0.0409, 0.0374, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:28:17,169 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36515.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:28:17,708 INFO [finetune.py:976] (3/7) Epoch 7, batch 2150, loss[loss=0.2681, simple_loss=0.3249, pruned_loss=0.1057, over 4912.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2677, pruned_loss=0.07084, over 956714.20 frames. ], batch size: 36, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:28:25,936 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.260e+02 1.822e+02 2.196e+02 2.668e+02 7.231e+02, threshold=4.392e+02, percent-clipped=1.0 +2023-04-26 19:28:30,879 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36537.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:28:51,287 INFO [finetune.py:976] (3/7) Epoch 7, batch 2200, loss[loss=0.1874, simple_loss=0.2565, pruned_loss=0.05918, over 4893.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.27, pruned_loss=0.0715, over 955933.87 frames. ], batch size: 35, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:29:28,457 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36605.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:29:30,345 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36608.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:29:38,210 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36612.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:29:41,094 INFO [finetune.py:976] (3/7) Epoch 7, batch 2250, loss[loss=0.2274, simple_loss=0.2931, pruned_loss=0.08086, over 4911.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2694, pruned_loss=0.07124, over 954968.17 frames. ], batch size: 36, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:30:00,405 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.830e+02 2.114e+02 2.598e+02 4.357e+02, threshold=4.229e+02, percent-clipped=0.0 +2023-04-26 19:30:40,442 INFO [finetune.py:976] (3/7) Epoch 7, batch 2300, loss[loss=0.1952, simple_loss=0.268, pruned_loss=0.06118, over 4810.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2695, pruned_loss=0.07121, over 954143.23 frames. ], batch size: 39, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:30:40,556 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36666.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:30:42,379 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36669.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:30:45,360 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36673.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:31:02,260 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36700.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:31:23,678 INFO [finetune.py:976] (3/7) Epoch 7, batch 2350, loss[loss=0.1953, simple_loss=0.2531, pruned_loss=0.06881, over 4808.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2666, pruned_loss=0.07007, over 954697.38 frames. ], batch size: 25, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:31:27,172 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-26 19:31:38,221 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.607e+01 1.814e+02 2.081e+02 2.499e+02 5.543e+02, threshold=4.163e+02, percent-clipped=2.0 +2023-04-26 19:32:21,174 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36761.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 19:32:30,354 INFO [finetune.py:976] (3/7) Epoch 7, batch 2400, loss[loss=0.1769, simple_loss=0.2483, pruned_loss=0.05282, over 4766.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.264, pruned_loss=0.06932, over 955496.63 frames. ], batch size: 28, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:32:53,302 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2384, 1.4770, 1.6295, 1.8547, 1.7150, 1.8089, 1.7273, 1.7460], + device='cuda:3'), covar=tensor([0.5771, 0.8070, 0.7231, 0.6166, 0.7867, 1.0865, 0.8551, 0.7404], + device='cuda:3'), in_proj_covar=tensor([0.0321, 0.0393, 0.0318, 0.0328, 0.0346, 0.0409, 0.0372, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:32:56,363 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36805.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:33:03,828 INFO [finetune.py:976] (3/7) Epoch 7, batch 2450, loss[loss=0.1405, simple_loss=0.2115, pruned_loss=0.03478, over 4781.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2613, pruned_loss=0.06873, over 952346.77 frames. ], batch size: 27, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:33:12,663 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.628e+02 1.856e+02 2.205e+02 3.828e+02, threshold=3.712e+02, percent-clipped=0.0 +2023-04-26 19:33:18,550 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36837.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:33:37,051 INFO [finetune.py:976] (3/7) Epoch 7, batch 2500, loss[loss=0.2203, simple_loss=0.2786, pruned_loss=0.08102, over 4813.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2625, pruned_loss=0.06913, over 954750.00 frames. ], batch size: 38, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:33:37,690 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36866.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:33:37,888 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-26 19:33:51,062 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36885.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:07,391 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36911.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:10,362 INFO [finetune.py:976] (3/7) Epoch 7, batch 2550, loss[loss=0.2013, simple_loss=0.2685, pruned_loss=0.06706, over 4893.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2659, pruned_loss=0.07005, over 951685.44 frames. ], batch size: 32, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:34:13,259 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2681, 1.5526, 1.5521, 1.7847, 1.5258, 1.6934, 1.7136, 1.6628], + device='cuda:3'), covar=tensor([0.6179, 0.9194, 0.7575, 0.6826, 0.8389, 1.1461, 0.8830, 0.8032], + device='cuda:3'), in_proj_covar=tensor([0.0321, 0.0393, 0.0319, 0.0328, 0.0346, 0.0409, 0.0373, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:34:20,132 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.245e+02 1.780e+02 2.108e+02 2.520e+02 3.900e+02, threshold=4.217e+02, percent-clipped=2.0 +2023-04-26 19:34:21,734 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-26 19:34:45,072 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36958.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:52,520 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36961.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:54,323 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36964.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:55,486 INFO [finetune.py:976] (3/7) Epoch 7, batch 2600, loss[loss=0.1656, simple_loss=0.2242, pruned_loss=0.05349, over 3946.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.268, pruned_loss=0.07041, over 949896.72 frames. ], batch size: 17, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:35:00,961 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36968.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:35:04,493 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36972.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:35:48,164 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37004.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:36:01,127 INFO [finetune.py:976] (3/7) Epoch 7, batch 2650, loss[loss=0.2519, simple_loss=0.3108, pruned_loss=0.09654, over 4808.00 frames. ], tot_loss[loss=0.205, simple_loss=0.269, pruned_loss=0.07048, over 949292.32 frames. ], batch size: 40, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:36:03,043 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37019.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:36:09,956 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.731e+02 2.035e+02 2.586e+02 4.493e+02, threshold=4.069e+02, percent-clipped=1.0 +2023-04-26 19:36:28,805 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37056.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:36:31,878 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4146, 1.6099, 1.5229, 1.8601, 1.6575, 2.0431, 1.4733, 3.1785], + device='cuda:3'), covar=tensor([0.0589, 0.0652, 0.0698, 0.1034, 0.0527, 0.0569, 0.0669, 0.0189], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 19:36:34,289 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37065.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:36:34,794 INFO [finetune.py:976] (3/7) Epoch 7, batch 2700, loss[loss=0.2147, simple_loss=0.2664, pruned_loss=0.08152, over 4785.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2683, pruned_loss=0.0703, over 949981.05 frames. ], batch size: 29, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:36:46,115 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8447, 1.8728, 2.2365, 2.3818, 2.3501, 1.9244, 1.5203, 1.8882], + device='cuda:3'), covar=tensor([0.0972, 0.1056, 0.0566, 0.0634, 0.0649, 0.0971, 0.0956, 0.0718], + device='cuda:3'), in_proj_covar=tensor([0.0201, 0.0207, 0.0182, 0.0179, 0.0181, 0.0194, 0.0164, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:37:19,212 INFO [finetune.py:976] (3/7) Epoch 7, batch 2750, loss[loss=0.2136, simple_loss=0.2639, pruned_loss=0.08162, over 4868.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2664, pruned_loss=0.06973, over 952574.39 frames. ], batch size: 34, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:37:32,276 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.627e+02 1.918e+02 2.292e+02 3.296e+02, threshold=3.835e+02, percent-clipped=0.0 +2023-04-26 19:38:17,426 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37161.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:38:20,451 INFO [finetune.py:976] (3/7) Epoch 7, batch 2800, loss[loss=0.1726, simple_loss=0.2392, pruned_loss=0.05304, over 4910.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2627, pruned_loss=0.06849, over 954250.27 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:38:34,704 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1288, 2.0971, 1.8321, 1.9017, 2.1996, 1.8110, 2.7116, 1.4477], + device='cuda:3'), covar=tensor([0.3862, 0.1787, 0.4833, 0.2858, 0.1869, 0.2516, 0.1506, 0.4718], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0354, 0.0436, 0.0366, 0.0390, 0.0386, 0.0387, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:39:00,022 INFO [finetune.py:976] (3/7) Epoch 7, batch 2850, loss[loss=0.1795, simple_loss=0.2415, pruned_loss=0.05874, over 4806.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2602, pruned_loss=0.06775, over 952433.00 frames. ], batch size: 25, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:39:08,542 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.213e+02 1.711e+02 1.920e+02 2.384e+02 5.364e+02, threshold=3.840e+02, percent-clipped=2.0 +2023-04-26 19:39:28,504 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2792, 1.4703, 1.5470, 1.7012, 1.5838, 1.7083, 1.6893, 1.6518], + device='cuda:3'), covar=tensor([0.5550, 0.7128, 0.6321, 0.5643, 0.7265, 1.0557, 0.7069, 0.6864], + device='cuda:3'), in_proj_covar=tensor([0.0320, 0.0389, 0.0316, 0.0327, 0.0342, 0.0407, 0.0370, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:39:30,799 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37261.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:32,615 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37264.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:33,742 INFO [finetune.py:976] (3/7) Epoch 7, batch 2900, loss[loss=0.2858, simple_loss=0.3376, pruned_loss=0.117, over 4804.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.2651, pruned_loss=0.07013, over 953392.16 frames. ], batch size: 51, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:39:34,410 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37267.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:35,024 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:35,037 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:56,264 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 19:40:02,585 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37309.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:10,643 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37312.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:11,891 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37314.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:13,065 INFO [finetune.py:976] (3/7) Epoch 7, batch 2950, loss[loss=0.2047, simple_loss=0.281, pruned_loss=0.06417, over 4863.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2679, pruned_loss=0.07116, over 953702.79 frames. ], batch size: 44, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:40:13,125 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37316.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:32,654 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37329.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:33,129 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.889e+02 2.237e+02 2.673e+02 8.962e+02, threshold=4.474e+02, percent-clipped=7.0 +2023-04-26 19:40:56,679 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5049, 1.6738, 1.6865, 1.8213, 1.6337, 1.7944, 1.8435, 1.7819], + device='cuda:3'), covar=tensor([0.5552, 0.8443, 0.7676, 0.6462, 0.7799, 1.1072, 0.8484, 0.7871], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0393, 0.0319, 0.0329, 0.0345, 0.0410, 0.0372, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 19:41:05,672 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37356.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:41:08,606 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37360.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:41:17,440 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37364.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:41:18,547 INFO [finetune.py:976] (3/7) Epoch 7, batch 3000, loss[loss=0.2111, simple_loss=0.2909, pruned_loss=0.06567, over 4816.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.2685, pruned_loss=0.071, over 952683.86 frames. ], batch size: 39, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:41:18,547 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 19:41:40,584 INFO [finetune.py:1010] (3/7) Epoch 7, validation: loss=0.1559, simple_loss=0.2289, pruned_loss=0.04148, over 2265189.00 frames. +2023-04-26 19:41:40,599 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 19:42:15,397 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37404.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:42:23,695 INFO [finetune.py:976] (3/7) Epoch 7, batch 3050, loss[loss=0.2252, simple_loss=0.2846, pruned_loss=0.08289, over 4885.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2689, pruned_loss=0.07009, over 953569.80 frames. ], batch size: 35, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:42:30,260 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37425.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:42:31,345 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9189, 2.5251, 1.5624, 1.7271, 2.3910, 1.8284, 1.7547, 1.9610], + device='cuda:3'), covar=tensor([0.0501, 0.0334, 0.0316, 0.0561, 0.0257, 0.0556, 0.0577, 0.0555], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0048, 0.0049], + device='cuda:3') +2023-04-26 19:42:34,151 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.257e+01 1.766e+02 2.163e+02 2.758e+02 4.245e+02, threshold=4.325e+02, percent-clipped=0.0 +2023-04-26 19:42:54,127 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 19:42:54,536 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37461.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:42:57,501 INFO [finetune.py:976] (3/7) Epoch 7, batch 3100, loss[loss=0.171, simple_loss=0.2362, pruned_loss=0.05286, over 4767.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2655, pruned_loss=0.0685, over 953002.08 frames. ], batch size: 26, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:43:26,085 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37509.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:43:31,232 INFO [finetune.py:976] (3/7) Epoch 7, batch 3150, loss[loss=0.2128, simple_loss=0.2701, pruned_loss=0.07774, over 4874.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2637, pruned_loss=0.06873, over 950935.03 frames. ], batch size: 31, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:43:52,226 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.638e+02 2.016e+02 2.443e+02 5.326e+02, threshold=4.032e+02, percent-clipped=1.0 +2023-04-26 19:44:05,174 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5521, 1.6377, 0.6748, 1.2707, 1.6740, 1.4041, 1.3403, 1.3660], + device='cuda:3'), covar=tensor([0.0540, 0.0397, 0.0444, 0.0600, 0.0298, 0.0562, 0.0542, 0.0626], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 19:44:25,681 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6138, 3.4876, 0.8618, 2.0338, 2.0327, 2.4761, 2.0346, 1.0320], + device='cuda:3'), covar=tensor([0.1300, 0.1000, 0.2112, 0.1262, 0.1072, 0.1085, 0.1480, 0.2112], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0254, 0.0143, 0.0125, 0.0135, 0.0155, 0.0120, 0.0123], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 19:44:36,863 INFO [finetune.py:976] (3/7) Epoch 7, batch 3200, loss[loss=0.2055, simple_loss=0.264, pruned_loss=0.07345, over 4716.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2607, pruned_loss=0.06772, over 951808.36 frames. ], batch size: 59, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:44:37,567 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37567.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:45:33,739 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 19:45:42,725 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37614.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:45:43,285 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37615.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:45:43,827 INFO [finetune.py:976] (3/7) Epoch 7, batch 3250, loss[loss=0.2329, simple_loss=0.2846, pruned_loss=0.09061, over 4905.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2616, pruned_loss=0.06818, over 950972.33 frames. ], batch size: 36, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:45:55,385 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37624.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:46:03,764 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.639e+01 1.714e+02 1.975e+02 2.453e+02 9.656e+02, threshold=3.950e+02, percent-clipped=3.0 +2023-04-26 19:46:41,407 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37660.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:46:42,585 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37662.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:46:45,458 INFO [finetune.py:976] (3/7) Epoch 7, batch 3300, loss[loss=0.2237, simple_loss=0.3015, pruned_loss=0.07299, over 4804.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2655, pruned_loss=0.06972, over 953016.79 frames. ], batch size: 41, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:47:48,146 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37708.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:47:57,367 INFO [finetune.py:976] (3/7) Epoch 7, batch 3350, loss[loss=0.2702, simple_loss=0.3265, pruned_loss=0.1069, over 4753.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2668, pruned_loss=0.06985, over 954908.70 frames. ], batch size: 54, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:48:00,381 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37720.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:48:10,678 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 19:48:12,885 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.789e+02 2.091e+02 2.525e+02 4.436e+02, threshold=4.182e+02, percent-clipped=1.0 +2023-04-26 19:48:36,823 INFO [finetune.py:976] (3/7) Epoch 7, batch 3400, loss[loss=0.2058, simple_loss=0.2773, pruned_loss=0.06714, over 4865.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2677, pruned_loss=0.07007, over 956056.73 frames. ], batch size: 31, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:49:36,094 INFO [finetune.py:976] (3/7) Epoch 7, batch 3450, loss[loss=0.1792, simple_loss=0.2501, pruned_loss=0.05413, over 4813.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.2665, pruned_loss=0.06883, over 956279.82 frames. ], batch size: 38, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:49:55,706 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.199e+01 1.617e+02 2.034e+02 2.503e+02 3.981e+02, threshold=4.069e+02, percent-clipped=0.0 +2023-04-26 19:50:33,685 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 19:50:37,108 INFO [finetune.py:976] (3/7) Epoch 7, batch 3500, loss[loss=0.1583, simple_loss=0.2203, pruned_loss=0.04813, over 4765.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.264, pruned_loss=0.06859, over 954435.39 frames. ], batch size: 27, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:51:06,011 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-26 19:51:17,010 INFO [finetune.py:976] (3/7) Epoch 7, batch 3550, loss[loss=0.1752, simple_loss=0.2372, pruned_loss=0.05658, over 4812.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2612, pruned_loss=0.06743, over 954843.45 frames. ], batch size: 41, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:51:27,506 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37924.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:51:34,160 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 19:51:34,697 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8521, 2.4373, 2.0037, 2.2010, 1.6837, 2.0097, 2.0625, 1.5435], + device='cuda:3'), covar=tensor([0.2118, 0.1497, 0.1120, 0.1386, 0.3444, 0.1308, 0.1900, 0.2589], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0321, 0.0234, 0.0295, 0.0315, 0.0273, 0.0262, 0.0286], + device='cuda:3'), out_proj_covar=tensor([1.2303e-04, 1.3002e-04, 9.4601e-05, 1.1812e-04, 1.2945e-04, 1.1045e-04, + 1.0762e-04, 1.1512e-04], device='cuda:3') +2023-04-26 19:51:36,392 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.609e+02 1.946e+02 2.387e+02 4.801e+02, threshold=3.892e+02, percent-clipped=2.0 +2023-04-26 19:51:48,491 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.04 vs. limit=5.0 +2023-04-26 19:52:01,356 INFO [finetune.py:976] (3/7) Epoch 7, batch 3600, loss[loss=0.1804, simple_loss=0.2413, pruned_loss=0.05974, over 4828.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2594, pruned_loss=0.06667, over 954156.33 frames. ], batch size: 30, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:52:05,027 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37972.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:52:09,406 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37979.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:52:35,121 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 19:52:35,871 INFO [finetune.py:976] (3/7) Epoch 7, batch 3650, loss[loss=0.2607, simple_loss=0.3212, pruned_loss=0.1, over 4824.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.2611, pruned_loss=0.06795, over 951936.99 frames. ], batch size: 51, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:52:38,361 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38020.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:52:44,355 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.213e+02 1.731e+02 1.989e+02 2.580e+02 1.075e+03, threshold=3.977e+02, percent-clipped=3.0 +2023-04-26 19:52:51,021 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38040.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:53:08,980 INFO [finetune.py:976] (3/7) Epoch 7, batch 3700, loss[loss=0.1965, simple_loss=0.2628, pruned_loss=0.06508, over 4928.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2642, pruned_loss=0.0688, over 952047.82 frames. ], batch size: 38, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:53:10,246 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38068.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:53:41,669 INFO [finetune.py:976] (3/7) Epoch 7, batch 3750, loss[loss=0.1844, simple_loss=0.2489, pruned_loss=0.06001, over 4809.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2666, pruned_loss=0.07007, over 952245.79 frames. ], batch size: 45, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:53:50,670 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.250e+02 1.891e+02 2.141e+02 2.716e+02 4.079e+02, threshold=4.281e+02, percent-clipped=1.0 +2023-04-26 19:53:52,627 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7034, 1.5464, 1.7248, 2.0942, 2.1132, 1.6856, 1.4911, 1.8108], + device='cuda:3'), covar=tensor([0.0729, 0.0963, 0.0566, 0.0428, 0.0457, 0.0760, 0.0727, 0.0503], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0203, 0.0180, 0.0176, 0.0178, 0.0192, 0.0161, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 19:53:53,485 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.50 vs. limit=5.0 +2023-04-26 19:54:31,494 INFO [finetune.py:976] (3/7) Epoch 7, batch 3800, loss[loss=0.1187, simple_loss=0.1902, pruned_loss=0.02362, over 4764.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2673, pruned_loss=0.07059, over 950535.97 frames. ], batch size: 23, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:54:55,813 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 19:55:36,932 INFO [finetune.py:976] (3/7) Epoch 7, batch 3850, loss[loss=0.227, simple_loss=0.285, pruned_loss=0.08454, over 4870.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2652, pruned_loss=0.06886, over 951751.93 frames. ], batch size: 34, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:55:56,087 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.647e+02 1.925e+02 2.333e+02 3.999e+02, threshold=3.850e+02, percent-clipped=0.0 +2023-04-26 19:56:30,064 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5641, 1.9853, 2.0506, 2.0090, 2.0627, 2.2513, 1.7768, 3.9280], + device='cuda:3'), covar=tensor([0.0587, 0.0682, 0.0670, 0.1114, 0.0547, 0.0435, 0.0678, 0.0155], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0041, 0.0040, 0.0039, 0.0060], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 19:56:41,864 INFO [finetune.py:976] (3/7) Epoch 7, batch 3900, loss[loss=0.1975, simple_loss=0.2683, pruned_loss=0.06338, over 4910.00 frames. ], tot_loss[loss=0.1989, simple_loss=0.2622, pruned_loss=0.06784, over 951526.87 frames. ], batch size: 35, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:57:06,343 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1409, 2.5664, 1.1274, 1.4190, 2.1151, 1.1674, 3.5526, 1.8430], + device='cuda:3'), covar=tensor([0.0679, 0.0746, 0.0834, 0.1226, 0.0469, 0.1017, 0.0215, 0.0610], + device='cuda:3'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 19:57:47,851 INFO [finetune.py:976] (3/7) Epoch 7, batch 3950, loss[loss=0.1761, simple_loss=0.2473, pruned_loss=0.05244, over 4766.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2586, pruned_loss=0.06623, over 951654.91 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:58:09,819 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.650e+02 1.940e+02 2.337e+02 4.288e+02, threshold=3.879e+02, percent-clipped=1.0 +2023-04-26 19:58:18,663 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38335.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:58:43,688 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38360.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:58:53,666 INFO [finetune.py:976] (3/7) Epoch 7, batch 4000, loss[loss=0.2062, simple_loss=0.2684, pruned_loss=0.07198, over 4809.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.259, pruned_loss=0.06687, over 950555.12 frames. ], batch size: 40, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:00:00,677 INFO [finetune.py:976] (3/7) Epoch 7, batch 4050, loss[loss=0.1919, simple_loss=0.2688, pruned_loss=0.05743, over 4763.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2633, pruned_loss=0.06895, over 948827.71 frames. ], batch size: 28, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:00:10,075 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38421.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:00:21,445 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.926e+01 1.887e+02 2.250e+02 2.854e+02 6.950e+02, threshold=4.500e+02, percent-clipped=7.0 +2023-04-26 20:00:25,153 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0602, 1.4724, 1.5559, 1.5996, 2.2114, 1.8100, 1.4756, 1.5048], + device='cuda:3'), covar=tensor([0.1539, 0.1866, 0.2280, 0.1705, 0.0852, 0.1700, 0.2891, 0.2432], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0329, 0.0358, 0.0305, 0.0342, 0.0328, 0.0310, 0.0356], + device='cuda:3'), out_proj_covar=tensor([6.5882e-05, 7.0161e-05, 7.7529e-05, 6.3420e-05, 7.2034e-05, 7.0816e-05, + 6.6940e-05, 7.6471e-05], device='cuda:3') +2023-04-26 20:01:01,406 INFO [finetune.py:976] (3/7) Epoch 7, batch 4100, loss[loss=0.2134, simple_loss=0.2747, pruned_loss=0.07608, over 4860.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2667, pruned_loss=0.07003, over 950191.67 frames. ], batch size: 31, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:02:13,352 INFO [finetune.py:976] (3/7) Epoch 7, batch 4150, loss[loss=0.2594, simple_loss=0.3182, pruned_loss=0.1002, over 4126.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2685, pruned_loss=0.07041, over 953255.03 frames. ], batch size: 66, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:02:23,484 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-26 20:02:28,364 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.719e+02 1.998e+02 2.310e+02 4.665e+02, threshold=3.995e+02, percent-clipped=1.0 +2023-04-26 20:02:42,324 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2490, 1.5887, 1.4554, 1.8117, 1.7875, 1.9539, 1.4044, 3.7449], + device='cuda:3'), covar=tensor([0.0693, 0.0837, 0.0816, 0.1163, 0.0657, 0.0598, 0.0766, 0.0143], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 20:02:51,270 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-04-26 20:02:57,365 INFO [finetune.py:976] (3/7) Epoch 7, batch 4200, loss[loss=0.2687, simple_loss=0.3142, pruned_loss=0.1116, over 4149.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2693, pruned_loss=0.07072, over 951651.38 frames. ], batch size: 65, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:03:29,907 INFO [finetune.py:976] (3/7) Epoch 7, batch 4250, loss[loss=0.2046, simple_loss=0.2721, pruned_loss=0.0686, over 4830.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2663, pruned_loss=0.06931, over 952720.83 frames. ], batch size: 39, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:03:40,413 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.629e+02 2.024e+02 2.405e+02 4.304e+02, threshold=4.048e+02, percent-clipped=1.0 +2023-04-26 20:03:44,058 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38635.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:03:45,886 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7044, 1.6674, 0.8958, 1.4107, 1.9454, 1.5936, 1.4892, 1.4980], + device='cuda:3'), covar=tensor([0.0526, 0.0404, 0.0368, 0.0587, 0.0281, 0.0529, 0.0521, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 20:04:03,669 INFO [finetune.py:976] (3/7) Epoch 7, batch 4300, loss[loss=0.1803, simple_loss=0.2468, pruned_loss=0.05688, over 4791.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2638, pruned_loss=0.06897, over 954291.89 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:04:05,610 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4570, 1.4093, 4.3444, 4.0427, 3.7642, 4.0417, 3.9931, 3.8047], + device='cuda:3'), covar=tensor([0.6992, 0.5417, 0.0946, 0.1609, 0.1043, 0.1395, 0.1656, 0.1427], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0309, 0.0411, 0.0417, 0.0352, 0.0407, 0.0320, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:04:26,427 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38683.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:04:49,583 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 20:04:59,130 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1269, 3.2816, 1.0399, 1.5851, 1.6192, 2.2735, 1.8282, 1.0041], + device='cuda:3'), covar=tensor([0.2099, 0.1519, 0.2449, 0.2041, 0.1591, 0.1481, 0.1820, 0.2373], + device='cuda:3'), in_proj_covar=tensor([0.0119, 0.0254, 0.0143, 0.0125, 0.0136, 0.0156, 0.0120, 0.0123], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:05:02,798 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1455, 2.0228, 2.3974, 2.5537, 2.5839, 2.0779, 1.5857, 2.1154], + device='cuda:3'), covar=tensor([0.0947, 0.0996, 0.0509, 0.0591, 0.0596, 0.0868, 0.0900, 0.0670], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0204, 0.0181, 0.0178, 0.0179, 0.0193, 0.0162, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:05:03,901 INFO [finetune.py:976] (3/7) Epoch 7, batch 4350, loss[loss=0.1851, simple_loss=0.245, pruned_loss=0.06256, over 4829.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2604, pruned_loss=0.06753, over 954479.12 frames. ], batch size: 25, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:05:08,720 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38716.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:05:10,669 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3535, 2.1554, 2.6187, 2.8535, 2.4508, 2.2267, 2.3877, 2.3074], + device='cuda:3'), covar=tensor([0.6911, 0.9706, 1.0081, 0.9063, 0.8515, 1.2314, 1.1722, 1.0459], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0429, 0.0513, 0.0535, 0.0441, 0.0460, 0.0471, 0.0468], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:05:17,131 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-04-26 20:05:18,152 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.707e+02 2.023e+02 2.548e+02 4.531e+02, threshold=4.045e+02, percent-clipped=2.0 +2023-04-26 20:05:52,728 INFO [finetune.py:976] (3/7) Epoch 7, batch 4400, loss[loss=0.237, simple_loss=0.2997, pruned_loss=0.08721, over 4120.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2613, pruned_loss=0.06803, over 951669.59 frames. ], batch size: 65, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:05:59,896 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3331, 1.5746, 1.5654, 1.7165, 1.5762, 1.7199, 1.7131, 1.6224], + device='cuda:3'), covar=tensor([0.5862, 0.8582, 0.7393, 0.6949, 0.8230, 1.2170, 0.8651, 0.7924], + device='cuda:3'), in_proj_covar=tensor([0.0324, 0.0393, 0.0317, 0.0328, 0.0344, 0.0408, 0.0371, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 20:06:01,613 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8988, 1.3344, 1.7059, 1.9972, 1.6960, 1.3319, 1.0224, 1.4633], + device='cuda:3'), covar=tensor([0.4331, 0.4215, 0.2129, 0.3086, 0.3429, 0.3525, 0.5342, 0.3048], + device='cuda:3'), in_proj_covar=tensor([0.0279, 0.0253, 0.0218, 0.0322, 0.0214, 0.0229, 0.0236, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 20:06:26,655 INFO [finetune.py:976] (3/7) Epoch 7, batch 4450, loss[loss=0.2236, simple_loss=0.2982, pruned_loss=0.07453, over 4743.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2656, pruned_loss=0.06937, over 950872.64 frames. ], batch size: 54, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:06:36,506 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38823.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:06:45,522 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.756e+02 2.089e+02 2.730e+02 5.109e+02, threshold=4.178e+02, percent-clipped=5.0 +2023-04-26 20:07:37,020 INFO [finetune.py:976] (3/7) Epoch 7, batch 4500, loss[loss=0.2378, simple_loss=0.2946, pruned_loss=0.09047, over 4767.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2671, pruned_loss=0.0696, over 951759.22 frames. ], batch size: 28, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:08:00,074 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38884.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:08:44,169 INFO [finetune.py:976] (3/7) Epoch 7, batch 4550, loss[loss=0.2289, simple_loss=0.2916, pruned_loss=0.08314, over 4847.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2678, pruned_loss=0.06961, over 952869.51 frames. ], batch size: 49, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:08:58,073 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.848e+02 2.064e+02 2.542e+02 5.076e+02, threshold=4.127e+02, percent-clipped=2.0 +2023-04-26 20:09:50,305 INFO [finetune.py:976] (3/7) Epoch 7, batch 4600, loss[loss=0.1678, simple_loss=0.2355, pruned_loss=0.05006, over 4728.00 frames. ], tot_loss[loss=0.2015, simple_loss=0.2663, pruned_loss=0.06831, over 954156.66 frames. ], batch size: 54, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:10:56,078 INFO [finetune.py:976] (3/7) Epoch 7, batch 4650, loss[loss=0.1589, simple_loss=0.2376, pruned_loss=0.04012, over 4898.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2632, pruned_loss=0.06747, over 954318.90 frames. ], batch size: 32, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:10:56,167 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39016.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:11:16,121 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.581e+02 1.907e+02 2.292e+02 5.308e+02, threshold=3.814e+02, percent-clipped=1.0 +2023-04-26 20:12:02,799 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39064.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:12:03,961 INFO [finetune.py:976] (3/7) Epoch 7, batch 4700, loss[loss=0.1712, simple_loss=0.2336, pruned_loss=0.05443, over 4765.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2612, pruned_loss=0.0669, over 955227.81 frames. ], batch size: 27, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:12:11,537 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 20:12:22,824 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9653, 1.8590, 2.2446, 2.5161, 2.0943, 1.9000, 2.0884, 2.0797], + device='cuda:3'), covar=tensor([0.8387, 1.1131, 1.2209, 1.0650, 1.0009, 1.4311, 1.4516, 1.2260], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0431, 0.0514, 0.0536, 0.0441, 0.0460, 0.0472, 0.0469], + device='cuda:3'), out_proj_covar=tensor([9.9900e-05, 1.0686e-04, 1.1588e-04, 1.2696e-04, 1.0729e-04, 1.1136e-04, + 1.1361e-04, 1.1415e-04], device='cuda:3') +2023-04-26 20:13:04,272 INFO [finetune.py:976] (3/7) Epoch 7, batch 4750, loss[loss=0.2286, simple_loss=0.286, pruned_loss=0.08555, over 4906.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2571, pruned_loss=0.0651, over 954361.69 frames. ], batch size: 32, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:13:24,485 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.720e+02 2.202e+02 2.598e+02 5.909e+02, threshold=4.403e+02, percent-clipped=7.0 +2023-04-26 20:14:16,832 INFO [finetune.py:976] (3/7) Epoch 7, batch 4800, loss[loss=0.2077, simple_loss=0.2861, pruned_loss=0.06463, over 4895.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2607, pruned_loss=0.06653, over 954912.72 frames. ], batch size: 35, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:14:31,376 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39179.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:14:40,497 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39185.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:15:12,275 INFO [finetune.py:976] (3/7) Epoch 7, batch 4850, loss[loss=0.2259, simple_loss=0.2887, pruned_loss=0.08152, over 4889.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2637, pruned_loss=0.0677, over 952801.39 frames. ], batch size: 35, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:15:21,782 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.823e+02 2.117e+02 2.667e+02 5.725e+02, threshold=4.234e+02, percent-clipped=1.0 +2023-04-26 20:15:24,396 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 20:15:31,649 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39246.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:15:45,478 INFO [finetune.py:976] (3/7) Epoch 7, batch 4900, loss[loss=0.2063, simple_loss=0.2906, pruned_loss=0.06096, over 4758.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2652, pruned_loss=0.06837, over 952708.74 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:16:09,155 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39284.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:16:50,958 INFO [finetune.py:976] (3/7) Epoch 7, batch 4950, loss[loss=0.177, simple_loss=0.2403, pruned_loss=0.05685, over 4895.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2671, pruned_loss=0.06982, over 950669.62 frames. ], batch size: 35, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:17:06,089 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.715e+02 2.035e+02 2.513e+02 5.677e+02, threshold=4.070e+02, percent-clipped=1.0 +2023-04-26 20:17:25,348 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39345.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:17:49,759 INFO [finetune.py:976] (3/7) Epoch 7, batch 5000, loss[loss=0.1319, simple_loss=0.2008, pruned_loss=0.03145, over 4779.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2647, pruned_loss=0.06828, over 953601.56 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:18:23,438 INFO [finetune.py:976] (3/7) Epoch 7, batch 5050, loss[loss=0.1937, simple_loss=0.2543, pruned_loss=0.06651, over 4833.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2614, pruned_loss=0.06759, over 953199.92 frames. ], batch size: 30, lr: 3.85e-03, grad_scale: 64.0 +2023-04-26 20:18:33,402 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.684e+02 2.008e+02 2.415e+02 4.173e+02, threshold=4.016e+02, percent-clipped=2.0 +2023-04-26 20:18:54,478 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39463.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:18:56,693 INFO [finetune.py:976] (3/7) Epoch 7, batch 5100, loss[loss=0.1932, simple_loss=0.2584, pruned_loss=0.06403, over 4918.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2593, pruned_loss=0.06663, over 953020.08 frames. ], batch size: 36, lr: 3.85e-03, grad_scale: 64.0 +2023-04-26 20:19:02,592 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.9574, 4.0279, 2.9963, 4.6121, 4.0317, 4.0468, 1.8473, 3.8804], + device='cuda:3'), covar=tensor([0.1929, 0.1069, 0.3154, 0.1400, 0.3432, 0.1711, 0.5860, 0.2270], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0219, 0.0254, 0.0310, 0.0304, 0.0255, 0.0277, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 20:19:04,445 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4183, 1.6123, 1.6863, 2.3398, 2.5729, 2.1302, 2.0016, 1.8095], + device='cuda:3'), covar=tensor([0.1908, 0.2176, 0.2349, 0.1322, 0.1349, 0.1984, 0.2901, 0.2363], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0325, 0.0352, 0.0299, 0.0338, 0.0324, 0.0306, 0.0352], + device='cuda:3'), out_proj_covar=tensor([6.5074e-05, 6.9097e-05, 7.6240e-05, 6.2074e-05, 7.1095e-05, 6.9765e-05, + 6.6095e-05, 7.5560e-05], device='cuda:3') +2023-04-26 20:19:05,637 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39479.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:19:28,174 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1907, 1.5352, 1.7797, 1.9029, 2.3618, 1.9937, 1.7052, 1.6890], + device='cuda:3'), covar=tensor([0.1715, 0.2074, 0.2343, 0.1354, 0.0942, 0.1479, 0.2392, 0.2036], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0326, 0.0354, 0.0300, 0.0338, 0.0325, 0.0307, 0.0353], + device='cuda:3'), out_proj_covar=tensor([6.5225e-05, 6.9290e-05, 7.6603e-05, 6.2214e-05, 7.1182e-05, 6.9999e-05, + 6.6280e-05, 7.5845e-05], device='cuda:3') +2023-04-26 20:19:29,889 INFO [finetune.py:976] (3/7) Epoch 7, batch 5150, loss[loss=0.2468, simple_loss=0.3135, pruned_loss=0.09006, over 4808.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2585, pruned_loss=0.06637, over 953477.16 frames. ], batch size: 41, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:19:35,831 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39524.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:19:38,069 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39527.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:19:40,454 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.309e+02 1.798e+02 2.123e+02 2.599e+02 5.486e+02, threshold=4.247e+02, percent-clipped=4.0 +2023-04-26 20:19:47,595 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39541.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:20:03,309 INFO [finetune.py:976] (3/7) Epoch 7, batch 5200, loss[loss=0.298, simple_loss=0.3421, pruned_loss=0.127, over 4190.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2604, pruned_loss=0.06646, over 950684.13 frames. ], batch size: 65, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:20:37,217 INFO [finetune.py:976] (3/7) Epoch 7, batch 5250, loss[loss=0.1878, simple_loss=0.2548, pruned_loss=0.06041, over 4814.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2622, pruned_loss=0.06662, over 949085.31 frames. ], batch size: 51, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:20:47,796 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.203e+02 1.776e+02 2.030e+02 2.603e+02 8.469e+02, threshold=4.060e+02, percent-clipped=1.0 +2023-04-26 20:20:53,194 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39639.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 20:20:53,792 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39640.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:21:10,371 INFO [finetune.py:976] (3/7) Epoch 7, batch 5300, loss[loss=0.2023, simple_loss=0.2593, pruned_loss=0.07267, over 4762.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2656, pruned_loss=0.06842, over 950442.15 frames. ], batch size: 27, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:21:19,578 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.86 vs. limit=5.0 +2023-04-26 20:21:50,864 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39700.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:22:11,714 INFO [finetune.py:976] (3/7) Epoch 7, batch 5350, loss[loss=0.2052, simple_loss=0.2692, pruned_loss=0.07062, over 4925.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2658, pruned_loss=0.068, over 952177.75 frames. ], batch size: 33, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:22:31,524 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.675e+02 2.001e+02 2.311e+02 3.893e+02, threshold=4.002e+02, percent-clipped=0.0 +2023-04-26 20:23:17,037 INFO [finetune.py:976] (3/7) Epoch 7, batch 5400, loss[loss=0.194, simple_loss=0.259, pruned_loss=0.06447, over 4817.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2631, pruned_loss=0.06685, over 951526.72 frames. ], batch size: 41, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:23:17,747 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39767.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:23:38,533 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8460, 2.1092, 1.1541, 1.6089, 2.3302, 1.7297, 1.6564, 1.6571], + device='cuda:3'), covar=tensor([0.0497, 0.0350, 0.0308, 0.0548, 0.0235, 0.0481, 0.0483, 0.0572], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0049], + device='cuda:3') +2023-04-26 20:24:18,123 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39815.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:24:18,613 INFO [finetune.py:976] (3/7) Epoch 7, batch 5450, loss[loss=0.1824, simple_loss=0.2466, pruned_loss=0.05912, over 4815.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2606, pruned_loss=0.06611, over 951742.11 frames. ], batch size: 51, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:24:20,494 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39819.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:24:31,590 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39828.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:24:33,271 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.624e+02 1.910e+02 2.215e+02 3.764e+02, threshold=3.819e+02, percent-clipped=0.0 +2023-04-26 20:24:40,458 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2972, 1.5272, 1.3849, 1.7378, 1.5958, 2.0329, 1.3372, 3.6234], + device='cuda:3'), covar=tensor([0.0632, 0.0789, 0.0815, 0.1190, 0.0670, 0.0504, 0.0776, 0.0139], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0040, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 20:24:40,465 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39841.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:24:57,926 INFO [finetune.py:976] (3/7) Epoch 7, batch 5500, loss[loss=0.1897, simple_loss=0.2418, pruned_loss=0.06877, over 4822.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2576, pruned_loss=0.0652, over 953443.74 frames. ], batch size: 30, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:25:04,135 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39876.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:25:12,055 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39889.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:25:31,741 INFO [finetune.py:976] (3/7) Epoch 7, batch 5550, loss[loss=0.1954, simple_loss=0.2682, pruned_loss=0.06129, over 4913.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2623, pruned_loss=0.06824, over 954490.27 frames. ], batch size: 43, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:25:40,907 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.695e+02 2.043e+02 2.631e+02 5.173e+02, threshold=4.085e+02, percent-clipped=3.0 +2023-04-26 20:25:46,475 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:26:02,970 INFO [finetune.py:976] (3/7) Epoch 7, batch 5600, loss[loss=0.2055, simple_loss=0.2586, pruned_loss=0.07615, over 4739.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2657, pruned_loss=0.06913, over 956001.39 frames. ], batch size: 23, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:26:07,778 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4290, 1.3465, 1.7103, 1.6731, 1.3503, 1.0874, 1.5164, 0.9945], + device='cuda:3'), covar=tensor([0.0746, 0.0717, 0.0459, 0.0630, 0.0756, 0.1160, 0.0644, 0.0859], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0074, 0.0073, 0.0067, 0.0077, 0.0096, 0.0080, 0.0076], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:26:15,866 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39988.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:26:19,989 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39995.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 20:26:38,931 INFO [finetune.py:976] (3/7) Epoch 7, batch 5650, loss[loss=0.2262, simple_loss=0.296, pruned_loss=0.07824, over 4802.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2689, pruned_loss=0.07019, over 954673.66 frames. ], batch size: 45, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:26:53,979 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.232e+02 1.674e+02 2.052e+02 2.439e+02 4.352e+02, threshold=4.105e+02, percent-clipped=2.0 +2023-04-26 20:27:27,068 INFO [finetune.py:976] (3/7) Epoch 7, batch 5700, loss[loss=0.1813, simple_loss=0.2373, pruned_loss=0.06266, over 4051.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.265, pruned_loss=0.07015, over 936129.61 frames. ], batch size: 17, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:27:38,787 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2846, 1.6230, 5.4783, 5.1076, 4.8307, 5.2474, 4.9260, 4.7515], + device='cuda:3'), covar=tensor([0.5887, 0.6227, 0.1076, 0.1980, 0.0978, 0.1063, 0.0856, 0.1544], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0307, 0.0408, 0.0413, 0.0348, 0.0406, 0.0318, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:28:14,054 INFO [finetune.py:976] (3/7) Epoch 8, batch 0, loss[loss=0.2215, simple_loss=0.281, pruned_loss=0.08105, over 4870.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.281, pruned_loss=0.08105, over 4870.00 frames. ], batch size: 32, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:28:14,054 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 20:28:25,776 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3804, 1.2462, 1.6305, 1.5698, 1.2791, 1.1373, 1.3193, 0.8292], + device='cuda:3'), covar=tensor([0.0731, 0.0915, 0.0583, 0.0804, 0.0936, 0.1288, 0.0791, 0.0873], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0074, 0.0073, 0.0067, 0.0077, 0.0096, 0.0080, 0.0075], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:28:30,560 INFO [finetune.py:1010] (3/7) Epoch 8, validation: loss=0.1574, simple_loss=0.2299, pruned_loss=0.04247, over 2265189.00 frames. +2023-04-26 20:28:30,561 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 20:29:03,018 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40119.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:03,420 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-26 20:29:05,430 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40123.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:10,787 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.752e+02 2.089e+02 2.536e+02 4.447e+02, threshold=4.178e+02, percent-clipped=1.0 +2023-04-26 20:29:20,979 INFO [finetune.py:976] (3/7) Epoch 8, batch 50, loss[loss=0.2044, simple_loss=0.2621, pruned_loss=0.07331, over 4843.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2719, pruned_loss=0.07055, over 217849.78 frames. ], batch size: 44, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:29:24,045 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7233, 1.9928, 1.1208, 1.5038, 2.2426, 1.5847, 1.4989, 1.6395], + device='cuda:3'), covar=tensor([0.0534, 0.0375, 0.0346, 0.0561, 0.0250, 0.0557, 0.0557, 0.0602], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 20:29:36,058 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40167.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:38,547 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40171.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:41,451 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-26 20:29:42,170 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6495, 2.6807, 2.2050, 3.0874, 2.6957, 2.6563, 1.3193, 2.6340], + device='cuda:3'), covar=tensor([0.2571, 0.1733, 0.4239, 0.3179, 0.2979, 0.2378, 0.5069, 0.2953], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0218, 0.0253, 0.0310, 0.0303, 0.0253, 0.0274, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 20:29:54,388 INFO [finetune.py:976] (3/7) Epoch 8, batch 100, loss[loss=0.1655, simple_loss=0.2419, pruned_loss=0.04459, over 4898.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2645, pruned_loss=0.06879, over 382309.00 frames. ], batch size: 35, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:29:55,590 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40195.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:01,237 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6308, 1.2797, 1.3081, 1.5075, 1.8614, 1.5467, 1.3803, 1.2597], + device='cuda:3'), covar=tensor([0.1363, 0.1362, 0.1688, 0.1256, 0.0678, 0.1290, 0.1845, 0.1706], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0328, 0.0357, 0.0303, 0.0341, 0.0328, 0.0309, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.5548e-05, 6.9670e-05, 7.7417e-05, 6.2802e-05, 7.1811e-05, 7.0622e-05, + 6.6724e-05, 7.5959e-05], device='cuda:3') +2023-04-26 20:30:10,530 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40218.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:18,316 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.606e+02 1.925e+02 2.421e+02 4.111e+02, threshold=3.850e+02, percent-clipped=0.0 +2023-04-26 20:30:28,097 INFO [finetune.py:976] (3/7) Epoch 8, batch 150, loss[loss=0.1802, simple_loss=0.235, pruned_loss=0.06275, over 4825.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2587, pruned_loss=0.0663, over 508439.55 frames. ], batch size: 25, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:30:36,430 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40256.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:46,201 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4077, 2.4565, 1.9861, 2.2281, 2.5237, 2.0388, 3.4668, 1.9371], + device='cuda:3'), covar=tensor([0.4781, 0.2579, 0.5765, 0.3795, 0.2276, 0.3153, 0.1680, 0.4528], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0350, 0.0432, 0.0361, 0.0386, 0.0380, 0.0382, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:30:46,203 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40272.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:50,429 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40279.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:31:01,726 INFO [finetune.py:976] (3/7) Epoch 8, batch 200, loss[loss=0.2502, simple_loss=0.3066, pruned_loss=0.09695, over 4800.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2579, pruned_loss=0.06654, over 607924.56 frames. ], batch size: 45, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:31:02,440 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40295.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:31:24,762 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40330.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:31:25,242 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.700e+02 1.993e+02 2.528e+02 5.564e+02, threshold=3.985e+02, percent-clipped=2.0 +2023-04-26 20:31:26,599 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40333.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:31:34,054 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40343.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:31:35,038 INFO [finetune.py:976] (3/7) Epoch 8, batch 250, loss[loss=0.2691, simple_loss=0.3264, pruned_loss=0.1058, over 4825.00 frames. ], tot_loss[loss=0.1988, simple_loss=0.262, pruned_loss=0.06778, over 685778.90 frames. ], batch size: 38, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:32:05,774 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40391.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:32:07,909 INFO [finetune.py:976] (3/7) Epoch 8, batch 300, loss[loss=0.2348, simple_loss=0.274, pruned_loss=0.09778, over 4879.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.2658, pruned_loss=0.06978, over 745485.86 frames. ], batch size: 31, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:32:27,221 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40423.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:32:31,997 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.648e+02 2.026e+02 2.422e+02 3.959e+02, threshold=4.051e+02, percent-clipped=0.0 +2023-04-26 20:32:45,857 INFO [finetune.py:976] (3/7) Epoch 8, batch 350, loss[loss=0.1562, simple_loss=0.2174, pruned_loss=0.04749, over 4709.00 frames. ], tot_loss[loss=0.2024, simple_loss=0.2659, pruned_loss=0.06948, over 792805.32 frames. ], batch size: 23, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:33:27,251 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-26 20:33:27,707 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:33:27,755 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:33:52,001 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40492.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:33:53,113 INFO [finetune.py:976] (3/7) Epoch 8, batch 400, loss[loss=0.2502, simple_loss=0.3147, pruned_loss=0.09282, over 4765.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2665, pruned_loss=0.06924, over 827279.99 frames. ], batch size: 28, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:34:03,314 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7032, 2.2570, 1.6713, 1.4421, 1.2689, 1.3246, 1.7102, 1.1561], + device='cuda:3'), covar=tensor([0.1806, 0.1553, 0.1610, 0.2147, 0.2592, 0.2178, 0.1213, 0.2326], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0218, 0.0173, 0.0205, 0.0206, 0.0186, 0.0163, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 20:34:24,706 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8397, 1.4432, 4.4347, 3.8140, 3.9356, 4.2577, 4.0518, 3.8539], + device='cuda:3'), covar=tensor([0.8147, 0.7803, 0.1467, 0.3279, 0.2162, 0.2564, 0.2806, 0.2849], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0307, 0.0408, 0.0414, 0.0349, 0.0405, 0.0317, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:34:32,788 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40519.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:34:33,466 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9132, 2.8134, 1.9794, 2.0211, 1.4301, 1.4438, 2.1334, 1.4124], + device='cuda:3'), covar=tensor([0.1841, 0.1599, 0.1667, 0.2076, 0.2582, 0.2094, 0.1201, 0.2181], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0218, 0.0173, 0.0205, 0.0207, 0.0186, 0.0164, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 20:34:45,210 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.623e+02 1.962e+02 2.378e+02 5.703e+02, threshold=3.923e+02, percent-clipped=1.0 +2023-04-26 20:35:05,023 INFO [finetune.py:976] (3/7) Epoch 8, batch 450, loss[loss=0.1633, simple_loss=0.2338, pruned_loss=0.04636, over 4816.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2658, pruned_loss=0.06909, over 856362.67 frames. ], batch size: 39, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:35:17,318 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:35:18,623 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40553.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:35:37,624 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9629, 1.8903, 2.1515, 2.4204, 1.8367, 1.4417, 1.9776, 1.0451], + device='cuda:3'), covar=tensor([0.0690, 0.0909, 0.0644, 0.0880, 0.0862, 0.1210, 0.0883, 0.1111], + device='cuda:3'), in_proj_covar=tensor([0.0065, 0.0074, 0.0072, 0.0066, 0.0076, 0.0095, 0.0080, 0.0075], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:35:38,790 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40574.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:35:55,303 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2723, 1.3459, 1.4165, 1.6157, 1.6731, 1.3322, 1.0058, 1.4455], + device='cuda:3'), covar=tensor([0.0800, 0.1154, 0.0771, 0.0603, 0.0558, 0.0770, 0.0787, 0.0581], + device='cuda:3'), in_proj_covar=tensor([0.0202, 0.0209, 0.0185, 0.0182, 0.0182, 0.0196, 0.0165, 0.0191], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:35:56,422 INFO [finetune.py:976] (3/7) Epoch 8, batch 500, loss[loss=0.1829, simple_loss=0.2477, pruned_loss=0.05902, over 4834.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2634, pruned_loss=0.06809, over 880676.47 frames. ], batch size: 30, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:36:37,616 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40628.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:36:39,369 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.684e+02 2.034e+02 2.425e+02 5.158e+02, threshold=4.068e+02, percent-clipped=3.0 +2023-04-26 20:36:52,502 INFO [finetune.py:976] (3/7) Epoch 8, batch 550, loss[loss=0.1872, simple_loss=0.2497, pruned_loss=0.0623, over 4819.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2612, pruned_loss=0.06755, over 896799.16 frames. ], batch size: 33, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:36:54,503 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7429, 1.7578, 1.9902, 2.2500, 2.2169, 1.8145, 1.4989, 1.8204], + device='cuda:3'), covar=tensor([0.0930, 0.1083, 0.0640, 0.0612, 0.0587, 0.0882, 0.0839, 0.0691], + device='cuda:3'), in_proj_covar=tensor([0.0201, 0.0208, 0.0184, 0.0182, 0.0181, 0.0195, 0.0165, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:36:56,970 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40651.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:37:25,624 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-26 20:37:29,153 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 20:37:47,177 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40686.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:37:52,037 INFO [finetune.py:976] (3/7) Epoch 8, batch 600, loss[loss=0.2231, simple_loss=0.2914, pruned_loss=0.07738, over 4922.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2623, pruned_loss=0.06753, over 911407.53 frames. ], batch size: 38, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:37:57,817 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8230, 3.7397, 2.8734, 4.3971, 3.7538, 3.7609, 1.5304, 3.8219], + device='cuda:3'), covar=tensor([0.1748, 0.1306, 0.3304, 0.1282, 0.3912, 0.1792, 0.6041, 0.2280], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0219, 0.0253, 0.0309, 0.0303, 0.0254, 0.0274, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 20:38:21,419 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40712.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:38:22,270 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-26 20:38:32,783 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5360, 1.7625, 2.3732, 2.9420, 2.2273, 1.8053, 1.6216, 2.2446], + device='cuda:3'), covar=tensor([0.4038, 0.4307, 0.1958, 0.3430, 0.3886, 0.3258, 0.4790, 0.2900], + device='cuda:3'), in_proj_covar=tensor([0.0279, 0.0253, 0.0218, 0.0320, 0.0214, 0.0228, 0.0237, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 20:38:44,483 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.285e+02 1.754e+02 2.110e+02 2.532e+02 4.405e+02, threshold=4.220e+02, percent-clipped=2.0 +2023-04-26 20:38:54,530 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4098, 1.3789, 1.7390, 1.6869, 1.3725, 1.0712, 1.5278, 1.0224], + device='cuda:3'), covar=tensor([0.0818, 0.0905, 0.0491, 0.0879, 0.0893, 0.1306, 0.0764, 0.0867], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0074, 0.0072, 0.0067, 0.0076, 0.0096, 0.0080, 0.0075], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:39:02,721 INFO [finetune.py:976] (3/7) Epoch 8, batch 650, loss[loss=0.1654, simple_loss=0.2307, pruned_loss=0.05008, over 4827.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2651, pruned_loss=0.0684, over 921709.72 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:39:14,086 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-26 20:39:36,214 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0047, 1.2347, 1.7588, 2.2870, 1.8078, 1.3877, 1.1073, 1.5655], + device='cuda:3'), covar=tensor([0.4931, 0.5410, 0.2526, 0.3499, 0.3953, 0.3627, 0.5650, 0.3421], + device='cuda:3'), in_proj_covar=tensor([0.0280, 0.0254, 0.0219, 0.0322, 0.0215, 0.0229, 0.0238, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 20:40:08,310 INFO [finetune.py:976] (3/7) Epoch 8, batch 700, loss[loss=0.2222, simple_loss=0.278, pruned_loss=0.08318, over 4798.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2651, pruned_loss=0.06788, over 928828.39 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:40:22,183 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 20:40:24,131 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6961, 1.4566, 1.8417, 1.9161, 1.4974, 1.2502, 1.5681, 0.9324], + device='cuda:3'), covar=tensor([0.0675, 0.0842, 0.0512, 0.0767, 0.0973, 0.1403, 0.0817, 0.0974], + device='cuda:3'), in_proj_covar=tensor([0.0065, 0.0074, 0.0072, 0.0067, 0.0076, 0.0095, 0.0080, 0.0075], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:40:55,212 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.764e+02 2.066e+02 2.587e+02 3.741e+02, threshold=4.132e+02, percent-clipped=0.0 +2023-04-26 20:41:14,341 INFO [finetune.py:976] (3/7) Epoch 8, batch 750, loss[loss=0.2201, simple_loss=0.2926, pruned_loss=0.07382, over 4894.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2656, pruned_loss=0.06806, over 933938.95 frames. ], batch size: 36, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:41:16,849 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:17,726 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-26 20:41:24,326 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40851.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:38,681 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40865.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:44,667 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40874.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:57,693 INFO [finetune.py:976] (3/7) Epoch 8, batch 800, loss[loss=0.2066, simple_loss=0.2629, pruned_loss=0.07517, over 4820.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2637, pruned_loss=0.06639, over 938442.12 frames. ], batch size: 39, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:42:00,808 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40899.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:01,456 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40900.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:16,691 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40922.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:19,180 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40926.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:20,387 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40928.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:22,148 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.615e+02 1.856e+02 2.266e+02 4.587e+02, threshold=3.712e+02, percent-clipped=2.0 +2023-04-26 20:42:31,002 INFO [finetune.py:976] (3/7) Epoch 8, batch 850, loss[loss=0.1546, simple_loss=0.2205, pruned_loss=0.04432, over 4908.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2608, pruned_loss=0.06544, over 942688.27 frames. ], batch size: 43, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:42:41,365 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40961.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:42:48,586 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 20:42:52,276 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40976.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:58,799 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40986.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:03,447 INFO [finetune.py:976] (3/7) Epoch 8, batch 900, loss[loss=0.21, simple_loss=0.2537, pruned_loss=0.08318, over 4811.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2584, pruned_loss=0.06489, over 947672.99 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:43:05,783 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40997.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:12,072 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41007.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:28,963 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.671e+02 1.992e+02 2.459e+02 6.072e+02, threshold=3.985e+02, percent-clipped=2.0 +2023-04-26 20:43:30,876 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41034.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:35,080 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-26 20:43:37,319 INFO [finetune.py:976] (3/7) Epoch 8, batch 950, loss[loss=0.2077, simple_loss=0.2724, pruned_loss=0.07152, over 4797.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2587, pruned_loss=0.06564, over 951164.85 frames. ], batch size: 29, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:43:38,029 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41045.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:46,385 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41058.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:10,578 INFO [finetune.py:976] (3/7) Epoch 8, batch 1000, loss[loss=0.3079, simple_loss=0.3432, pruned_loss=0.1363, over 4109.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2622, pruned_loss=0.06737, over 952961.45 frames. ], batch size: 65, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:44:16,011 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41102.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:18,377 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41106.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:20,498 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 20:44:35,825 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.783e+02 2.181e+02 2.591e+02 5.810e+02, threshold=4.362e+02, percent-clipped=3.0 +2023-04-26 20:44:44,111 INFO [finetune.py:976] (3/7) Epoch 8, batch 1050, loss[loss=0.1589, simple_loss=0.232, pruned_loss=0.04287, over 4770.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.265, pruned_loss=0.068, over 953728.28 frames. ], batch size: 28, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:44:46,606 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41148.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:55,277 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 20:44:56,156 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41163.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:45:00,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6276, 0.9733, 1.5445, 2.0873, 1.7728, 1.5375, 1.5710, 1.6358], + device='cuda:3'), covar=tensor([0.7249, 0.9886, 1.0384, 0.9926, 0.9094, 1.1937, 1.1958, 0.9560], + device='cuda:3'), in_proj_covar=tensor([0.0410, 0.0429, 0.0512, 0.0532, 0.0440, 0.0459, 0.0472, 0.0469], + device='cuda:3'), out_proj_covar=tensor([9.9578e-05, 1.0633e-04, 1.1558e-04, 1.2623e-04, 1.0695e-04, 1.1116e-04, + 1.1364e-04, 1.1390e-04], device='cuda:3') +2023-04-26 20:45:27,979 INFO [finetune.py:976] (3/7) Epoch 8, batch 1100, loss[loss=0.2139, simple_loss=0.2694, pruned_loss=0.07922, over 4767.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2658, pruned_loss=0.06816, over 951848.09 frames. ], batch size: 28, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:45:29,273 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41196.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:45:50,740 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41221.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:45:58,234 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.713e+02 2.125e+02 2.514e+02 5.728e+02, threshold=4.249e+02, percent-clipped=2.0 +2023-04-26 20:46:17,813 INFO [finetune.py:976] (3/7) Epoch 8, batch 1150, loss[loss=0.2334, simple_loss=0.2925, pruned_loss=0.08715, over 4814.00 frames. ], tot_loss[loss=0.2023, simple_loss=0.2674, pruned_loss=0.06858, over 954324.17 frames. ], batch size: 33, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:46:27,973 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 20:46:29,756 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6328, 1.4704, 1.9235, 1.9839, 1.4460, 1.2284, 1.6877, 1.0490], + device='cuda:3'), covar=tensor([0.0576, 0.0944, 0.0449, 0.0869, 0.0946, 0.1318, 0.0917, 0.0927], + device='cuda:3'), in_proj_covar=tensor([0.0065, 0.0073, 0.0072, 0.0066, 0.0076, 0.0095, 0.0080, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:46:36,332 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41256.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 20:46:51,088 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5395, 1.4591, 1.7864, 1.8047, 1.4859, 1.2443, 1.6050, 1.0480], + device='cuda:3'), covar=tensor([0.0731, 0.0799, 0.0560, 0.0756, 0.0803, 0.1231, 0.0766, 0.0895], + device='cuda:3'), in_proj_covar=tensor([0.0065, 0.0073, 0.0072, 0.0066, 0.0076, 0.0095, 0.0079, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 20:47:24,549 INFO [finetune.py:976] (3/7) Epoch 8, batch 1200, loss[loss=0.1614, simple_loss=0.2325, pruned_loss=0.04513, over 4766.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2661, pruned_loss=0.06822, over 954856.50 frames. ], batch size: 26, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:47:44,084 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41307.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:47:49,502 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2401, 1.6875, 1.6628, 2.0484, 1.8727, 1.9360, 1.5214, 4.3519], + device='cuda:3'), covar=tensor([0.0618, 0.0781, 0.0751, 0.1135, 0.0633, 0.0598, 0.0737, 0.0103], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0041, 0.0040, 0.0039, 0.0060], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 20:47:51,050 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-26 20:47:55,032 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7148, 1.8372, 0.8129, 1.4077, 1.9247, 1.5796, 1.4497, 1.6073], + device='cuda:3'), covar=tensor([0.0505, 0.0383, 0.0384, 0.0584, 0.0263, 0.0569, 0.0571, 0.0551], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 20:47:59,150 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.672e+02 1.949e+02 2.366e+02 5.490e+02, threshold=3.899e+02, percent-clipped=2.0 +2023-04-26 20:48:08,614 INFO [finetune.py:976] (3/7) Epoch 8, batch 1250, loss[loss=0.191, simple_loss=0.2537, pruned_loss=0.06412, over 4942.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2636, pruned_loss=0.06742, over 955741.68 frames. ], batch size: 38, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:48:14,609 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41353.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:48:15,821 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41355.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:48:42,247 INFO [finetune.py:976] (3/7) Epoch 8, batch 1300, loss[loss=0.2121, simple_loss=0.2769, pruned_loss=0.07367, over 4894.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2607, pruned_loss=0.06629, over 955713.57 frames. ], batch size: 35, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:48:46,732 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41401.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:48:53,976 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0189, 2.3418, 1.1672, 1.3876, 1.8489, 1.1849, 3.1248, 1.6672], + device='cuda:3'), covar=tensor([0.0650, 0.0791, 0.0773, 0.1134, 0.0467, 0.0962, 0.0224, 0.0602], + device='cuda:3'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0052, 0.0053, 0.0080, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 20:49:05,847 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.753e+02 2.019e+02 2.733e+02 6.111e+02, threshold=4.038e+02, percent-clipped=8.0 +2023-04-26 20:49:15,123 INFO [finetune.py:976] (3/7) Epoch 8, batch 1350, loss[loss=0.1444, simple_loss=0.2154, pruned_loss=0.03667, over 4814.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2604, pruned_loss=0.06618, over 955548.59 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:49:25,038 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:49:48,139 INFO [finetune.py:976] (3/7) Epoch 8, batch 1400, loss[loss=0.2291, simple_loss=0.2943, pruned_loss=0.082, over 4751.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2639, pruned_loss=0.06703, over 957526.44 frames. ], batch size: 27, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:50:04,311 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 20:50:06,414 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41521.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:50:12,254 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-26 20:50:12,930 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.248e+02 1.632e+02 2.112e+02 2.588e+02 6.346e+02, threshold=4.225e+02, percent-clipped=4.0 +2023-04-26 20:50:21,307 INFO [finetune.py:976] (3/7) Epoch 8, batch 1450, loss[loss=0.2721, simple_loss=0.3139, pruned_loss=0.1151, over 4864.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2644, pruned_loss=0.06679, over 956455.49 frames. ], batch size: 49, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:50:30,555 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41556.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:50:38,928 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41569.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:50:54,400 INFO [finetune.py:976] (3/7) Epoch 8, batch 1500, loss[loss=0.1717, simple_loss=0.249, pruned_loss=0.04719, over 4764.00 frames. ], tot_loss[loss=0.198, simple_loss=0.264, pruned_loss=0.06607, over 957455.88 frames. ], batch size: 28, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:51:02,367 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41604.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:51:25,120 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.252e+02 1.649e+02 2.032e+02 2.388e+02 6.025e+02, threshold=4.063e+02, percent-clipped=2.0 +2023-04-26 20:51:38,889 INFO [finetune.py:976] (3/7) Epoch 8, batch 1550, loss[loss=0.2078, simple_loss=0.2836, pruned_loss=0.06594, over 4750.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2655, pruned_loss=0.06646, over 957354.75 frames. ], batch size: 26, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:51:51,057 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41653.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:51:58,439 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:52:47,105 INFO [finetune.py:976] (3/7) Epoch 8, batch 1600, loss[loss=0.1946, simple_loss=0.2542, pruned_loss=0.06746, over 4913.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2633, pruned_loss=0.06604, over 955790.68 frames. ], batch size: 36, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:52:56,991 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:52:57,035 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:53:12,222 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41715.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:53:15,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5693, 1.9819, 1.7068, 1.9169, 1.5180, 1.5845, 1.7021, 1.3356], + device='cuda:3'), covar=tensor([0.1790, 0.1324, 0.0889, 0.1082, 0.3084, 0.1282, 0.1724, 0.2242], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0325, 0.0235, 0.0296, 0.0321, 0.0277, 0.0264, 0.0288], + device='cuda:3'), out_proj_covar=tensor([1.2375e-04, 1.3118e-04, 9.5080e-05, 1.1854e-04, 1.3213e-04, 1.1179e-04, + 1.0828e-04, 1.1586e-04], device='cuda:3') +2023-04-26 20:53:22,287 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.620e+02 1.919e+02 2.261e+02 4.784e+02, threshold=3.838e+02, percent-clipped=1.0 +2023-04-26 20:53:30,671 INFO [finetune.py:976] (3/7) Epoch 8, batch 1650, loss[loss=0.2405, simple_loss=0.2843, pruned_loss=0.09834, over 4905.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2602, pruned_loss=0.06523, over 956765.48 frames. ], batch size: 37, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:53:33,799 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41749.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:53:40,291 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41758.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:54:03,921 INFO [finetune.py:976] (3/7) Epoch 8, batch 1700, loss[loss=0.1687, simple_loss=0.2309, pruned_loss=0.05319, over 4817.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2574, pruned_loss=0.06424, over 956186.88 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:54:11,554 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41806.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:54:16,321 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7273, 1.5920, 1.7915, 2.0170, 2.0986, 1.6012, 1.3343, 1.7876], + device='cuda:3'), covar=tensor([0.0842, 0.1090, 0.0669, 0.0531, 0.0543, 0.0869, 0.0814, 0.0605], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0206, 0.0183, 0.0179, 0.0179, 0.0193, 0.0162, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:54:29,248 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.826e+02 2.142e+02 2.567e+02 5.646e+02, threshold=4.284e+02, percent-clipped=4.0 +2023-04-26 20:54:36,815 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-26 20:54:37,046 INFO [finetune.py:976] (3/7) Epoch 8, batch 1750, loss[loss=0.1958, simple_loss=0.2726, pruned_loss=0.05954, over 4901.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2606, pruned_loss=0.06574, over 957926.78 frames. ], batch size: 37, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:54:49,219 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 20:55:10,267 INFO [finetune.py:976] (3/7) Epoch 8, batch 1800, loss[loss=0.2038, simple_loss=0.2656, pruned_loss=0.07097, over 4896.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2637, pruned_loss=0.06709, over 958853.53 frames. ], batch size: 35, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:55:18,165 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41906.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:55:18,376 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-26 20:55:35,584 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.841e+02 2.087e+02 2.578e+02 5.698e+02, threshold=4.173e+02, percent-clipped=3.0 +2023-04-26 20:55:40,971 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9732, 1.9744, 1.8919, 1.6988, 2.1846, 1.6053, 2.7614, 1.5872], + device='cuda:3'), covar=tensor([0.4149, 0.1970, 0.5047, 0.3021, 0.1796, 0.2766, 0.1425, 0.4652], + device='cuda:3'), in_proj_covar=tensor([0.0350, 0.0354, 0.0436, 0.0365, 0.0392, 0.0386, 0.0385, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 20:55:43,856 INFO [finetune.py:976] (3/7) Epoch 8, batch 1850, loss[loss=0.1797, simple_loss=0.2568, pruned_loss=0.05132, over 4897.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2646, pruned_loss=0.06738, over 957804.28 frames. ], batch size: 35, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:55:46,369 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41948.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 20:55:58,820 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41967.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:56:17,001 INFO [finetune.py:976] (3/7) Epoch 8, batch 1900, loss[loss=0.1547, simple_loss=0.2293, pruned_loss=0.04001, over 4751.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2644, pruned_loss=0.06724, over 954115.36 frames. ], batch size: 26, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:56:27,977 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42009.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 20:56:28,528 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42010.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 20:56:49,877 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.688e+02 2.018e+02 2.426e+02 3.814e+02, threshold=4.036e+02, percent-clipped=0.0 +2023-04-26 20:57:08,688 INFO [finetune.py:976] (3/7) Epoch 8, batch 1950, loss[loss=0.1911, simple_loss=0.2673, pruned_loss=0.05743, over 4760.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2638, pruned_loss=0.06669, over 952826.77 frames. ], batch size: 27, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:57:08,849 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8224, 1.0485, 1.3722, 1.5093, 1.4377, 1.6133, 1.4151, 1.3975], + device='cuda:3'), covar=tensor([0.5323, 0.6695, 0.5864, 0.5556, 0.6799, 0.9899, 0.7073, 0.6372], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0387, 0.0317, 0.0328, 0.0343, 0.0407, 0.0369, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 20:57:32,057 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7616, 1.9513, 0.7961, 1.3874, 2.0873, 1.6218, 1.5312, 1.5998], + device='cuda:3'), covar=tensor([0.0533, 0.0386, 0.0391, 0.0582, 0.0262, 0.0554, 0.0519, 0.0629], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 20:58:13,786 INFO [finetune.py:976] (3/7) Epoch 8, batch 2000, loss[loss=0.1721, simple_loss=0.2405, pruned_loss=0.05188, over 4724.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2615, pruned_loss=0.06601, over 953901.46 frames. ], batch size: 59, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:58:36,349 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9054, 2.2705, 0.9371, 1.1949, 1.6010, 1.0597, 2.4058, 1.3711], + device='cuda:3'), covar=tensor([0.0674, 0.0636, 0.0687, 0.1264, 0.0427, 0.0974, 0.0324, 0.0680], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 20:59:06,780 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.619e+02 1.918e+02 2.281e+02 4.896e+02, threshold=3.837e+02, percent-clipped=1.0 +2023-04-26 20:59:17,301 INFO [finetune.py:976] (3/7) Epoch 8, batch 2050, loss[loss=0.165, simple_loss=0.2319, pruned_loss=0.04908, over 4776.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2583, pruned_loss=0.06514, over 954884.70 frames. ], batch size: 26, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:59:17,519 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 20:59:36,105 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 20:59:50,407 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1410, 1.5098, 2.0079, 2.4423, 1.9169, 1.5503, 1.1679, 1.7971], + device='cuda:3'), covar=tensor([0.3886, 0.4106, 0.1864, 0.2719, 0.3256, 0.3161, 0.4915, 0.2664], + device='cuda:3'), in_proj_covar=tensor([0.0278, 0.0251, 0.0217, 0.0320, 0.0212, 0.0227, 0.0234, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 21:00:12,508 INFO [finetune.py:976] (3/7) Epoch 8, batch 2100, loss[loss=0.254, simple_loss=0.3165, pruned_loss=0.09579, over 4825.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2582, pruned_loss=0.0657, over 954312.66 frames. ], batch size: 40, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:00:17,036 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 21:00:38,854 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.610e+02 1.994e+02 2.459e+02 7.557e+02, threshold=3.988e+02, percent-clipped=2.0 +2023-04-26 21:00:46,720 INFO [finetune.py:976] (3/7) Epoch 8, batch 2150, loss[loss=0.1509, simple_loss=0.2265, pruned_loss=0.03767, over 4737.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2624, pruned_loss=0.06684, over 954293.09 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:00:50,389 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7157, 2.2809, 1.6864, 1.5331, 1.2649, 1.2847, 1.7746, 1.1927], + device='cuda:3'), covar=tensor([0.1846, 0.1442, 0.1718, 0.2062, 0.2598, 0.2208, 0.1239, 0.2232], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0217, 0.0173, 0.0205, 0.0206, 0.0184, 0.0163, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 21:00:58,116 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42262.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:01:15,139 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4166, 0.8946, 0.4374, 1.0764, 1.0085, 1.3049, 1.2095, 1.1828], + device='cuda:3'), covar=tensor([0.0571, 0.0465, 0.0451, 0.0622, 0.0351, 0.0560, 0.0557, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 21:01:18,925 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-04-26 21:01:20,321 INFO [finetune.py:976] (3/7) Epoch 8, batch 2200, loss[loss=0.2283, simple_loss=0.2913, pruned_loss=0.08269, over 4824.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2652, pruned_loss=0.06785, over 954197.02 frames. ], batch size: 33, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:01:26,916 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42304.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 21:01:29,348 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1007, 2.0911, 1.7757, 1.9017, 2.3276, 1.7653, 2.7796, 1.6173], + device='cuda:3'), covar=tensor([0.4318, 0.1945, 0.4890, 0.3563, 0.1850, 0.2839, 0.1631, 0.4436], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0351, 0.0435, 0.0364, 0.0390, 0.0384, 0.0384, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:01:30,520 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42310.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:01:35,296 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 21:01:39,080 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0604, 2.7673, 2.3647, 2.4830, 1.9616, 2.3231, 2.4457, 1.9459], + device='cuda:3'), covar=tensor([0.2222, 0.1284, 0.0887, 0.1288, 0.2889, 0.1173, 0.1973, 0.2480], + device='cuda:3'), in_proj_covar=tensor([0.0301, 0.0322, 0.0233, 0.0295, 0.0317, 0.0276, 0.0263, 0.0286], + device='cuda:3'), out_proj_covar=tensor([1.2195e-04, 1.3017e-04, 9.4140e-05, 1.1795e-04, 1.3050e-04, 1.1130e-04, + 1.0776e-04, 1.1493e-04], device='cuda:3') +2023-04-26 21:01:44,811 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.211e+02 1.704e+02 2.094e+02 2.427e+02 3.760e+02, threshold=4.188e+02, percent-clipped=0.0 +2023-04-26 21:01:48,153 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-26 21:01:52,972 INFO [finetune.py:976] (3/7) Epoch 8, batch 2250, loss[loss=0.2228, simple_loss=0.2814, pruned_loss=0.08208, over 4722.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2665, pruned_loss=0.06872, over 952721.36 frames. ], batch size: 54, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:01:56,648 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1652, 1.5655, 1.9684, 2.4496, 1.9060, 1.5653, 1.1138, 1.7348], + device='cuda:3'), covar=tensor([0.3773, 0.4084, 0.1845, 0.2540, 0.3235, 0.3156, 0.5176, 0.2658], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0254, 0.0220, 0.0324, 0.0215, 0.0230, 0.0237, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 21:02:02,510 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42358.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:02:15,805 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42379.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:02:26,180 INFO [finetune.py:976] (3/7) Epoch 8, batch 2300, loss[loss=0.165, simple_loss=0.2448, pruned_loss=0.04257, over 4786.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2656, pruned_loss=0.06763, over 955099.89 frames. ], batch size: 51, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:02:50,919 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.665e+02 1.908e+02 2.261e+02 6.563e+02, threshold=3.817e+02, percent-clipped=1.0 +2023-04-26 21:02:56,805 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42440.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:02:59,687 INFO [finetune.py:976] (3/7) Epoch 8, batch 2350, loss[loss=0.1637, simple_loss=0.2284, pruned_loss=0.04952, over 4229.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2626, pruned_loss=0.06635, over 955378.09 frames. ], batch size: 65, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:03:32,291 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9481, 1.2426, 5.0424, 4.6402, 4.4206, 4.8265, 4.4833, 4.4462], + device='cuda:3'), covar=tensor([0.7426, 0.6799, 0.1135, 0.2078, 0.1061, 0.1225, 0.1374, 0.1540], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0310, 0.0411, 0.0418, 0.0353, 0.0409, 0.0319, 0.0372], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:03:45,326 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 21:03:48,469 INFO [finetune.py:976] (3/7) Epoch 8, batch 2400, loss[loss=0.1771, simple_loss=0.247, pruned_loss=0.05362, over 4809.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2593, pruned_loss=0.06557, over 956342.86 frames. ], batch size: 25, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:04:08,565 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:04:40,674 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.697e+02 2.030e+02 2.440e+02 3.399e+02, threshold=4.060e+02, percent-clipped=0.0 +2023-04-26 21:04:54,580 INFO [finetune.py:976] (3/7) Epoch 8, batch 2450, loss[loss=0.1658, simple_loss=0.2322, pruned_loss=0.04974, over 4762.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2574, pruned_loss=0.06572, over 956297.92 frames. ], batch size: 27, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:04:55,136 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 21:05:24,358 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42562.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:05:27,955 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42567.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:05:35,538 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1103, 1.8101, 2.1051, 2.4941, 2.4700, 1.9840, 1.6511, 2.0708], + device='cuda:3'), covar=tensor([0.0940, 0.1093, 0.0647, 0.0572, 0.0622, 0.0904, 0.0869, 0.0656], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0205, 0.0183, 0.0178, 0.0179, 0.0191, 0.0161, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:06:00,485 INFO [finetune.py:976] (3/7) Epoch 8, batch 2500, loss[loss=0.2005, simple_loss=0.2673, pruned_loss=0.0669, over 4845.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2601, pruned_loss=0.0671, over 953898.60 frames. ], batch size: 49, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:06:20,333 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42604.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:06:29,255 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42610.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:06:54,510 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.868e+02 2.156e+02 2.482e+02 4.292e+02, threshold=4.312e+02, percent-clipped=2.0 +2023-04-26 21:07:08,190 INFO [finetune.py:976] (3/7) Epoch 8, batch 2550, loss[loss=0.2058, simple_loss=0.2649, pruned_loss=0.07331, over 4924.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2663, pruned_loss=0.06986, over 955526.22 frames. ], batch size: 33, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:07:11,891 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42649.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:07:14,686 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42652.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 21:07:20,200 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5989, 1.1594, 1.6554, 2.1418, 1.7399, 1.5513, 1.6087, 1.6667], + device='cuda:3'), covar=tensor([0.7049, 1.0229, 1.0261, 0.9626, 0.8742, 1.2340, 1.2176, 1.0255], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0427, 0.0509, 0.0531, 0.0440, 0.0459, 0.0470, 0.0469], + device='cuda:3'), out_proj_covar=tensor([9.9471e-05, 1.0593e-04, 1.1504e-04, 1.2611e-04, 1.0694e-04, 1.1100e-04, + 1.1322e-04, 1.1372e-04], device='cuda:3') +2023-04-26 21:07:28,542 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4802, 1.4239, 4.1226, 3.8714, 3.6383, 3.9397, 3.8951, 3.6702], + device='cuda:3'), covar=tensor([0.6579, 0.5241, 0.1090, 0.1614, 0.1050, 0.1579, 0.1391, 0.1368], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0310, 0.0412, 0.0417, 0.0354, 0.0409, 0.0319, 0.0373], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:07:41,704 INFO [finetune.py:976] (3/7) Epoch 8, batch 2600, loss[loss=0.2039, simple_loss=0.2706, pruned_loss=0.06863, over 4905.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.267, pruned_loss=0.06992, over 955622.92 frames. ], batch size: 36, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:07:53,043 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42710.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:08:07,496 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.967e+01 1.665e+02 1.922e+02 2.451e+02 5.774e+02, threshold=3.843e+02, percent-clipped=4.0 +2023-04-26 21:08:09,416 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42735.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:08:14,788 INFO [finetune.py:976] (3/7) Epoch 8, batch 2650, loss[loss=0.1759, simple_loss=0.2468, pruned_loss=0.0525, over 4789.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2679, pruned_loss=0.06953, over 956021.68 frames. ], batch size: 29, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:08:42,363 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4914, 1.0910, 1.2034, 1.0766, 1.6669, 1.3041, 1.0926, 1.1840], + device='cuda:3'), covar=tensor([0.1467, 0.1474, 0.2154, 0.1619, 0.0841, 0.1404, 0.1977, 0.2107], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0325, 0.0354, 0.0302, 0.0342, 0.0326, 0.0311, 0.0355], + device='cuda:3'), out_proj_covar=tensor([6.5317e-05, 6.9120e-05, 7.6706e-05, 6.2534e-05, 7.2011e-05, 7.0144e-05, + 6.7132e-05, 7.6163e-05], device='cuda:3') +2023-04-26 21:08:48,362 INFO [finetune.py:976] (3/7) Epoch 8, batch 2700, loss[loss=0.2019, simple_loss=0.2706, pruned_loss=0.0666, over 4824.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2667, pruned_loss=0.06824, over 956743.49 frames. ], batch size: 39, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:08:52,072 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3452, 1.5884, 1.7289, 1.8795, 1.7470, 1.8862, 1.8177, 1.7968], + device='cuda:3'), covar=tensor([0.5669, 0.6953, 0.5706, 0.5554, 0.6622, 0.9164, 0.6757, 0.6597], + device='cuda:3'), in_proj_covar=tensor([0.0321, 0.0383, 0.0315, 0.0325, 0.0341, 0.0404, 0.0366, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 21:09:12,285 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1262, 2.1790, 1.9186, 1.8785, 2.3845, 1.8332, 2.8098, 1.6222], + device='cuda:3'), covar=tensor([0.4234, 0.1938, 0.4846, 0.3361, 0.1781, 0.2736, 0.1659, 0.4719], + device='cuda:3'), in_proj_covar=tensor([0.0350, 0.0352, 0.0437, 0.0367, 0.0393, 0.0386, 0.0387, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:09:14,557 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.600e+01 1.577e+02 1.842e+02 2.218e+02 2.993e+02, threshold=3.685e+02, percent-clipped=0.0 +2023-04-26 21:09:21,928 INFO [finetune.py:976] (3/7) Epoch 8, batch 2750, loss[loss=0.2213, simple_loss=0.2698, pruned_loss=0.08639, over 4853.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2637, pruned_loss=0.06753, over 957637.88 frames. ], batch size: 44, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:09:27,291 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.0425, 3.7861, 2.8650, 4.6292, 4.0669, 3.9459, 1.7928, 3.9682], + device='cuda:3'), covar=tensor([0.1810, 0.1417, 0.3080, 0.1483, 0.2942, 0.2086, 0.5911, 0.2388], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0217, 0.0249, 0.0309, 0.0299, 0.0250, 0.0272, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 21:09:30,359 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.6456, 4.3567, 3.2109, 5.2769, 4.5988, 4.5763, 2.0833, 4.5733], + device='cuda:3'), covar=tensor([0.1638, 0.0880, 0.3180, 0.0912, 0.2938, 0.1686, 0.5900, 0.1999], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0217, 0.0249, 0.0309, 0.0299, 0.0250, 0.0272, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 21:09:31,572 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:09:34,270 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42862.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:09:54,707 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-26 21:09:57,297 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42879.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:10:16,714 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 21:10:17,379 INFO [finetune.py:976] (3/7) Epoch 8, batch 2800, loss[loss=0.181, simple_loss=0.244, pruned_loss=0.05894, over 4895.00 frames. ], tot_loss[loss=0.1954, simple_loss=0.2595, pruned_loss=0.06565, over 954928.09 frames. ], batch size: 32, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:10:51,340 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 21:10:51,624 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42920.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:11:10,644 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.665e+02 1.974e+02 2.383e+02 4.846e+02, threshold=3.948e+02, percent-clipped=3.0 +2023-04-26 21:11:16,203 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:11:18,534 INFO [finetune.py:976] (3/7) Epoch 8, batch 2850, loss[loss=0.1936, simple_loss=0.2582, pruned_loss=0.0645, over 4808.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2599, pruned_loss=0.06653, over 956266.86 frames. ], batch size: 45, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:11:37,195 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1403, 1.5003, 1.2884, 1.7282, 1.6080, 1.6622, 1.3741, 3.0623], + device='cuda:3'), covar=tensor([0.0687, 0.0774, 0.0843, 0.1258, 0.0642, 0.0515, 0.0747, 0.0181], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 21:11:56,701 INFO [finetune.py:976] (3/7) Epoch 8, batch 2900, loss[loss=0.2094, simple_loss=0.2746, pruned_loss=0.0721, over 4911.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2639, pruned_loss=0.0682, over 956544.19 frames. ], batch size: 37, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:12:15,326 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43005.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:12:51,491 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.213e+02 1.808e+02 2.091e+02 2.545e+02 3.466e+02, threshold=4.182e+02, percent-clipped=0.0 +2023-04-26 21:12:59,360 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43035.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:13:10,713 INFO [finetune.py:976] (3/7) Epoch 8, batch 2950, loss[loss=0.2217, simple_loss=0.2928, pruned_loss=0.07529, over 4835.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2653, pruned_loss=0.06774, over 957139.54 frames. ], batch size: 47, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:13:35,178 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8181, 2.4227, 1.8006, 1.5986, 1.2994, 1.3790, 1.8736, 1.2783], + device='cuda:3'), covar=tensor([0.1805, 0.1459, 0.1689, 0.2121, 0.2662, 0.2141, 0.1155, 0.2230], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0219, 0.0173, 0.0207, 0.0207, 0.0185, 0.0164, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 21:13:37,394 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43083.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:13:44,494 INFO [finetune.py:976] (3/7) Epoch 8, batch 3000, loss[loss=0.24, simple_loss=0.2949, pruned_loss=0.09257, over 4921.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2666, pruned_loss=0.06852, over 956198.50 frames. ], batch size: 33, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:13:44,494 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 21:13:49,533 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4626, 3.0799, 0.9787, 1.7680, 1.9224, 2.3952, 1.9629, 1.0233], + device='cuda:3'), covar=tensor([0.1273, 0.0938, 0.1883, 0.1257, 0.0972, 0.0805, 0.1342, 0.1649], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0255, 0.0144, 0.0125, 0.0137, 0.0157, 0.0122, 0.0123], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:13:50,438 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6258, 1.8581, 1.6951, 1.3326, 1.9003, 1.4474, 2.3145, 1.4536], + device='cuda:3'), covar=tensor([0.4126, 0.1897, 0.5635, 0.3364, 0.1689, 0.2611, 0.1627, 0.5682], + device='cuda:3'), in_proj_covar=tensor([0.0349, 0.0351, 0.0433, 0.0365, 0.0391, 0.0385, 0.0385, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:13:50,559 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8658, 2.2479, 1.9674, 2.0772, 1.8182, 1.8755, 1.9691, 1.6193], + device='cuda:3'), covar=tensor([0.1552, 0.1134, 0.0728, 0.1183, 0.2989, 0.1224, 0.1575, 0.2091], + device='cuda:3'), in_proj_covar=tensor([0.0302, 0.0322, 0.0233, 0.0296, 0.0320, 0.0276, 0.0262, 0.0287], + device='cuda:3'), out_proj_covar=tensor([1.2254e-04, 1.3010e-04, 9.4147e-05, 1.1835e-04, 1.3156e-04, 1.1149e-04, + 1.0732e-04, 1.1528e-04], device='cuda:3') +2023-04-26 21:13:54,962 INFO [finetune.py:1010] (3/7) Epoch 8, validation: loss=0.1551, simple_loss=0.2273, pruned_loss=0.04149, over 2265189.00 frames. +2023-04-26 21:13:54,963 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 21:14:02,821 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 21:14:18,553 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.720e+02 2.087e+02 2.508e+02 4.995e+02, threshold=4.174e+02, percent-clipped=1.0 +2023-04-26 21:14:27,803 INFO [finetune.py:976] (3/7) Epoch 8, batch 3050, loss[loss=0.2162, simple_loss=0.2448, pruned_loss=0.09381, over 4017.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.268, pruned_loss=0.06889, over 956479.12 frames. ], batch size: 17, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:14:39,671 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43162.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:14:43,976 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6286, 1.6289, 0.9808, 1.3289, 1.8026, 1.5011, 1.4175, 1.4488], + device='cuda:3'), covar=tensor([0.0553, 0.0385, 0.0377, 0.0587, 0.0293, 0.0545, 0.0553, 0.0596], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 21:15:00,985 INFO [finetune.py:976] (3/7) Epoch 8, batch 3100, loss[loss=0.1558, simple_loss=0.2334, pruned_loss=0.03907, over 4897.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2647, pruned_loss=0.06802, over 957408.91 frames. ], batch size: 36, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:15:11,989 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43210.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:15:15,104 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43215.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:15:25,435 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.200e+02 1.610e+02 1.868e+02 2.269e+02 4.698e+02, threshold=3.736e+02, percent-clipped=1.0 +2023-04-26 21:15:27,886 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43235.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:15:34,192 INFO [finetune.py:976] (3/7) Epoch 8, batch 3150, loss[loss=0.1981, simple_loss=0.2611, pruned_loss=0.06756, over 4823.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2607, pruned_loss=0.0664, over 954178.35 frames. ], batch size: 38, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:16:18,110 INFO [finetune.py:976] (3/7) Epoch 8, batch 3200, loss[loss=0.2064, simple_loss=0.2675, pruned_loss=0.07263, over 4739.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2572, pruned_loss=0.06457, over 956218.67 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:16:30,941 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43305.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:16:51,016 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3602, 3.2583, 1.1343, 1.7975, 1.7768, 2.4565, 1.8978, 0.9979], + device='cuda:3'), covar=tensor([0.1436, 0.0962, 0.1852, 0.1342, 0.1152, 0.0894, 0.1441, 0.1921], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0251, 0.0141, 0.0123, 0.0135, 0.0154, 0.0120, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:17:03,561 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.615e+02 2.007e+02 2.471e+02 4.400e+02, threshold=4.014e+02, percent-clipped=2.0 +2023-04-26 21:17:20,614 INFO [finetune.py:976] (3/7) Epoch 8, batch 3250, loss[loss=0.2506, simple_loss=0.3087, pruned_loss=0.09624, over 4743.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2582, pruned_loss=0.06521, over 956247.87 frames. ], batch size: 54, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:17:24,324 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0126, 1.8004, 2.0767, 2.4539, 2.4773, 1.8872, 1.5140, 1.9950], + device='cuda:3'), covar=tensor([0.0973, 0.1046, 0.0669, 0.0568, 0.0538, 0.0982, 0.0917, 0.0634], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0205, 0.0180, 0.0178, 0.0178, 0.0191, 0.0160, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:17:33,418 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43353.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:17:58,311 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-04-26 21:18:26,602 INFO [finetune.py:976] (3/7) Epoch 8, batch 3300, loss[loss=0.1837, simple_loss=0.2249, pruned_loss=0.07127, over 4421.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.2617, pruned_loss=0.06656, over 953157.79 frames. ], batch size: 19, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:18:41,228 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4315, 1.6315, 1.6988, 2.2294, 2.3512, 1.9684, 1.8313, 1.7556], + device='cuda:3'), covar=tensor([0.1422, 0.1961, 0.2084, 0.1794, 0.1426, 0.2176, 0.2617, 0.2353], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0329, 0.0357, 0.0306, 0.0346, 0.0329, 0.0313, 0.0359], + device='cuda:3'), out_proj_covar=tensor([6.6092e-05, 6.9861e-05, 7.7274e-05, 6.3532e-05, 7.2749e-05, 7.0846e-05, + 6.7615e-05, 7.7101e-05], device='cuda:3') +2023-04-26 21:18:45,366 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2468, 3.0689, 2.4642, 2.6777, 2.0572, 2.5471, 2.6847, 2.0321], + device='cuda:3'), covar=tensor([0.2771, 0.1480, 0.1052, 0.1525, 0.3695, 0.1660, 0.2005, 0.3246], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0323, 0.0233, 0.0295, 0.0321, 0.0277, 0.0262, 0.0288], + device='cuda:3'), out_proj_covar=tensor([1.2304e-04, 1.3040e-04, 9.4112e-05, 1.1816e-04, 1.3183e-04, 1.1181e-04, + 1.0756e-04, 1.1561e-04], device='cuda:3') +2023-04-26 21:18:59,197 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.201e+02 1.779e+02 2.083e+02 2.593e+02 4.743e+02, threshold=4.166e+02, percent-clipped=4.0 +2023-04-26 21:19:06,531 INFO [finetune.py:976] (3/7) Epoch 8, batch 3350, loss[loss=0.1809, simple_loss=0.2579, pruned_loss=0.05194, over 4815.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2641, pruned_loss=0.06756, over 951278.91 frames. ], batch size: 38, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:19:26,521 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7263, 1.6169, 3.8723, 3.6149, 3.3977, 3.5509, 3.4991, 3.4713], + device='cuda:3'), covar=tensor([0.6519, 0.5013, 0.1100, 0.1708, 0.1166, 0.1644, 0.3611, 0.1331], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0308, 0.0410, 0.0414, 0.0352, 0.0406, 0.0318, 0.0372], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:19:40,384 INFO [finetune.py:976] (3/7) Epoch 8, batch 3400, loss[loss=0.1804, simple_loss=0.2508, pruned_loss=0.05503, over 4703.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2661, pruned_loss=0.06811, over 952186.73 frames. ], batch size: 59, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:19:54,921 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43514.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:19:55,542 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43515.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:00,421 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-26 21:20:06,651 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.715e+02 2.177e+02 2.420e+02 5.380e+02, threshold=4.353e+02, percent-clipped=2.0 +2023-04-26 21:20:07,482 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-04-26 21:20:08,579 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43535.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:13,909 INFO [finetune.py:976] (3/7) Epoch 8, batch 3450, loss[loss=0.1842, simple_loss=0.253, pruned_loss=0.05773, over 4828.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2668, pruned_loss=0.06845, over 953048.79 frames. ], batch size: 47, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:20:26,913 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43563.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:35,644 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43575.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:40,951 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43583.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:47,607 INFO [finetune.py:976] (3/7) Epoch 8, batch 3500, loss[loss=0.1996, simple_loss=0.2461, pruned_loss=0.07651, over 4732.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2647, pruned_loss=0.06817, over 953287.86 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:21:09,584 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6369, 1.6600, 0.9100, 1.3187, 1.8425, 1.5206, 1.4143, 1.4301], + device='cuda:3'), covar=tensor([0.0527, 0.0378, 0.0379, 0.0555, 0.0284, 0.0541, 0.0541, 0.0620], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 21:21:11,478 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3783, 1.4869, 1.6329, 1.8288, 1.6954, 1.7947, 1.7672, 1.7480], + device='cuda:3'), covar=tensor([0.4852, 0.7257, 0.6164, 0.5334, 0.7141, 1.0114, 0.7039, 0.6858], + device='cuda:3'), in_proj_covar=tensor([0.0324, 0.0388, 0.0318, 0.0329, 0.0345, 0.0410, 0.0370, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 21:21:13,732 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.646e+02 1.959e+02 2.415e+02 4.063e+02, threshold=3.918e+02, percent-clipped=0.0 +2023-04-26 21:21:21,526 INFO [finetune.py:976] (3/7) Epoch 8, batch 3550, loss[loss=0.142, simple_loss=0.2096, pruned_loss=0.03715, over 4875.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.261, pruned_loss=0.06644, over 955637.21 frames. ], batch size: 31, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:22:22,681 INFO [finetune.py:976] (3/7) Epoch 8, batch 3600, loss[loss=0.174, simple_loss=0.2438, pruned_loss=0.05215, over 4765.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2587, pruned_loss=0.06584, over 956197.67 frames. ], batch size: 28, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:23:15,351 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.186e+02 1.664e+02 2.003e+02 2.703e+02 5.988e+02, threshold=4.006e+02, percent-clipped=3.0 +2023-04-26 21:23:16,894 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.81 vs. limit=2.0 +2023-04-26 21:23:34,070 INFO [finetune.py:976] (3/7) Epoch 8, batch 3650, loss[loss=0.2629, simple_loss=0.3271, pruned_loss=0.09936, over 4833.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2612, pruned_loss=0.06661, over 955828.41 frames. ], batch size: 47, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:23:45,339 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2729, 1.6624, 2.0579, 2.6692, 2.0784, 1.6311, 1.2691, 2.0345], + device='cuda:3'), covar=tensor([0.3766, 0.4443, 0.2114, 0.2674, 0.3556, 0.3436, 0.4887, 0.2712], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0253, 0.0220, 0.0323, 0.0215, 0.0230, 0.0237, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 21:24:19,027 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8595, 2.3470, 2.0438, 2.1656, 1.5909, 1.9658, 2.1303, 1.6802], + device='cuda:3'), covar=tensor([0.2300, 0.1362, 0.0863, 0.1505, 0.3495, 0.1499, 0.1935, 0.2616], + device='cuda:3'), in_proj_covar=tensor([0.0302, 0.0322, 0.0232, 0.0294, 0.0319, 0.0276, 0.0262, 0.0286], + device='cuda:3'), out_proj_covar=tensor([1.2240e-04, 1.2987e-04, 9.3700e-05, 1.1783e-04, 1.3134e-04, 1.1149e-04, + 1.0723e-04, 1.1496e-04], device='cuda:3') +2023-04-26 21:24:29,569 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43787.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:24:33,740 INFO [finetune.py:976] (3/7) Epoch 8, batch 3700, loss[loss=0.2059, simple_loss=0.2762, pruned_loss=0.06782, over 4932.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2643, pruned_loss=0.06771, over 955488.25 frames. ], batch size: 42, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:25:03,746 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.187e+02 1.778e+02 2.044e+02 2.450e+02 4.108e+02, threshold=4.087e+02, percent-clipped=2.0 +2023-04-26 21:25:11,091 INFO [finetune.py:976] (3/7) Epoch 8, batch 3750, loss[loss=0.1888, simple_loss=0.2657, pruned_loss=0.056, over 4835.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2649, pruned_loss=0.06759, over 953588.89 frames. ], batch size: 47, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:25:14,040 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:25:27,346 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43870.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:25:44,752 INFO [finetune.py:976] (3/7) Epoch 8, batch 3800, loss[loss=0.165, simple_loss=0.2316, pruned_loss=0.04923, over 4801.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2648, pruned_loss=0.06728, over 953417.56 frames. ], batch size: 25, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:25:54,505 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43909.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:26:03,080 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7982, 1.6148, 1.9807, 2.0949, 1.9101, 1.7012, 1.8239, 1.8827], + device='cuda:3'), covar=tensor([0.8436, 1.0606, 1.1501, 1.1656, 0.9336, 1.3935, 1.4300, 1.3301], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0424, 0.0509, 0.0530, 0.0439, 0.0457, 0.0470, 0.0467], + device='cuda:3'), out_proj_covar=tensor([9.9590e-05, 1.0516e-04, 1.1500e-04, 1.2576e-04, 1.0666e-04, 1.1038e-04, + 1.1305e-04, 1.1335e-04], device='cuda:3') +2023-04-26 21:26:09,807 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.636e+02 2.019e+02 2.329e+02 3.676e+02, threshold=4.038e+02, percent-clipped=0.0 +2023-04-26 21:26:18,576 INFO [finetune.py:976] (3/7) Epoch 8, batch 3850, loss[loss=0.2279, simple_loss=0.2898, pruned_loss=0.08305, over 4822.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2631, pruned_loss=0.06661, over 953648.63 frames. ], batch size: 30, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:26:28,313 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43959.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:26:35,138 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:26:50,776 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3692, 1.7693, 1.5536, 2.1897, 1.8528, 2.1240, 1.5210, 4.4157], + device='cuda:3'), covar=tensor([0.0610, 0.0788, 0.0828, 0.1179, 0.0634, 0.0568, 0.0794, 0.0098], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 21:27:02,336 INFO [finetune.py:976] (3/7) Epoch 8, batch 3900, loss[loss=0.198, simple_loss=0.2523, pruned_loss=0.0719, over 4906.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2597, pruned_loss=0.06541, over 955802.28 frames. ], batch size: 35, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:27:34,282 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8829, 1.1824, 3.2781, 3.0081, 2.9043, 3.2243, 3.1870, 2.8327], + device='cuda:3'), covar=tensor([0.7825, 0.5790, 0.1509, 0.2214, 0.1499, 0.1748, 0.1601, 0.1908], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0304, 0.0407, 0.0409, 0.0348, 0.0402, 0.0315, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:27:35,505 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44020.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:27:48,403 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.747e+02 1.960e+02 2.337e+02 4.630e+02, threshold=3.919e+02, percent-clipped=3.0 +2023-04-26 21:28:08,674 INFO [finetune.py:976] (3/7) Epoch 8, batch 3950, loss[loss=0.1882, simple_loss=0.256, pruned_loss=0.06023, over 4865.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2567, pruned_loss=0.06452, over 955543.25 frames. ], batch size: 31, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:28:58,844 INFO [finetune.py:976] (3/7) Epoch 8, batch 4000, loss[loss=0.1796, simple_loss=0.2476, pruned_loss=0.05581, over 4767.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2561, pruned_loss=0.06431, over 954992.44 frames. ], batch size: 27, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:29:27,322 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-04-26 21:29:49,241 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.897e+01 1.638e+02 2.029e+02 2.445e+02 5.584e+02, threshold=4.057e+02, percent-clipped=1.0 +2023-04-26 21:30:02,120 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44143.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:30:02,664 INFO [finetune.py:976] (3/7) Epoch 8, batch 4050, loss[loss=0.2094, simple_loss=0.2756, pruned_loss=0.07157, over 4718.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2602, pruned_loss=0.066, over 953814.60 frames. ], batch size: 59, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:30:11,778 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1943, 1.1883, 3.7968, 3.4908, 3.3495, 3.6550, 3.6627, 3.2500], + device='cuda:3'), covar=tensor([0.7774, 0.6026, 0.1319, 0.2000, 0.1363, 0.2218, 0.1420, 0.1740], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0307, 0.0410, 0.0412, 0.0350, 0.0405, 0.0317, 0.0371], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:30:22,497 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-26 21:30:35,440 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44170.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:31:07,887 INFO [finetune.py:976] (3/7) Epoch 8, batch 4100, loss[loss=0.2461, simple_loss=0.3049, pruned_loss=0.09364, over 4903.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2638, pruned_loss=0.06751, over 952592.97 frames. ], batch size: 36, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:31:41,627 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44218.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:32:00,831 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.648e+02 2.017e+02 2.380e+02 4.436e+02, threshold=4.034e+02, percent-clipped=1.0 +2023-04-26 21:32:09,631 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0247, 1.2977, 1.1421, 1.5619, 1.3983, 1.3583, 1.2674, 2.4384], + device='cuda:3'), covar=tensor([0.0643, 0.0824, 0.0823, 0.1219, 0.0658, 0.0534, 0.0768, 0.0226], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 21:32:13,830 INFO [finetune.py:976] (3/7) Epoch 8, batch 4150, loss[loss=0.2191, simple_loss=0.2919, pruned_loss=0.07315, over 4875.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2649, pruned_loss=0.06714, over 953183.97 frames. ], batch size: 32, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:32:34,618 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:32:52,670 INFO [finetune.py:976] (3/7) Epoch 8, batch 4200, loss[loss=0.1963, simple_loss=0.2632, pruned_loss=0.06464, over 4916.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2647, pruned_loss=0.0668, over 951508.29 frames. ], batch size: 38, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:33:08,261 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44315.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:33:18,929 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.191e+02 1.745e+02 1.959e+02 2.417e+02 4.699e+02, threshold=3.917e+02, percent-clipped=3.0 +2023-04-26 21:33:26,339 INFO [finetune.py:976] (3/7) Epoch 8, batch 4250, loss[loss=0.1999, simple_loss=0.2537, pruned_loss=0.07298, over 4908.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2631, pruned_loss=0.06635, over 954262.28 frames. ], batch size: 46, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:33:57,243 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8620, 2.0815, 1.8984, 2.1185, 1.9127, 2.1799, 1.9674, 1.9071], + device='cuda:3'), covar=tensor([0.5752, 0.8523, 0.7652, 0.6499, 0.7811, 1.0256, 0.9668, 0.8298], + device='cuda:3'), in_proj_covar=tensor([0.0321, 0.0385, 0.0315, 0.0325, 0.0341, 0.0406, 0.0366, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 21:34:00,167 INFO [finetune.py:976] (3/7) Epoch 8, batch 4300, loss[loss=0.1529, simple_loss=0.2243, pruned_loss=0.04078, over 4743.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2593, pruned_loss=0.0651, over 953796.12 frames. ], batch size: 27, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:34:13,751 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7606, 1.8430, 1.2257, 1.5912, 1.8864, 1.7098, 1.6357, 1.6482], + device='cuda:3'), covar=tensor([0.0416, 0.0298, 0.0335, 0.0451, 0.0273, 0.0416, 0.0392, 0.0469], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:3') +2023-04-26 21:34:26,193 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.667e+02 1.928e+02 2.256e+02 4.744e+02, threshold=3.855e+02, percent-clipped=2.0 +2023-04-26 21:34:44,419 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44443.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:34:44,910 INFO [finetune.py:976] (3/7) Epoch 8, batch 4350, loss[loss=0.2703, simple_loss=0.3095, pruned_loss=0.1155, over 4841.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.257, pruned_loss=0.06479, over 950865.37 frames. ], batch size: 51, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:34:55,242 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-04-26 21:35:10,441 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9528, 2.2289, 2.1187, 2.2214, 2.0408, 2.3403, 2.2392, 2.1649], + device='cuda:3'), covar=tensor([0.5055, 0.8906, 0.7435, 0.6743, 0.7850, 1.0539, 0.8938, 0.7932], + device='cuda:3'), in_proj_covar=tensor([0.0322, 0.0385, 0.0315, 0.0325, 0.0341, 0.0405, 0.0366, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 21:35:19,940 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6084, 1.2388, 1.7567, 2.0925, 1.7792, 1.5720, 1.6455, 1.6754], + device='cuda:3'), covar=tensor([0.6261, 0.8491, 0.8104, 0.8172, 0.7457, 0.9913, 0.9785, 0.8980], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0424, 0.0507, 0.0529, 0.0439, 0.0459, 0.0470, 0.0466], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:35:46,506 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44491.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:35:48,247 INFO [finetune.py:976] (3/7) Epoch 8, batch 4400, loss[loss=0.2033, simple_loss=0.2683, pruned_loss=0.06911, over 4864.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2584, pruned_loss=0.06569, over 952983.38 frames. ], batch size: 34, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:36:05,285 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6109, 1.0949, 1.1637, 1.3767, 1.8256, 1.4142, 1.2077, 1.0568], + device='cuda:3'), covar=tensor([0.1486, 0.1818, 0.1828, 0.1494, 0.0961, 0.1594, 0.2009, 0.2335], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0322, 0.0350, 0.0299, 0.0336, 0.0321, 0.0307, 0.0353], + device='cuda:3'), out_proj_covar=tensor([6.4592e-05, 6.8400e-05, 7.5563e-05, 6.1960e-05, 7.0465e-05, 6.9011e-05, + 6.6178e-05, 7.5805e-05], device='cuda:3') +2023-04-26 21:36:07,090 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 21:36:11,272 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44527.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:36:13,050 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 21:36:14,647 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.807e+02 2.142e+02 2.505e+02 5.650e+02, threshold=4.284e+02, percent-clipped=2.0 +2023-04-26 21:36:33,286 INFO [finetune.py:976] (3/7) Epoch 8, batch 4450, loss[loss=0.2434, simple_loss=0.3093, pruned_loss=0.08874, over 4809.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2633, pruned_loss=0.06752, over 953985.60 frames. ], batch size: 45, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:36:42,779 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44550.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:36:46,731 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-04-26 21:36:58,011 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:37:30,501 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44588.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:37:39,768 INFO [finetune.py:976] (3/7) Epoch 8, batch 4500, loss[loss=0.2439, simple_loss=0.3044, pruned_loss=0.09167, over 4802.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2647, pruned_loss=0.06761, over 954492.85 frames. ], batch size: 41, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:37:42,354 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4421, 0.7261, 1.3480, 1.8684, 1.5861, 1.3739, 1.3814, 1.4782], + device='cuda:3'), covar=tensor([0.6101, 0.7804, 0.7326, 0.8630, 0.6894, 0.9314, 0.8886, 0.8433], + device='cuda:3'), in_proj_covar=tensor([0.0410, 0.0424, 0.0505, 0.0528, 0.0438, 0.0457, 0.0469, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.9884e-05, 1.0515e-04, 1.1417e-04, 1.2534e-04, 1.0644e-04, 1.1040e-04, + 1.1295e-04, 1.1262e-04], device='cuda:3') +2023-04-26 21:37:57,810 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44611.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:37:58,975 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44613.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:38:00,231 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44615.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:38:30,723 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.764e+02 2.099e+02 2.413e+02 4.489e+02, threshold=4.197e+02, percent-clipped=1.0 +2023-04-26 21:38:30,971 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 21:38:43,668 INFO [finetune.py:976] (3/7) Epoch 8, batch 4550, loss[loss=0.2042, simple_loss=0.2674, pruned_loss=0.07051, over 4679.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2657, pruned_loss=0.06786, over 954405.05 frames. ], batch size: 23, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:38:46,854 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5619, 3.0212, 0.9847, 1.4948, 2.1610, 1.5260, 4.2391, 1.8227], + device='cuda:3'), covar=tensor([0.0621, 0.0800, 0.0950, 0.1395, 0.0550, 0.1022, 0.0399, 0.0709], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-26 21:38:55,368 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44663.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:39:03,125 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1761, 1.4695, 1.9660, 2.5786, 2.0206, 1.5527, 1.3804, 1.8789], + device='cuda:3'), covar=tensor([0.4014, 0.4482, 0.2056, 0.3577, 0.3688, 0.3138, 0.5331, 0.2967], + device='cuda:3'), in_proj_covar=tensor([0.0279, 0.0249, 0.0218, 0.0319, 0.0212, 0.0227, 0.0234, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 21:39:17,128 INFO [finetune.py:976] (3/7) Epoch 8, batch 4600, loss[loss=0.18, simple_loss=0.254, pruned_loss=0.05298, over 4908.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2634, pruned_loss=0.06667, over 952918.69 frames. ], batch size: 37, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:39:17,214 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4880, 1.3771, 4.4645, 4.1516, 3.9249, 4.2624, 4.0691, 3.9240], + device='cuda:3'), covar=tensor([0.7600, 0.5901, 0.1027, 0.1591, 0.1071, 0.1479, 0.1312, 0.1362], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0307, 0.0408, 0.0412, 0.0348, 0.0404, 0.0315, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:39:17,238 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2283, 1.5829, 1.3810, 1.8347, 1.6555, 1.9381, 1.3767, 3.6992], + device='cuda:3'), covar=tensor([0.0650, 0.0796, 0.0824, 0.1209, 0.0657, 0.0565, 0.0769, 0.0145], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 21:39:43,235 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.725e+02 1.987e+02 2.350e+02 3.510e+02, threshold=3.973e+02, percent-clipped=0.0 +2023-04-26 21:39:55,905 INFO [finetune.py:976] (3/7) Epoch 8, batch 4650, loss[loss=0.198, simple_loss=0.2602, pruned_loss=0.06796, over 4817.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2603, pruned_loss=0.06562, over 954462.03 frames. ], batch size: 30, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:40:51,264 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7955, 1.3201, 1.4131, 1.4426, 1.9612, 1.5488, 1.3365, 1.3189], + device='cuda:3'), covar=tensor([0.1313, 0.1377, 0.1566, 0.1320, 0.0669, 0.1437, 0.1892, 0.1677], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0324, 0.0355, 0.0303, 0.0340, 0.0326, 0.0309, 0.0358], + device='cuda:3'), out_proj_covar=tensor([6.5534e-05, 6.8778e-05, 7.6766e-05, 6.2824e-05, 7.1531e-05, 6.9994e-05, + 6.6633e-05, 7.6927e-05], device='cuda:3') +2023-04-26 21:41:01,770 INFO [finetune.py:976] (3/7) Epoch 8, batch 4700, loss[loss=0.1716, simple_loss=0.2311, pruned_loss=0.05609, over 4873.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2572, pruned_loss=0.06451, over 956512.15 frames. ], batch size: 34, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:41:02,386 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.3748, 1.4140, 1.3886, 1.0387, 1.4855, 1.2243, 1.7960, 1.3124], + device='cuda:3'), covar=tensor([0.4211, 0.1881, 0.5273, 0.3114, 0.1578, 0.2242, 0.1702, 0.5427], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0349, 0.0430, 0.0364, 0.0386, 0.0382, 0.0383, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:41:19,052 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-04-26 21:41:26,394 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.673e+02 1.965e+02 2.397e+02 5.509e+02, threshold=3.929e+02, percent-clipped=2.0 +2023-04-26 21:41:28,748 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6361, 1.3197, 1.7134, 2.0793, 1.7932, 1.5865, 1.6780, 1.6742], + device='cuda:3'), covar=tensor([0.7159, 0.9603, 0.9501, 0.9192, 0.8143, 1.1252, 1.1590, 1.0915], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0426, 0.0508, 0.0530, 0.0440, 0.0459, 0.0471, 0.0467], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:41:35,058 INFO [finetune.py:976] (3/7) Epoch 8, batch 4750, loss[loss=0.1946, simple_loss=0.266, pruned_loss=0.06159, over 4836.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2546, pruned_loss=0.06342, over 956722.83 frames. ], batch size: 33, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:41:38,717 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 21:41:42,314 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4800, 1.7021, 1.3883, 1.1435, 1.1322, 1.1338, 1.3528, 1.0837], + device='cuda:3'), covar=tensor([0.1876, 0.1391, 0.1659, 0.1933, 0.2613, 0.2069, 0.1231, 0.2187], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0217, 0.0172, 0.0206, 0.0206, 0.0185, 0.0163, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 21:41:59,217 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44883.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:42:08,192 INFO [finetune.py:976] (3/7) Epoch 8, batch 4800, loss[loss=0.2028, simple_loss=0.2761, pruned_loss=0.06468, over 4822.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2594, pruned_loss=0.06602, over 956763.98 frames. ], batch size: 40, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:42:13,935 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 21:42:16,079 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44906.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:42:32,274 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.212e+02 1.751e+02 2.130e+02 2.730e+02 4.658e+02, threshold=4.261e+02, percent-clipped=2.0 +2023-04-26 21:42:40,904 INFO [finetune.py:976] (3/7) Epoch 8, batch 4850, loss[loss=0.2069, simple_loss=0.2723, pruned_loss=0.07073, over 4816.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2628, pruned_loss=0.06658, over 957646.46 frames. ], batch size: 30, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:42:58,207 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6756, 1.2475, 1.7508, 2.1762, 1.8067, 1.6940, 1.7887, 1.7728], + device='cuda:3'), covar=tensor([0.6802, 0.8917, 0.9041, 0.8790, 0.8147, 1.0603, 1.0043, 0.9047], + device='cuda:3'), in_proj_covar=tensor([0.0410, 0.0424, 0.0506, 0.0528, 0.0438, 0.0458, 0.0468, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.9737e-05, 1.0523e-04, 1.1441e-04, 1.2520e-04, 1.0635e-04, 1.1065e-04, + 1.1270e-04, 1.1273e-04], device='cuda:3') +2023-04-26 21:43:28,014 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.5983, 1.6526, 1.7506, 1.2306, 1.7979, 1.4173, 2.3175, 1.4847], + device='cuda:3'), covar=tensor([0.3721, 0.1605, 0.4189, 0.2997, 0.1549, 0.2350, 0.1409, 0.4472], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0349, 0.0430, 0.0363, 0.0386, 0.0382, 0.0382, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:43:28,021 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4402, 1.2139, 1.5764, 1.5902, 1.3145, 1.1707, 1.3447, 0.7989], + device='cuda:3'), covar=tensor([0.0564, 0.1162, 0.0647, 0.0754, 0.0898, 0.1296, 0.0720, 0.0931], + device='cuda:3'), in_proj_covar=tensor([0.0065, 0.0073, 0.0072, 0.0067, 0.0076, 0.0095, 0.0079, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:43:40,816 INFO [finetune.py:976] (3/7) Epoch 8, batch 4900, loss[loss=0.2036, simple_loss=0.2796, pruned_loss=0.06378, over 4855.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.265, pruned_loss=0.06714, over 957155.01 frames. ], batch size: 44, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:44:04,822 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0179, 0.9574, 1.1946, 1.1056, 0.9918, 0.9149, 0.9452, 0.5304], + device='cuda:3'), covar=tensor([0.0653, 0.0760, 0.0639, 0.0751, 0.0871, 0.1341, 0.0654, 0.0990], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0077, 0.0096, 0.0079, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:44:28,706 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6100, 2.0161, 1.5143, 1.3228, 1.1880, 1.2263, 1.5836, 1.1667], + device='cuda:3'), covar=tensor([0.1817, 0.1399, 0.1706, 0.2021, 0.2467, 0.2111, 0.1135, 0.2207], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0218, 0.0173, 0.0207, 0.0207, 0.0186, 0.0163, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 21:44:34,742 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.241e+02 1.669e+02 1.942e+02 2.230e+02 4.872e+02, threshold=3.883e+02, percent-clipped=3.0 +2023-04-26 21:44:46,772 INFO [finetune.py:976] (3/7) Epoch 8, batch 4950, loss[loss=0.1863, simple_loss=0.256, pruned_loss=0.05832, over 4851.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2658, pruned_loss=0.06717, over 956942.31 frames. ], batch size: 44, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:44:47,373 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.4192, 1.4565, 1.3663, 1.0971, 1.4813, 1.1777, 1.8000, 1.2540], + device='cuda:3'), covar=tensor([0.3409, 0.1530, 0.4841, 0.2406, 0.1572, 0.1927, 0.1561, 0.4627], + device='cuda:3'), in_proj_covar=tensor([0.0348, 0.0353, 0.0433, 0.0366, 0.0390, 0.0385, 0.0384, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:45:08,584 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45056.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:45:08,876 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 21:45:39,787 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45081.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:45:47,620 INFO [finetune.py:976] (3/7) Epoch 8, batch 5000, loss[loss=0.1973, simple_loss=0.2501, pruned_loss=0.07221, over 4804.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2638, pruned_loss=0.06654, over 958505.14 frames. ], batch size: 40, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:46:04,172 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45117.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:46:13,682 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.645e+02 2.005e+02 2.467e+02 4.350e+02, threshold=4.011e+02, percent-clipped=3.0 +2023-04-26 21:46:19,274 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45142.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:46:20,393 INFO [finetune.py:976] (3/7) Epoch 8, batch 5050, loss[loss=0.1539, simple_loss=0.2204, pruned_loss=0.04369, over 4905.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2609, pruned_loss=0.06554, over 957214.45 frames. ], batch size: 32, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:46:46,396 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1454, 1.3697, 1.2401, 1.6289, 1.4183, 1.4785, 1.2573, 2.4188], + device='cuda:3'), covar=tensor([0.0638, 0.0869, 0.0847, 0.1238, 0.0697, 0.0537, 0.0791, 0.0231], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 21:46:47,001 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45183.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:46:53,636 INFO [finetune.py:976] (3/7) Epoch 8, batch 5100, loss[loss=0.1899, simple_loss=0.2527, pruned_loss=0.06355, over 4132.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2579, pruned_loss=0.06462, over 957514.86 frames. ], batch size: 18, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:46:53,727 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9877, 2.2312, 0.7468, 1.3036, 1.3820, 1.5858, 1.4549, 0.7334], + device='cuda:3'), covar=tensor([0.1471, 0.1474, 0.1773, 0.1393, 0.1108, 0.1055, 0.1541, 0.1608], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0252, 0.0142, 0.0124, 0.0136, 0.0155, 0.0120, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:47:02,072 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45206.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:47:02,732 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8147, 2.3786, 1.9628, 2.0761, 1.5759, 1.8884, 2.0415, 1.6424], + device='cuda:3'), covar=tensor([0.1948, 0.1032, 0.0897, 0.1249, 0.3284, 0.1265, 0.1714, 0.2334], + device='cuda:3'), in_proj_covar=tensor([0.0300, 0.0320, 0.0230, 0.0293, 0.0317, 0.0274, 0.0260, 0.0281], + device='cuda:3'), out_proj_covar=tensor([1.2139e-04, 1.2924e-04, 9.2821e-05, 1.1727e-04, 1.3029e-04, 1.1049e-04, + 1.0638e-04, 1.1284e-04], device='cuda:3') +2023-04-26 21:47:18,724 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45231.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:47:19,884 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.617e+02 1.880e+02 2.299e+02 5.532e+02, threshold=3.760e+02, percent-clipped=2.0 +2023-04-26 21:47:20,193 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-26 21:47:25,037 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-26 21:47:26,587 INFO [finetune.py:976] (3/7) Epoch 8, batch 5150, loss[loss=0.2133, simple_loss=0.2766, pruned_loss=0.075, over 4842.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2572, pruned_loss=0.06463, over 955706.97 frames. ], batch size: 49, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:47:32,726 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45254.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:47:35,217 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-26 21:47:49,845 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-26 21:48:00,050 INFO [finetune.py:976] (3/7) Epoch 8, batch 5200, loss[loss=0.1653, simple_loss=0.2386, pruned_loss=0.04595, over 4865.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2603, pruned_loss=0.06559, over 953218.91 frames. ], batch size: 31, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:48:32,830 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4163, 2.3352, 2.6672, 2.9175, 2.8886, 2.1986, 1.9839, 2.4140], + device='cuda:3'), covar=tensor([0.0957, 0.1022, 0.0639, 0.0681, 0.0564, 0.1013, 0.0867, 0.0675], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0206, 0.0180, 0.0177, 0.0177, 0.0191, 0.0161, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 21:48:37,714 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8045, 2.1892, 1.9189, 2.0341, 1.5456, 1.8327, 1.8808, 1.5405], + device='cuda:3'), covar=tensor([0.1891, 0.1314, 0.0901, 0.1430, 0.3345, 0.1238, 0.1811, 0.2476], + device='cuda:3'), in_proj_covar=tensor([0.0301, 0.0323, 0.0232, 0.0295, 0.0320, 0.0277, 0.0261, 0.0284], + device='cuda:3'), out_proj_covar=tensor([1.2212e-04, 1.3016e-04, 9.3576e-05, 1.1812e-04, 1.3163e-04, 1.1174e-04, + 1.0704e-04, 1.1378e-04], device='cuda:3') +2023-04-26 21:48:44,628 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.724e+02 2.073e+02 2.442e+02 4.626e+02, threshold=4.147e+02, percent-clipped=2.0 +2023-04-26 21:48:56,370 INFO [finetune.py:976] (3/7) Epoch 8, batch 5250, loss[loss=0.2117, simple_loss=0.2724, pruned_loss=0.0755, over 4903.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.262, pruned_loss=0.06556, over 954131.06 frames. ], batch size: 37, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:48:58,948 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45348.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 21:49:39,192 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3640, 3.0159, 2.5090, 2.4943, 1.6517, 1.8658, 2.7672, 1.7402], + device='cuda:3'), covar=tensor([0.1682, 0.1498, 0.1388, 0.1688, 0.2379, 0.1883, 0.0937, 0.2087], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0218, 0.0173, 0.0207, 0.0207, 0.0186, 0.0163, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 21:50:01,144 INFO [finetune.py:976] (3/7) Epoch 8, batch 5300, loss[loss=0.2198, simple_loss=0.2852, pruned_loss=0.07719, over 4742.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2628, pruned_loss=0.06565, over 953247.93 frames. ], batch size: 54, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:50:21,701 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45409.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 21:50:23,489 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45412.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 21:50:35,534 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5048, 1.3902, 0.5427, 1.2169, 1.4193, 1.3548, 1.2522, 1.3281], + device='cuda:3'), covar=tensor([0.0581, 0.0441, 0.0475, 0.0641, 0.0327, 0.0615, 0.0583, 0.0671], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 21:50:55,376 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.314e+02 1.711e+02 1.928e+02 2.396e+02 5.196e+02, threshold=3.857e+02, percent-clipped=2.0 +2023-04-26 21:50:57,915 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45437.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:51:07,772 INFO [finetune.py:976] (3/7) Epoch 8, batch 5350, loss[loss=0.1873, simple_loss=0.2534, pruned_loss=0.06057, over 4727.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2625, pruned_loss=0.06543, over 954123.72 frames. ], batch size: 54, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:51:45,336 INFO [finetune.py:976] (3/7) Epoch 8, batch 5400, loss[loss=0.1795, simple_loss=0.2393, pruned_loss=0.05988, over 4779.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2605, pruned_loss=0.06498, over 955491.64 frames. ], batch size: 26, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:52:08,924 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4406, 1.2846, 1.6365, 1.5637, 1.3216, 1.2021, 1.3875, 0.7878], + device='cuda:3'), covar=tensor([0.0578, 0.0687, 0.0478, 0.0680, 0.0696, 0.1316, 0.0549, 0.0845], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0076, 0.0096, 0.0078, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:52:11,221 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.671e+02 1.868e+02 2.359e+02 4.891e+02, threshold=3.736e+02, percent-clipped=4.0 +2023-04-26 21:52:18,384 INFO [finetune.py:976] (3/7) Epoch 8, batch 5450, loss[loss=0.1872, simple_loss=0.2519, pruned_loss=0.06126, over 4870.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2579, pruned_loss=0.06405, over 956619.72 frames. ], batch size: 31, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:52:33,877 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.02 vs. limit=5.0 +2023-04-26 21:52:46,916 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6064, 3.4729, 0.8938, 1.8875, 2.0263, 2.3791, 1.9554, 1.0484], + device='cuda:3'), covar=tensor([0.1295, 0.1103, 0.2121, 0.1290, 0.0957, 0.1060, 0.1460, 0.1977], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0251, 0.0141, 0.0123, 0.0134, 0.0155, 0.0119, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:52:51,572 INFO [finetune.py:976] (3/7) Epoch 8, batch 5500, loss[loss=0.185, simple_loss=0.2494, pruned_loss=0.06025, over 4831.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2568, pruned_loss=0.06432, over 957976.29 frames. ], batch size: 30, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:53:07,064 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0506, 2.5126, 1.0164, 1.2643, 1.8457, 1.1741, 3.0884, 1.5183], + device='cuda:3'), covar=tensor([0.0708, 0.0573, 0.0772, 0.1458, 0.0502, 0.1120, 0.0340, 0.0720], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 21:53:16,427 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.725e+02 1.934e+02 2.387e+02 4.697e+02, threshold=3.868e+02, percent-clipped=2.0 +2023-04-26 21:53:24,570 INFO [finetune.py:976] (3/7) Epoch 8, batch 5550, loss[loss=0.171, simple_loss=0.2366, pruned_loss=0.05266, over 4887.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2592, pruned_loss=0.06529, over 956401.01 frames. ], batch size: 32, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:54:06,992 INFO [finetune.py:976] (3/7) Epoch 8, batch 5600, loss[loss=0.232, simple_loss=0.2873, pruned_loss=0.08833, over 4764.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2634, pruned_loss=0.06674, over 956007.54 frames. ], batch size: 54, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:54:12,146 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45702.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:13,287 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45704.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 21:54:17,989 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45712.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:54:30,163 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.723e+02 2.023e+02 2.514e+02 5.159e+02, threshold=4.046e+02, percent-clipped=6.0 +2023-04-26 21:54:32,583 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45737.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:32,619 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7938, 1.5605, 2.0457, 2.0619, 1.6134, 1.3922, 1.7458, 1.0777], + device='cuda:3'), covar=tensor([0.0610, 0.1009, 0.0550, 0.0736, 0.0771, 0.1338, 0.0724, 0.0941], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0074, 0.0073, 0.0068, 0.0077, 0.0097, 0.0079, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:54:36,957 INFO [finetune.py:976] (3/7) Epoch 8, batch 5650, loss[loss=0.2148, simple_loss=0.29, pruned_loss=0.06974, over 4816.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2653, pruned_loss=0.06679, over 955210.97 frames. ], batch size: 38, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:54:46,840 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45760.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:48,652 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:54,993 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45774.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:57,367 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2664, 1.8143, 2.1507, 2.5550, 2.1503, 1.7762, 1.3499, 1.9125], + device='cuda:3'), covar=tensor([0.4075, 0.3626, 0.1893, 0.2595, 0.3174, 0.3246, 0.4806, 0.2460], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0254, 0.0221, 0.0321, 0.0215, 0.0230, 0.0237, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 21:55:06,106 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-04-26 21:55:07,629 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45785.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:55:16,187 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5998, 2.5226, 2.0342, 2.9765, 2.5394, 2.6023, 1.0124, 2.5343], + device='cuda:3'), covar=tensor([0.2011, 0.1608, 0.2920, 0.2337, 0.3002, 0.2283, 0.5734, 0.2599], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0217, 0.0251, 0.0306, 0.0299, 0.0250, 0.0271, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 21:55:18,509 INFO [finetune.py:976] (3/7) Epoch 8, batch 5700, loss[loss=0.2153, simple_loss=0.2654, pruned_loss=0.08264, over 4418.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2623, pruned_loss=0.06719, over 939133.65 frames. ], batch size: 19, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:55:59,017 INFO [finetune.py:976] (3/7) Epoch 9, batch 0, loss[loss=0.2184, simple_loss=0.2815, pruned_loss=0.07758, over 4841.00 frames. ], tot_loss[loss=0.2184, simple_loss=0.2815, pruned_loss=0.07758, over 4841.00 frames. ], batch size: 49, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:55:59,017 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 21:56:04,661 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4455, 2.9853, 0.9301, 1.7355, 1.9131, 2.2533, 1.8538, 1.0532], + device='cuda:3'), covar=tensor([0.1278, 0.0999, 0.2029, 0.1276, 0.0979, 0.0936, 0.1466, 0.1788], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0253, 0.0142, 0.0124, 0.0135, 0.0156, 0.0120, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 21:56:14,949 INFO [finetune.py:1010] (3/7) Epoch 9, validation: loss=0.1554, simple_loss=0.2289, pruned_loss=0.04093, over 2265189.00 frames. +2023-04-26 21:56:14,949 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 21:56:33,615 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.748e+02 2.064e+02 2.356e+02 2.984e+02, threshold=4.129e+02, percent-clipped=0.0 +2023-04-26 21:56:34,969 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45835.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:56:46,618 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1278, 2.5626, 1.0422, 1.3921, 1.9009, 1.2064, 3.4978, 1.5961], + device='cuda:3'), covar=tensor([0.0649, 0.0642, 0.0736, 0.1283, 0.0506, 0.1014, 0.0329, 0.0651], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-26 21:57:05,568 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:57:19,435 INFO [finetune.py:976] (3/7) Epoch 9, batch 50, loss[loss=0.179, simple_loss=0.2543, pruned_loss=0.05188, over 4919.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2648, pruned_loss=0.06835, over 215138.65 frames. ], batch size: 46, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:57:38,682 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5929, 1.3087, 1.2982, 1.3344, 1.8544, 1.5010, 1.2823, 1.2446], + device='cuda:3'), covar=tensor([0.1771, 0.1394, 0.2094, 0.1520, 0.0700, 0.1626, 0.2054, 0.2193], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0326, 0.0354, 0.0303, 0.0341, 0.0327, 0.0311, 0.0357], + device='cuda:3'), out_proj_covar=tensor([6.6122e-05, 6.9124e-05, 7.6568e-05, 6.2735e-05, 7.1757e-05, 7.0199e-05, + 6.6910e-05, 7.6727e-05], device='cuda:3') +2023-04-26 21:57:51,858 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45920.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:57:52,935 INFO [finetune.py:976] (3/7) Epoch 9, batch 100, loss[loss=0.1941, simple_loss=0.2482, pruned_loss=0.06997, over 4692.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2564, pruned_loss=0.06418, over 379986.93 frames. ], batch size: 59, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:58:01,518 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.187e+02 1.655e+02 1.945e+02 2.383e+02 5.251e+02, threshold=3.889e+02, percent-clipped=1.0 +2023-04-26 21:58:24,833 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 21:58:26,128 INFO [finetune.py:976] (3/7) Epoch 9, batch 150, loss[loss=0.163, simple_loss=0.2372, pruned_loss=0.04435, over 4771.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2525, pruned_loss=0.06249, over 508648.57 frames. ], batch size: 28, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:58:47,804 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46004.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 21:58:59,027 INFO [finetune.py:976] (3/7) Epoch 9, batch 200, loss[loss=0.1595, simple_loss=0.2335, pruned_loss=0.04274, over 4853.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2511, pruned_loss=0.06242, over 606448.31 frames. ], batch size: 49, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:59:07,991 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.665e+01 1.625e+02 2.034e+02 2.353e+02 6.037e+02, threshold=4.068e+02, percent-clipped=2.0 +2023-04-26 21:59:19,779 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46052.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:59:23,463 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46058.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:59:32,393 INFO [finetune.py:976] (3/7) Epoch 9, batch 250, loss[loss=0.1861, simple_loss=0.2628, pruned_loss=0.05463, over 4705.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2559, pruned_loss=0.06384, over 684725.97 frames. ], batch size: 59, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:59:49,085 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.99 vs. limit=5.0 +2023-04-26 21:59:51,841 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 21:59:53,243 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 21:59:59,754 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 22:00:05,317 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 22:00:05,473 INFO [finetune.py:976] (3/7) Epoch 9, batch 300, loss[loss=0.2431, simple_loss=0.2952, pruned_loss=0.09552, over 4839.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.26, pruned_loss=0.06554, over 744306.20 frames. ], batch size: 44, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 22:00:10,883 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:00:12,630 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.245e+02 1.689e+02 1.934e+02 2.285e+02 6.162e+02, threshold=3.867e+02, percent-clipped=1.0 +2023-04-26 22:00:24,315 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2361, 1.4510, 1.6477, 1.7918, 1.6548, 1.7406, 1.7769, 1.7174], + device='cuda:3'), covar=tensor([0.5059, 0.7647, 0.6321, 0.5826, 0.7133, 1.0133, 0.7110, 0.6759], + device='cuda:3'), in_proj_covar=tensor([0.0320, 0.0381, 0.0314, 0.0324, 0.0339, 0.0400, 0.0361, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 22:00:30,407 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5693, 2.0522, 1.5948, 1.3625, 1.2265, 1.1999, 1.6432, 1.2197], + device='cuda:3'), covar=tensor([0.1777, 0.1380, 0.1530, 0.1964, 0.2551, 0.2094, 0.1080, 0.2121], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0216, 0.0171, 0.0204, 0.0205, 0.0185, 0.0162, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 22:00:42,789 INFO [finetune.py:976] (3/7) Epoch 9, batch 350, loss[loss=0.2364, simple_loss=0.3013, pruned_loss=0.08578, over 4925.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2625, pruned_loss=0.06657, over 789636.30 frames. ], batch size: 33, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 22:01:38,763 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46215.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:01:48,014 INFO [finetune.py:976] (3/7) Epoch 9, batch 400, loss[loss=0.1913, simple_loss=0.2757, pruned_loss=0.0535, over 4887.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2641, pruned_loss=0.06641, over 827886.48 frames. ], batch size: 43, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:02:01,096 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.329e+02 1.830e+02 2.182e+02 2.472e+02 8.213e+02, threshold=4.364e+02, percent-clipped=3.0 +2023-04-26 22:02:26,905 INFO [finetune.py:976] (3/7) Epoch 9, batch 450, loss[loss=0.1784, simple_loss=0.2427, pruned_loss=0.05703, over 4831.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2625, pruned_loss=0.06579, over 855980.38 frames. ], batch size: 30, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:02:29,462 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46276.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:02:54,297 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.26 vs. limit=5.0 +2023-04-26 22:03:00,649 INFO [finetune.py:976] (3/7) Epoch 9, batch 500, loss[loss=0.2494, simple_loss=0.2997, pruned_loss=0.09958, over 4863.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2607, pruned_loss=0.0656, over 880562.19 frames. ], batch size: 44, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:03:07,761 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.712e+02 1.942e+02 2.347e+02 4.298e+02, threshold=3.884e+02, percent-clipped=0.0 +2023-04-26 22:03:10,768 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46337.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:03:26,001 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46358.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:03:34,346 INFO [finetune.py:976] (3/7) Epoch 9, batch 550, loss[loss=0.2614, simple_loss=0.311, pruned_loss=0.1059, over 4856.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2574, pruned_loss=0.0644, over 897756.18 frames. ], batch size: 44, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:03:57,941 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46406.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:04:07,673 INFO [finetune.py:976] (3/7) Epoch 9, batch 600, loss[loss=0.1941, simple_loss=0.2605, pruned_loss=0.06379, over 4817.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2569, pruned_loss=0.06436, over 909263.19 frames. ], batch size: 51, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:04:10,861 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 22:04:12,576 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46430.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:04:14,309 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.294e+02 1.692e+02 1.978e+02 2.446e+02 4.653e+02, threshold=3.955e+02, percent-clipped=2.0 +2023-04-26 22:04:40,995 INFO [finetune.py:976] (3/7) Epoch 9, batch 650, loss[loss=0.1799, simple_loss=0.2415, pruned_loss=0.05912, over 4770.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2589, pruned_loss=0.06496, over 918433.86 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:04:44,728 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46478.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:05:10,666 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46515.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:05:14,903 INFO [finetune.py:976] (3/7) Epoch 9, batch 700, loss[loss=0.2775, simple_loss=0.3325, pruned_loss=0.1112, over 4130.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2623, pruned_loss=0.06634, over 926345.39 frames. ], batch size: 65, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:05:21,595 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.297e+02 1.714e+02 2.090e+02 2.739e+02 4.841e+02, threshold=4.179e+02, percent-clipped=4.0 +2023-04-26 22:05:25,031 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-26 22:05:43,204 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46563.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:05:53,548 INFO [finetune.py:976] (3/7) Epoch 9, batch 750, loss[loss=0.198, simple_loss=0.2751, pruned_loss=0.06043, over 4875.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2633, pruned_loss=0.0661, over 933961.86 frames. ], batch size: 43, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:05:56,503 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-26 22:06:18,448 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46595.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:06:28,901 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46602.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:06:36,883 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 22:06:40,584 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6827, 1.3791, 1.9140, 1.8689, 1.5065, 1.3945, 1.5509, 1.0439], + device='cuda:3'), covar=tensor([0.0689, 0.1176, 0.0524, 0.1021, 0.0955, 0.1318, 0.0858, 0.0881], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0074, 0.0073, 0.0068, 0.0076, 0.0097, 0.0079, 0.0075], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:06:58,618 INFO [finetune.py:976] (3/7) Epoch 9, batch 800, loss[loss=0.1982, simple_loss=0.241, pruned_loss=0.07763, over 4466.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2622, pruned_loss=0.06542, over 937495.05 frames. ], batch size: 19, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:07:09,959 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46632.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:07:10,489 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.620e+02 1.991e+02 2.380e+02 4.805e+02, threshold=3.982e+02, percent-clipped=1.0 +2023-04-26 22:07:26,204 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46656.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:07:27,410 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-26 22:07:31,867 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46663.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:07:37,206 INFO [finetune.py:976] (3/7) Epoch 9, batch 850, loss[loss=0.189, simple_loss=0.2525, pruned_loss=0.06277, over 4161.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.26, pruned_loss=0.06431, over 942958.18 frames. ], batch size: 65, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:07:40,970 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6695, 1.5240, 0.7638, 1.2833, 1.8622, 1.5047, 1.3488, 1.4574], + device='cuda:3'), covar=tensor([0.0513, 0.0391, 0.0390, 0.0572, 0.0266, 0.0554, 0.0505, 0.0603], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 22:07:44,593 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8030, 1.8844, 1.6642, 1.4988, 2.0015, 1.5358, 2.4486, 1.5266], + device='cuda:3'), covar=tensor([0.3692, 0.1630, 0.4670, 0.2873, 0.1742, 0.2517, 0.1483, 0.4314], + device='cuda:3'), in_proj_covar=tensor([0.0348, 0.0352, 0.0434, 0.0366, 0.0391, 0.0386, 0.0385, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 22:08:10,355 INFO [finetune.py:976] (3/7) Epoch 9, batch 900, loss[loss=0.2153, simple_loss=0.2804, pruned_loss=0.07509, over 4839.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2577, pruned_loss=0.06349, over 946825.66 frames. ], batch size: 47, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:08:17,020 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.197e+02 1.643e+02 2.086e+02 2.450e+02 5.487e+02, threshold=4.172e+02, percent-clipped=3.0 +2023-04-26 22:08:43,450 INFO [finetune.py:976] (3/7) Epoch 9, batch 950, loss[loss=0.1965, simple_loss=0.2552, pruned_loss=0.06884, over 4933.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2563, pruned_loss=0.0632, over 949859.24 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:08:49,092 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 22:09:08,152 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46810.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:09:16,377 INFO [finetune.py:976] (3/7) Epoch 9, batch 1000, loss[loss=0.1599, simple_loss=0.2311, pruned_loss=0.04432, over 4882.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2595, pruned_loss=0.06436, over 951558.53 frames. ], batch size: 32, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:09:22,998 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.599e+02 1.946e+02 2.521e+02 4.625e+02, threshold=3.893e+02, percent-clipped=2.0 +2023-04-26 22:09:40,481 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7974, 1.6970, 2.0425, 2.2201, 1.7016, 1.4476, 1.8531, 0.9830], + device='cuda:3'), covar=tensor([0.0869, 0.0829, 0.0674, 0.0832, 0.0867, 0.1231, 0.0837, 0.1166], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0076, 0.0096, 0.0079, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:09:49,152 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46871.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:09:49,627 INFO [finetune.py:976] (3/7) Epoch 9, batch 1050, loss[loss=0.1883, simple_loss=0.2501, pruned_loss=0.06327, over 4884.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2625, pruned_loss=0.06526, over 953339.68 frames. ], batch size: 32, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:09:53,393 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1192, 2.5676, 1.1121, 1.3404, 2.0604, 1.2262, 3.5483, 1.6962], + device='cuda:3'), covar=tensor([0.0669, 0.0657, 0.0757, 0.1339, 0.0511, 0.1008, 0.0233, 0.0686], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0067, 0.0050, 0.0048, 0.0052, 0.0052, 0.0078, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-26 22:10:08,586 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0218, 2.4212, 1.0313, 1.3582, 1.8519, 1.2261, 3.2987, 1.7557], + device='cuda:3'), covar=tensor([0.0712, 0.0642, 0.0781, 0.1316, 0.0556, 0.1003, 0.0254, 0.0635], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0067, 0.0050, 0.0048, 0.0052, 0.0052, 0.0078, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-26 22:10:22,800 INFO [finetune.py:976] (3/7) Epoch 9, batch 1100, loss[loss=0.1665, simple_loss=0.2552, pruned_loss=0.03892, over 4924.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2634, pruned_loss=0.06546, over 953987.83 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:10:29,453 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46932.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:10:30,585 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.585e+02 1.927e+02 2.625e+02 4.611e+02, threshold=3.853e+02, percent-clipped=4.0 +2023-04-26 22:10:40,917 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46951.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 22:10:45,717 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46958.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:10:56,081 INFO [finetune.py:976] (3/7) Epoch 9, batch 1150, loss[loss=0.1726, simple_loss=0.2535, pruned_loss=0.04591, over 4889.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2646, pruned_loss=0.0659, over 955497.65 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:11:00,334 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8912, 2.4338, 1.9807, 2.2001, 1.8091, 2.0264, 2.0307, 1.4820], + device='cuda:3'), covar=tensor([0.2371, 0.1431, 0.1055, 0.1470, 0.3680, 0.1494, 0.2287, 0.3275], + device='cuda:3'), in_proj_covar=tensor([0.0301, 0.0321, 0.0230, 0.0293, 0.0319, 0.0274, 0.0261, 0.0286], + device='cuda:3'), out_proj_covar=tensor([1.2202e-04, 1.2918e-04, 9.2765e-05, 1.1734e-04, 1.3100e-04, 1.1041e-04, + 1.0679e-04, 1.1489e-04], device='cuda:3') +2023-04-26 22:11:01,459 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46980.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:11:30,332 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7547, 1.7664, 2.0041, 2.1715, 1.7315, 1.3524, 1.8241, 1.0691], + device='cuda:3'), covar=tensor([0.0944, 0.0841, 0.0607, 0.0913, 0.0795, 0.1220, 0.0885, 0.1002], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0074, 0.0072, 0.0067, 0.0076, 0.0096, 0.0079, 0.0075], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:11:43,960 INFO [finetune.py:976] (3/7) Epoch 9, batch 1200, loss[loss=0.2055, simple_loss=0.2688, pruned_loss=0.07109, over 4923.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2628, pruned_loss=0.0652, over 955828.87 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:12:03,718 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.560e+02 1.899e+02 2.282e+02 5.310e+02, threshold=3.798e+02, percent-clipped=1.0 +2023-04-26 22:12:55,840 INFO [finetune.py:976] (3/7) Epoch 9, batch 1250, loss[loss=0.1828, simple_loss=0.2496, pruned_loss=0.05806, over 4860.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2598, pruned_loss=0.0645, over 955457.65 frames. ], batch size: 34, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:14:00,073 INFO [finetune.py:976] (3/7) Epoch 9, batch 1300, loss[loss=0.1967, simple_loss=0.2595, pruned_loss=0.06702, over 4902.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2563, pruned_loss=0.06335, over 956587.17 frames. ], batch size: 37, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:14:15,266 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.673e+02 1.914e+02 2.315e+02 4.014e+02, threshold=3.828e+02, percent-clipped=2.0 +2023-04-26 22:14:58,580 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47166.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:15:07,950 INFO [finetune.py:976] (3/7) Epoch 9, batch 1350, loss[loss=0.2134, simple_loss=0.2915, pruned_loss=0.06768, over 4725.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2565, pruned_loss=0.06334, over 958225.48 frames. ], batch size: 59, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:15:58,006 INFO [finetune.py:976] (3/7) Epoch 9, batch 1400, loss[loss=0.1975, simple_loss=0.2606, pruned_loss=0.06726, over 4765.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2599, pruned_loss=0.06456, over 958303.53 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:16:01,101 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9629, 1.7248, 2.0882, 2.3339, 2.0359, 1.9146, 2.0276, 2.0447], + device='cuda:3'), covar=tensor([0.6692, 0.8210, 0.8762, 0.9186, 0.8406, 1.0313, 1.0560, 0.8663], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0419, 0.0503, 0.0523, 0.0437, 0.0454, 0.0467, 0.0464], + device='cuda:3'), out_proj_covar=tensor([9.9473e-05, 1.0393e-04, 1.1352e-04, 1.2425e-04, 1.0619e-04, 1.0991e-04, + 1.1235e-04, 1.1230e-04], device='cuda:3') +2023-04-26 22:16:06,774 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.248e+02 1.819e+02 2.122e+02 2.454e+02 4.582e+02, threshold=4.244e+02, percent-clipped=5.0 +2023-04-26 22:16:11,717 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47241.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:17,789 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47251.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:22,002 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47258.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:30,887 INFO [finetune.py:976] (3/7) Epoch 9, batch 1450, loss[loss=0.21, simple_loss=0.2729, pruned_loss=0.07358, over 4897.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.262, pruned_loss=0.06513, over 957648.47 frames. ], batch size: 37, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:16:50,057 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47299.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:51,885 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47302.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:53,688 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8346, 3.7822, 2.7182, 4.4217, 3.9003, 3.8412, 1.6587, 3.7388], + device='cuda:3'), covar=tensor([0.1624, 0.1256, 0.3086, 0.1426, 0.2303, 0.1767, 0.5576, 0.2280], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0217, 0.0251, 0.0306, 0.0300, 0.0252, 0.0271, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 22:16:54,271 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47306.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:17:03,914 INFO [finetune.py:976] (3/7) Epoch 9, batch 1500, loss[loss=0.2152, simple_loss=0.2874, pruned_loss=0.07151, over 4882.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2647, pruned_loss=0.06672, over 956086.56 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:17:12,630 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.650e+02 1.825e+02 2.290e+02 4.290e+02, threshold=3.651e+02, percent-clipped=1.0 +2023-04-26 22:18:03,818 INFO [finetune.py:976] (3/7) Epoch 9, batch 1550, loss[loss=0.2489, simple_loss=0.3046, pruned_loss=0.0966, over 4884.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2649, pruned_loss=0.06706, over 954534.90 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:18:13,661 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47379.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:18:48,894 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 22:18:49,435 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0775, 1.9548, 2.2214, 2.4825, 2.5146, 2.0269, 1.5834, 2.1716], + device='cuda:3'), covar=tensor([0.0839, 0.0937, 0.0592, 0.0565, 0.0493, 0.0800, 0.0828, 0.0560], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0205, 0.0182, 0.0177, 0.0178, 0.0190, 0.0161, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 22:18:49,866 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-26 22:19:10,274 INFO [finetune.py:976] (3/7) Epoch 9, batch 1600, loss[loss=0.1828, simple_loss=0.2477, pruned_loss=0.05894, over 4853.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2626, pruned_loss=0.06642, over 955525.27 frames. ], batch size: 44, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:19:21,112 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-26 22:19:23,854 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.649e+02 1.912e+02 2.310e+02 4.649e+02, threshold=3.824e+02, percent-clipped=3.0 +2023-04-26 22:19:34,494 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47440.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:19:35,645 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0438, 1.4770, 1.8622, 2.0206, 1.8381, 1.4463, 0.8903, 1.5267], + device='cuda:3'), covar=tensor([0.3621, 0.3968, 0.1863, 0.2823, 0.3095, 0.3122, 0.5241, 0.2695], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0252, 0.0220, 0.0320, 0.0214, 0.0229, 0.0235, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 22:20:08,740 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47466.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:20:17,937 INFO [finetune.py:976] (3/7) Epoch 9, batch 1650, loss[loss=0.1395, simple_loss=0.2115, pruned_loss=0.03379, over 4822.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2587, pruned_loss=0.06489, over 953847.62 frames. ], batch size: 39, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:20:46,558 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 22:21:00,461 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47514.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:21:05,270 INFO [finetune.py:976] (3/7) Epoch 9, batch 1700, loss[loss=0.1562, simple_loss=0.225, pruned_loss=0.04367, over 4901.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2567, pruned_loss=0.06447, over 955801.98 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:21:12,553 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.638e+02 2.054e+02 2.588e+02 3.948e+02, threshold=4.108e+02, percent-clipped=3.0 +2023-04-26 22:21:14,961 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9552, 1.5906, 1.5642, 1.6643, 2.2025, 1.8204, 1.4324, 1.4970], + device='cuda:3'), covar=tensor([0.1748, 0.1374, 0.1992, 0.1238, 0.0816, 0.1336, 0.2309, 0.2161], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0328, 0.0358, 0.0303, 0.0340, 0.0328, 0.0311, 0.0358], + device='cuda:3'), out_proj_covar=tensor([6.6300e-05, 6.9546e-05, 7.7312e-05, 6.2706e-05, 7.1399e-05, 7.0452e-05, + 6.6853e-05, 7.6904e-05], device='cuda:3') +2023-04-26 22:21:38,481 INFO [finetune.py:976] (3/7) Epoch 9, batch 1750, loss[loss=0.2192, simple_loss=0.2986, pruned_loss=0.06992, over 4903.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2599, pruned_loss=0.06528, over 956293.06 frames. ], batch size: 43, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:21:50,064 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7806, 2.3260, 1.7863, 1.6314, 1.2727, 1.3524, 1.9240, 1.2622], + device='cuda:3'), covar=tensor([0.1821, 0.1576, 0.1691, 0.2091, 0.2686, 0.2305, 0.1170, 0.2298], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0215, 0.0170, 0.0204, 0.0204, 0.0183, 0.0160, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 22:21:54,722 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47597.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:22:11,821 INFO [finetune.py:976] (3/7) Epoch 9, batch 1800, loss[loss=0.1681, simple_loss=0.2496, pruned_loss=0.04332, over 4775.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.264, pruned_loss=0.06673, over 957096.60 frames. ], batch size: 29, lr: 3.79e-03, grad_scale: 32.0 +2023-04-26 22:22:19,164 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.815e+02 2.187e+02 2.572e+02 4.070e+02, threshold=4.375e+02, percent-clipped=0.0 +2023-04-26 22:22:45,318 INFO [finetune.py:976] (3/7) Epoch 9, batch 1850, loss[loss=0.1849, simple_loss=0.2608, pruned_loss=0.05446, over 4749.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.265, pruned_loss=0.06679, over 955182.83 frames. ], batch size: 27, lr: 3.79e-03, grad_scale: 32.0 +2023-04-26 22:23:44,012 INFO [finetune.py:976] (3/7) Epoch 9, batch 1900, loss[loss=0.2299, simple_loss=0.2886, pruned_loss=0.08563, over 4881.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2668, pruned_loss=0.0671, over 956899.35 frames. ], batch size: 35, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:24:02,110 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.717e+02 1.936e+02 2.333e+02 4.969e+02, threshold=3.871e+02, percent-clipped=1.0 +2023-04-26 22:24:02,210 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47735.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:04,114 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47738.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:14,278 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9749, 1.0411, 1.1764, 1.2025, 0.9991, 0.8308, 1.0026, 0.6070], + device='cuda:3'), covar=tensor([0.0722, 0.0930, 0.0816, 0.0711, 0.0869, 0.1631, 0.0728, 0.0988], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0076, 0.0096, 0.0078, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:24:33,493 INFO [finetune.py:976] (3/7) Epoch 9, batch 1950, loss[loss=0.2019, simple_loss=0.2658, pruned_loss=0.06901, over 4913.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2641, pruned_loss=0.06603, over 955250.47 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:24:43,270 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47788.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:50,420 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47799.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:57,021 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47809.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:25:11,660 INFO [finetune.py:976] (3/7) Epoch 9, batch 2000, loss[loss=0.2127, simple_loss=0.2708, pruned_loss=0.07726, over 4914.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2601, pruned_loss=0.06455, over 954095.47 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:25:30,747 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.686e+02 2.019e+02 2.338e+02 4.477e+02, threshold=4.037e+02, percent-clipped=4.0 +2023-04-26 22:25:44,884 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:25:51,491 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 22:26:17,388 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47870.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 22:26:24,623 INFO [finetune.py:976] (3/7) Epoch 9, batch 2050, loss[loss=0.1442, simple_loss=0.2127, pruned_loss=0.0379, over 4791.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2563, pruned_loss=0.06308, over 954740.58 frames. ], batch size: 29, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:26:24,728 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9662, 1.4393, 4.8359, 4.5056, 4.2662, 4.5408, 4.2551, 4.2806], + device='cuda:3'), covar=tensor([0.6897, 0.6408, 0.1003, 0.1873, 0.1108, 0.1327, 0.1673, 0.1523], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0312, 0.0411, 0.0416, 0.0354, 0.0411, 0.0320, 0.0373], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 22:26:57,906 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47897.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:27:12,930 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2236, 1.5529, 2.0798, 2.5674, 2.1034, 1.6185, 1.2171, 1.8618], + device='cuda:3'), covar=tensor([0.3590, 0.3852, 0.1826, 0.2790, 0.3273, 0.2988, 0.5072, 0.2687], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0252, 0.0221, 0.0321, 0.0214, 0.0229, 0.0236, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 22:27:30,786 INFO [finetune.py:976] (3/7) Epoch 9, batch 2100, loss[loss=0.1666, simple_loss=0.2336, pruned_loss=0.04978, over 4782.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2555, pruned_loss=0.06258, over 953309.99 frames. ], batch size: 26, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:27:39,270 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.719e+02 2.114e+02 2.573e+02 5.213e+02, threshold=4.228e+02, percent-clipped=2.0 +2023-04-26 22:27:42,816 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-04-26 22:27:43,691 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8227, 1.3332, 1.3470, 1.6239, 1.9989, 1.6338, 1.3649, 1.2755], + device='cuda:3'), covar=tensor([0.1467, 0.1466, 0.1799, 0.1093, 0.0808, 0.1388, 0.2079, 0.1918], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0324, 0.0352, 0.0298, 0.0337, 0.0325, 0.0306, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.5429e-05, 6.8875e-05, 7.6104e-05, 6.1639e-05, 7.0678e-05, 6.9867e-05, + 6.5901e-05, 7.5887e-05], device='cuda:3') +2023-04-26 22:27:45,501 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47945.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:28:20,274 INFO [finetune.py:976] (3/7) Epoch 9, batch 2150, loss[loss=0.2347, simple_loss=0.305, pruned_loss=0.08219, over 4801.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2606, pruned_loss=0.06482, over 953817.07 frames. ], batch size: 51, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:28:31,442 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 22:28:48,482 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4826, 1.5922, 0.6684, 1.2458, 1.4781, 1.3200, 1.2738, 1.3674], + device='cuda:3'), covar=tensor([0.0635, 0.0357, 0.0433, 0.0649, 0.0325, 0.0710, 0.0701, 0.0659], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:3') +2023-04-26 22:29:17,268 INFO [finetune.py:976] (3/7) Epoch 9, batch 2200, loss[loss=0.1685, simple_loss=0.2402, pruned_loss=0.04847, over 4820.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2625, pruned_loss=0.06579, over 953819.63 frames. ], batch size: 40, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:29:28,720 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5849, 1.7355, 0.6800, 1.2406, 1.7198, 1.4092, 1.3479, 1.3580], + device='cuda:3'), covar=tensor([0.0649, 0.0361, 0.0418, 0.0676, 0.0297, 0.0726, 0.0716, 0.0707], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:3') +2023-04-26 22:29:32,259 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.341e+02 1.685e+02 2.065e+02 2.367e+02 5.445e+02, threshold=4.130e+02, percent-clipped=2.0 +2023-04-26 22:29:32,374 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48035.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:29:54,597 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-26 22:30:13,117 INFO [finetune.py:976] (3/7) Epoch 9, batch 2250, loss[loss=0.2836, simple_loss=0.3333, pruned_loss=0.117, over 4135.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2638, pruned_loss=0.0664, over 952368.84 frames. ], batch size: 65, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:30:18,364 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48072.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:30:32,861 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48083.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:30:45,758 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48094.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:31:21,272 INFO [finetune.py:976] (3/7) Epoch 9, batch 2300, loss[loss=0.1592, simple_loss=0.2277, pruned_loss=0.04535, over 4783.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2637, pruned_loss=0.06564, over 954099.71 frames. ], batch size: 25, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:31:35,724 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48133.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:31:36,817 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.203e+02 1.671e+02 1.981e+02 2.243e+02 3.920e+02, threshold=3.963e+02, percent-clipped=0.0 +2023-04-26 22:31:50,252 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48144.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:32:14,429 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48165.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 22:32:24,175 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4427, 1.6289, 1.6624, 2.2521, 2.4586, 2.0499, 1.8649, 1.7523], + device='cuda:3'), covar=tensor([0.1521, 0.2022, 0.1947, 0.1723, 0.1202, 0.1990, 0.2634, 0.2299], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0325, 0.0353, 0.0297, 0.0336, 0.0324, 0.0306, 0.0355], + device='cuda:3'), out_proj_covar=tensor([6.5184e-05, 6.8940e-05, 7.6227e-05, 6.1256e-05, 7.0361e-05, 6.9582e-05, + 6.5858e-05, 7.6167e-05], device='cuda:3') +2023-04-26 22:32:25,265 INFO [finetune.py:976] (3/7) Epoch 9, batch 2350, loss[loss=0.1504, simple_loss=0.2323, pruned_loss=0.03432, over 4910.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2608, pruned_loss=0.06476, over 955134.94 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:33:19,203 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48212.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 22:33:21,096 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3824, 1.1692, 1.6316, 1.4853, 1.2581, 1.1090, 1.2823, 0.8313], + device='cuda:3'), covar=tensor([0.0560, 0.0741, 0.0555, 0.0818, 0.0892, 0.1366, 0.0660, 0.0837], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0076, 0.0096, 0.0078, 0.0074], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:33:30,439 INFO [finetune.py:976] (3/7) Epoch 9, batch 2400, loss[loss=0.1585, simple_loss=0.2356, pruned_loss=0.04065, over 4904.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2581, pruned_loss=0.06449, over 953365.09 frames. ], batch size: 35, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:33:45,096 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.712e+02 1.962e+02 2.337e+02 6.803e+02, threshold=3.925e+02, percent-clipped=4.0 +2023-04-26 22:34:34,223 INFO [finetune.py:976] (3/7) Epoch 9, batch 2450, loss[loss=0.1949, simple_loss=0.2615, pruned_loss=0.06413, over 4826.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2561, pruned_loss=0.06404, over 953210.85 frames. ], batch size: 33, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:34:34,979 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48273.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 22:34:39,045 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9748, 2.6114, 1.9944, 2.2960, 1.8955, 2.0727, 1.9983, 1.8237], + device='cuda:3'), covar=tensor([0.2205, 0.1150, 0.1021, 0.1338, 0.2732, 0.1193, 0.2084, 0.2449], + device='cuda:3'), in_proj_covar=tensor([0.0296, 0.0316, 0.0228, 0.0290, 0.0313, 0.0271, 0.0258, 0.0281], + device='cuda:3'), out_proj_covar=tensor([1.1981e-04, 1.2736e-04, 9.1730e-05, 1.1600e-04, 1.2843e-04, 1.0915e-04, + 1.0567e-04, 1.1263e-04], device='cuda:3') +2023-04-26 22:35:23,668 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6515, 1.7459, 1.8133, 1.9356, 1.7312, 1.9180, 1.9949, 1.8948], + device='cuda:3'), covar=tensor([0.4785, 0.7915, 0.6682, 0.5987, 0.7420, 1.0441, 0.7184, 0.7386], + device='cuda:3'), in_proj_covar=tensor([0.0324, 0.0385, 0.0317, 0.0327, 0.0343, 0.0406, 0.0365, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 22:35:25,353 INFO [finetune.py:976] (3/7) Epoch 9, batch 2500, loss[loss=0.183, simple_loss=0.254, pruned_loss=0.05594, over 4859.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2582, pruned_loss=0.06491, over 954082.19 frames. ], batch size: 31, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:35:26,619 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7464, 4.0704, 1.1143, 1.9232, 2.1467, 2.8203, 2.3398, 0.8757], + device='cuda:3'), covar=tensor([0.1426, 0.0773, 0.1984, 0.1503, 0.1106, 0.1015, 0.1463, 0.2325], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0252, 0.0142, 0.0124, 0.0136, 0.0155, 0.0118, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:35:26,643 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48324.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:35:34,245 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.734e+02 2.061e+02 2.480e+02 6.002e+02, threshold=4.122e+02, percent-clipped=6.0 +2023-04-26 22:35:58,790 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.41 vs. limit=5.0 +2023-04-26 22:36:14,629 INFO [finetune.py:976] (3/7) Epoch 9, batch 2550, loss[loss=0.1977, simple_loss=0.2657, pruned_loss=0.06484, over 4925.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2614, pruned_loss=0.06508, over 954858.14 frames. ], batch size: 38, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:36:34,735 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48385.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:36:53,727 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48394.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:36:55,536 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1289, 4.1623, 0.7412, 2.1195, 2.5254, 2.8672, 2.4763, 0.9343], + device='cuda:3'), covar=tensor([0.1231, 0.1136, 0.2436, 0.1431, 0.0989, 0.1083, 0.1504, 0.2318], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0250, 0.0141, 0.0123, 0.0135, 0.0154, 0.0118, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:36:55,560 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48397.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:28,005 INFO [finetune.py:976] (3/7) Epoch 9, batch 2600, loss[loss=0.2101, simple_loss=0.2799, pruned_loss=0.07015, over 4725.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2623, pruned_loss=0.0651, over 953074.80 frames. ], batch size: 59, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:37:28,726 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48423.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:31,743 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48428.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:41,882 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.714e+02 2.123e+02 2.498e+02 4.905e+02, threshold=4.246e+02, percent-clipped=1.0 +2023-04-26 22:37:52,940 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48442.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:54,676 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48444.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:54,744 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6575, 1.3506, 1.7896, 2.1366, 1.7935, 1.6508, 1.7074, 1.7831], + device='cuda:3'), covar=tensor([0.6928, 0.9259, 0.9017, 0.9135, 0.8463, 1.1850, 1.1497, 1.0427], + device='cuda:3'), in_proj_covar=tensor([0.0407, 0.0419, 0.0501, 0.0521, 0.0437, 0.0455, 0.0468, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.9199e-05, 1.0391e-04, 1.1320e-04, 1.2388e-04, 1.0609e-04, 1.1015e-04, + 1.1245e-04, 1.1252e-04], device='cuda:3') +2023-04-26 22:38:14,005 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:38:24,203 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48465.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 22:38:34,662 INFO [finetune.py:976] (3/7) Epoch 9, batch 2650, loss[loss=0.1811, simple_loss=0.2485, pruned_loss=0.05687, over 4860.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.2637, pruned_loss=0.06554, over 955899.75 frames. ], batch size: 34, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:38:46,960 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48484.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:38:57,119 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48492.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:39:21,187 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48513.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:39:32,428 INFO [finetune.py:976] (3/7) Epoch 9, batch 2700, loss[loss=0.1532, simple_loss=0.2209, pruned_loss=0.0428, over 4812.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2622, pruned_loss=0.06458, over 956031.56 frames. ], batch size: 40, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:39:40,875 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.732e+02 1.990e+02 2.443e+02 3.754e+02, threshold=3.980e+02, percent-clipped=0.0 +2023-04-26 22:40:20,118 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48568.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:40:28,389 INFO [finetune.py:976] (3/7) Epoch 9, batch 2750, loss[loss=0.1331, simple_loss=0.2041, pruned_loss=0.03101, over 4748.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2583, pruned_loss=0.06368, over 953052.92 frames. ], batch size: 26, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:41:33,773 INFO [finetune.py:976] (3/7) Epoch 9, batch 2800, loss[loss=0.1615, simple_loss=0.2286, pruned_loss=0.04719, over 4831.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2551, pruned_loss=0.06239, over 954096.84 frames. ], batch size: 38, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:41:46,303 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.615e+02 1.895e+02 2.284e+02 5.384e+02, threshold=3.791e+02, percent-clipped=3.0 +2023-04-26 22:42:28,874 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1064, 2.2008, 1.9506, 1.8127, 2.4803, 1.8312, 2.9476, 1.7210], + device='cuda:3'), covar=tensor([0.3673, 0.1663, 0.4467, 0.3177, 0.1557, 0.2706, 0.1255, 0.4379], + device='cuda:3'), in_proj_covar=tensor([0.0348, 0.0352, 0.0433, 0.0364, 0.0392, 0.0386, 0.0382, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 22:42:38,718 INFO [finetune.py:976] (3/7) Epoch 9, batch 2850, loss[loss=0.1877, simple_loss=0.2507, pruned_loss=0.06236, over 4891.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2535, pruned_loss=0.06157, over 954000.76 frames. ], batch size: 35, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:42:49,420 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48680.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:43:00,134 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4310, 0.6442, 1.2963, 1.8684, 1.5759, 1.3638, 1.3443, 1.4152], + device='cuda:3'), covar=tensor([0.5779, 0.7915, 0.7689, 0.8481, 0.6910, 0.9143, 0.8879, 0.8738], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0421, 0.0504, 0.0522, 0.0438, 0.0458, 0.0469, 0.0466], + device='cuda:3'), out_proj_covar=tensor([9.9549e-05, 1.0445e-04, 1.1384e-04, 1.2413e-04, 1.0628e-04, 1.1068e-04, + 1.1285e-04, 1.1269e-04], device='cuda:3') +2023-04-26 22:43:11,125 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7658, 1.3173, 1.6253, 1.6362, 1.6109, 1.2876, 0.7178, 1.2851], + device='cuda:3'), covar=tensor([0.3642, 0.3498, 0.1777, 0.2441, 0.2767, 0.2773, 0.4706, 0.2496], + device='cuda:3'), in_proj_covar=tensor([0.0280, 0.0249, 0.0219, 0.0317, 0.0212, 0.0227, 0.0232, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 22:43:44,377 INFO [finetune.py:976] (3/7) Epoch 9, batch 2900, loss[loss=0.1837, simple_loss=0.2715, pruned_loss=0.04792, over 4814.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2555, pruned_loss=0.06202, over 953695.89 frames. ], batch size: 40, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:43:48,143 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48728.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:43:52,840 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.707e+02 2.011e+02 2.462e+02 4.094e+02, threshold=4.022e+02, percent-clipped=3.0 +2023-04-26 22:44:04,269 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48753.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:44:14,552 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 22:44:17,828 INFO [finetune.py:976] (3/7) Epoch 9, batch 2950, loss[loss=0.1662, simple_loss=0.2401, pruned_loss=0.04611, over 4896.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2593, pruned_loss=0.06325, over 953174.64 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:44:20,345 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48776.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:44:22,223 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48779.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:45:10,533 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48821.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:45:11,036 INFO [finetune.py:976] (3/7) Epoch 9, batch 3000, loss[loss=0.2677, simple_loss=0.3141, pruned_loss=0.1107, over 4908.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2605, pruned_loss=0.06343, over 953842.64 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:45:11,036 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 22:45:27,360 INFO [finetune.py:1010] (3/7) Epoch 9, validation: loss=0.1543, simple_loss=0.2267, pruned_loss=0.04097, over 2265189.00 frames. +2023-04-26 22:45:27,361 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 22:45:46,031 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.720e+02 1.968e+02 2.331e+02 3.766e+02, threshold=3.936e+02, percent-clipped=0.0 +2023-04-26 22:45:47,995 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48838.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:46:21,738 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-26 22:46:30,381 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48868.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 22:46:32,696 INFO [finetune.py:976] (3/7) Epoch 9, batch 3050, loss[loss=0.2249, simple_loss=0.2789, pruned_loss=0.0854, over 4824.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2627, pruned_loss=0.06486, over 952003.79 frames. ], batch size: 47, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:46:51,109 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48882.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:47:12,138 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48899.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:47:28,834 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48916.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 22:47:32,444 INFO [finetune.py:976] (3/7) Epoch 9, batch 3100, loss[loss=0.1674, simple_loss=0.2159, pruned_loss=0.05942, over 4201.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2607, pruned_loss=0.06478, over 952765.54 frames. ], batch size: 18, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:47:42,245 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.703e+02 2.023e+02 2.523e+02 5.306e+02, threshold=4.046e+02, percent-clipped=4.0 +2023-04-26 22:48:00,708 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 22:48:04,665 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-04-26 22:48:05,673 INFO [finetune.py:976] (3/7) Epoch 9, batch 3150, loss[loss=0.1786, simple_loss=0.2406, pruned_loss=0.05835, over 4793.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2584, pruned_loss=0.06408, over 952389.66 frames. ], batch size: 51, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:48:11,550 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48980.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:48:38,602 INFO [finetune.py:976] (3/7) Epoch 9, batch 3200, loss[loss=0.1969, simple_loss=0.2579, pruned_loss=0.06797, over 4718.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2556, pruned_loss=0.06279, over 952293.95 frames. ], batch size: 54, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:48:42,797 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49028.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:48:47,938 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.546e+02 1.866e+02 2.430e+02 3.518e+02, threshold=3.733e+02, percent-clipped=0.0 +2023-04-26 22:48:59,915 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49053.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:49:12,011 INFO [finetune.py:976] (3/7) Epoch 9, batch 3250, loss[loss=0.2147, simple_loss=0.2801, pruned_loss=0.07468, over 4761.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2561, pruned_loss=0.06319, over 952790.51 frames. ], batch size: 59, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:49:16,895 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49079.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:49:19,311 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4427, 3.4698, 0.7800, 1.9194, 1.9664, 2.2994, 1.9038, 0.9532], + device='cuda:3'), covar=tensor([0.1529, 0.1113, 0.2204, 0.1360, 0.1084, 0.1143, 0.1606, 0.2126], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0252, 0.0142, 0.0123, 0.0136, 0.0155, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:49:27,874 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0942, 2.6448, 1.0552, 1.3981, 1.9528, 1.4118, 3.4856, 1.9755], + device='cuda:3'), covar=tensor([0.0680, 0.0639, 0.0840, 0.1272, 0.0526, 0.0966, 0.0300, 0.0594], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0067, 0.0050, 0.0047, 0.0051, 0.0053, 0.0078, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-26 22:49:32,076 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49101.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:49:45,283 INFO [finetune.py:976] (3/7) Epoch 9, batch 3300, loss[loss=0.1278, simple_loss=0.1919, pruned_loss=0.03187, over 4724.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2602, pruned_loss=0.06483, over 953765.48 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:49:48,902 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49127.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:49:54,142 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.776e+02 1.982e+02 2.505e+02 4.368e+02, threshold=3.963e+02, percent-clipped=2.0 +2023-04-26 22:50:18,719 INFO [finetune.py:976] (3/7) Epoch 9, batch 3350, loss[loss=0.1592, simple_loss=0.2265, pruned_loss=0.04593, over 4773.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.263, pruned_loss=0.06585, over 954972.76 frames. ], batch size: 28, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:50:20,309 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 22:50:21,873 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49177.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:50:38,274 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-26 22:50:45,201 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4227, 0.6764, 1.3397, 1.8726, 1.5630, 1.3919, 1.3920, 1.4770], + device='cuda:3'), covar=tensor([0.5823, 0.7932, 0.7692, 0.8099, 0.7117, 0.9490, 0.9289, 0.8773], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0421, 0.0502, 0.0523, 0.0438, 0.0457, 0.0468, 0.0464], + device='cuda:3'), out_proj_covar=tensor([9.9418e-05, 1.0417e-04, 1.1343e-04, 1.2404e-04, 1.0618e-04, 1.1043e-04, + 1.1241e-04, 1.1237e-04], device='cuda:3') +2023-04-26 22:50:46,870 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49194.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:51:00,028 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.2187, 4.1967, 3.0733, 4.8532, 4.2401, 4.2254, 1.9433, 4.1267], + device='cuda:3'), covar=tensor([0.1481, 0.1004, 0.3489, 0.0863, 0.2521, 0.1497, 0.5219, 0.1933], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0217, 0.0251, 0.0306, 0.0302, 0.0252, 0.0271, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 22:51:19,528 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5089, 1.2974, 0.5361, 1.2168, 1.4236, 1.3625, 1.2901, 1.2813], + device='cuda:3'), covar=tensor([0.0529, 0.0422, 0.0451, 0.0600, 0.0323, 0.0552, 0.0501, 0.0616], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 22:51:20,614 INFO [finetune.py:976] (3/7) Epoch 9, batch 3400, loss[loss=0.2027, simple_loss=0.2649, pruned_loss=0.07022, over 4898.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2642, pruned_loss=0.06632, over 954304.49 frames. ], batch size: 35, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:51:38,492 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.560e+02 1.880e+02 2.304e+02 5.617e+02, threshold=3.759e+02, percent-clipped=2.0 +2023-04-26 22:51:39,224 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:51:53,748 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5506, 0.9722, 1.3226, 1.2696, 1.7373, 1.4132, 1.0968, 1.2974], + device='cuda:3'), covar=tensor([0.1961, 0.1900, 0.2449, 0.1690, 0.1018, 0.1613, 0.2459, 0.2392], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0324, 0.0354, 0.0296, 0.0335, 0.0321, 0.0307, 0.0355], + device='cuda:3'), out_proj_covar=tensor([6.5006e-05, 6.8722e-05, 7.6320e-05, 6.1246e-05, 7.0316e-05, 6.9002e-05, + 6.6136e-05, 7.6297e-05], device='cuda:3') +2023-04-26 22:51:54,293 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.4809, 4.3965, 3.1007, 5.1748, 4.5269, 4.5042, 1.9899, 4.4391], + device='cuda:3'), covar=tensor([0.1556, 0.1040, 0.3130, 0.0950, 0.2885, 0.1636, 0.5580, 0.2215], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0219, 0.0253, 0.0308, 0.0304, 0.0253, 0.0272, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 22:51:54,346 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2520, 2.9640, 2.2806, 2.6064, 2.0884, 2.4144, 2.5643, 2.0275], + device='cuda:3'), covar=tensor([0.2078, 0.1199, 0.0894, 0.1237, 0.3033, 0.1263, 0.2044, 0.2824], + device='cuda:3'), in_proj_covar=tensor([0.0295, 0.0315, 0.0227, 0.0287, 0.0311, 0.0270, 0.0256, 0.0280], + device='cuda:3'), out_proj_covar=tensor([1.1941e-04, 1.2710e-04, 9.1349e-05, 1.1507e-04, 1.2795e-04, 1.0904e-04, + 1.0502e-04, 1.1248e-04], device='cuda:3') +2023-04-26 22:52:05,037 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 22:52:13,320 INFO [finetune.py:976] (3/7) Epoch 9, batch 3450, loss[loss=0.1502, simple_loss=0.2205, pruned_loss=0.03989, over 4750.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.264, pruned_loss=0.06588, over 954486.51 frames. ], batch size: 27, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:52:14,699 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8600, 2.3509, 1.8952, 2.1688, 1.7033, 1.8475, 1.8949, 1.4928], + device='cuda:3'), covar=tensor([0.1983, 0.1180, 0.0907, 0.1132, 0.2885, 0.1372, 0.1810, 0.2520], + device='cuda:3'), in_proj_covar=tensor([0.0296, 0.0316, 0.0227, 0.0288, 0.0312, 0.0271, 0.0257, 0.0281], + device='cuda:3'), out_proj_covar=tensor([1.1978e-04, 1.2748e-04, 9.1471e-05, 1.1541e-04, 1.2826e-04, 1.0932e-04, + 1.0520e-04, 1.1276e-04], device='cuda:3') +2023-04-26 22:52:20,761 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-26 22:52:30,771 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49297.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:53:02,579 INFO [finetune.py:976] (3/7) Epoch 9, batch 3500, loss[loss=0.1984, simple_loss=0.2628, pruned_loss=0.06702, over 4823.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2621, pruned_loss=0.06552, over 955791.24 frames. ], batch size: 38, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:53:15,827 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.587e+02 1.893e+02 2.359e+02 3.521e+02, threshold=3.785e+02, percent-clipped=0.0 +2023-04-26 22:53:27,431 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 22:54:00,998 INFO [finetune.py:976] (3/7) Epoch 9, batch 3550, loss[loss=0.1473, simple_loss=0.2089, pruned_loss=0.04286, over 4827.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2586, pruned_loss=0.06442, over 955844.01 frames. ], batch size: 25, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:55:07,552 INFO [finetune.py:976] (3/7) Epoch 9, batch 3600, loss[loss=0.1628, simple_loss=0.2392, pruned_loss=0.04326, over 4760.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2569, pruned_loss=0.0642, over 956756.53 frames. ], batch size: 27, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:55:07,653 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49422.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:55:17,834 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0859, 1.2354, 1.4989, 1.6338, 1.5392, 1.7128, 1.6199, 1.5682], + device='cuda:3'), covar=tensor([0.4920, 0.6563, 0.5511, 0.5261, 0.6340, 0.8897, 0.6122, 0.5556], + device='cuda:3'), in_proj_covar=tensor([0.0321, 0.0383, 0.0314, 0.0324, 0.0340, 0.0401, 0.0361, 0.0323], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 22:55:24,176 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49432.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:55:25,884 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.649e+02 1.885e+02 2.480e+02 5.265e+02, threshold=3.771e+02, percent-clipped=3.0 +2023-04-26 22:56:18,972 INFO [finetune.py:976] (3/7) Epoch 9, batch 3650, loss[loss=0.219, simple_loss=0.288, pruned_loss=0.075, over 4837.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2603, pruned_loss=0.06542, over 956200.46 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:56:22,184 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49477.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:56:30,061 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49483.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:56:35,422 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-26 22:56:36,683 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49493.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 22:56:37,270 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49494.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:56:54,482 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-26 22:56:56,124 INFO [finetune.py:976] (3/7) Epoch 9, batch 3700, loss[loss=0.1998, simple_loss=0.2677, pruned_loss=0.06596, over 4931.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.264, pruned_loss=0.06641, over 953164.44 frames. ], batch size: 38, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:56:58,030 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49525.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:57:09,399 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.745e+02 1.976e+02 2.490e+02 4.358e+02, threshold=3.952e+02, percent-clipped=2.0 +2023-04-26 22:57:13,689 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49542.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:57:56,798 INFO [finetune.py:976] (3/7) Epoch 9, batch 3750, loss[loss=0.2336, simple_loss=0.2994, pruned_loss=0.08384, over 4852.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2659, pruned_loss=0.06708, over 954979.17 frames. ], batch size: 44, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:58:23,846 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49592.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:58:29,005 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4704, 3.4806, 0.9502, 1.9233, 1.8877, 2.3838, 1.8996, 1.0326], + device='cuda:3'), covar=tensor([0.1295, 0.0814, 0.1982, 0.1228, 0.1065, 0.1008, 0.1419, 0.1962], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0249, 0.0140, 0.0122, 0.0135, 0.0153, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 22:58:44,875 INFO [finetune.py:976] (3/7) Epoch 9, batch 3800, loss[loss=0.2253, simple_loss=0.3006, pruned_loss=0.07499, over 4814.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2665, pruned_loss=0.06711, over 955944.32 frames. ], batch size: 38, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:58:52,834 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.711e+02 2.053e+02 2.525e+02 6.309e+02, threshold=4.105e+02, percent-clipped=4.0 +2023-04-26 22:59:05,028 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3046, 1.8694, 1.6206, 2.1026, 2.0436, 2.1157, 1.6243, 4.2773], + device='cuda:3'), covar=tensor([0.0649, 0.0786, 0.0810, 0.1105, 0.0614, 0.0583, 0.0753, 0.0139], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 22:59:11,667 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0397, 1.4891, 1.6252, 1.8520, 2.2706, 1.9290, 1.6121, 1.5709], + device='cuda:3'), covar=tensor([0.1720, 0.1665, 0.1945, 0.1459, 0.0926, 0.1621, 0.2053, 0.1982], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0325, 0.0354, 0.0298, 0.0335, 0.0322, 0.0308, 0.0357], + device='cuda:3'), out_proj_covar=tensor([6.5235e-05, 6.8980e-05, 7.6384e-05, 6.1570e-05, 7.0310e-05, 6.9167e-05, + 6.6312e-05, 7.6580e-05], device='cuda:3') +2023-04-26 22:59:17,908 INFO [finetune.py:976] (3/7) Epoch 9, batch 3850, loss[loss=0.1712, simple_loss=0.2362, pruned_loss=0.0531, over 4719.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2629, pruned_loss=0.06495, over 955405.91 frames. ], batch size: 59, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:59:30,107 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9519, 2.6768, 2.0265, 2.0238, 1.6205, 1.5761, 2.0358, 1.5872], + device='cuda:3'), covar=tensor([0.1331, 0.1296, 0.1355, 0.1805, 0.2087, 0.1816, 0.0984, 0.1828], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0217, 0.0171, 0.0206, 0.0206, 0.0184, 0.0161, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 22:59:49,730 INFO [finetune.py:976] (3/7) Epoch 9, batch 3900, loss[loss=0.1792, simple_loss=0.2399, pruned_loss=0.05923, over 4866.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2589, pruned_loss=0.06328, over 953576.87 frames. ], batch size: 31, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 22:59:58,108 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.654e+02 1.955e+02 2.414e+02 4.265e+02, threshold=3.910e+02, percent-clipped=1.0 +2023-04-26 23:00:04,313 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7228, 1.8751, 0.8206, 1.4131, 1.9715, 1.5810, 1.4625, 1.5328], + device='cuda:3'), covar=tensor([0.0502, 0.0356, 0.0384, 0.0553, 0.0267, 0.0523, 0.0529, 0.0571], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:00:18,064 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49766.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:00:21,570 INFO [finetune.py:976] (3/7) Epoch 9, batch 3950, loss[loss=0.2185, simple_loss=0.271, pruned_loss=0.08296, over 4774.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2553, pruned_loss=0.06234, over 954374.55 frames. ], batch size: 54, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:00:27,206 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49778.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:00:33,310 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49788.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 23:00:55,286 INFO [finetune.py:976] (3/7) Epoch 9, batch 4000, loss[loss=0.1843, simple_loss=0.2432, pruned_loss=0.06272, over 4798.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2562, pruned_loss=0.06308, over 956474.36 frames. ], batch size: 29, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:00:57,208 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1044, 2.0847, 1.7305, 1.6837, 2.1549, 1.7337, 2.6518, 1.5641], + device='cuda:3'), covar=tensor([0.3327, 0.1615, 0.4516, 0.2855, 0.1573, 0.2247, 0.1304, 0.4071], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0349, 0.0432, 0.0362, 0.0389, 0.0384, 0.0381, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:00:57,267 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-26 23:00:59,440 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49827.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:01:05,133 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.602e+02 1.886e+02 2.314e+02 3.383e+02, threshold=3.771e+02, percent-clipped=0.0 +2023-04-26 23:01:41,477 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6797, 2.4755, 1.7602, 1.7395, 1.1953, 1.2705, 1.7585, 1.1787], + device='cuda:3'), covar=tensor([0.1630, 0.1339, 0.1513, 0.1866, 0.2475, 0.2105, 0.1132, 0.2117], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0216, 0.0171, 0.0205, 0.0205, 0.0184, 0.0161, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 23:01:44,304 INFO [finetune.py:976] (3/7) Epoch 9, batch 4050, loss[loss=0.2243, simple_loss=0.2887, pruned_loss=0.07995, over 4867.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2591, pruned_loss=0.06424, over 957840.47 frames. ], batch size: 31, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:02:14,670 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49892.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:02:14,724 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5634, 2.0341, 2.4650, 3.0714, 2.3677, 1.9213, 1.8693, 2.4205], + device='cuda:3'), covar=tensor([0.3833, 0.3669, 0.1660, 0.3066, 0.3124, 0.3018, 0.4439, 0.2605], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0252, 0.0220, 0.0319, 0.0214, 0.0229, 0.0233, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 23:02:49,742 INFO [finetune.py:976] (3/7) Epoch 9, batch 4100, loss[loss=0.2007, simple_loss=0.2634, pruned_loss=0.06897, over 4810.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2621, pruned_loss=0.0654, over 954697.44 frames. ], batch size: 25, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:03:10,346 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.345e+02 1.694e+02 2.037e+02 2.558e+02 4.844e+02, threshold=4.074e+02, percent-clipped=3.0 +2023-04-26 23:03:19,225 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:03:53,586 INFO [finetune.py:976] (3/7) Epoch 9, batch 4150, loss[loss=0.2203, simple_loss=0.2792, pruned_loss=0.08066, over 4162.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2628, pruned_loss=0.06526, over 954881.79 frames. ], batch size: 65, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:04:33,309 INFO [finetune.py:976] (3/7) Epoch 9, batch 4200, loss[loss=0.1652, simple_loss=0.239, pruned_loss=0.04572, over 4787.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2628, pruned_loss=0.06474, over 953094.11 frames. ], batch size: 29, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:04:41,637 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.800e+02 2.150e+02 2.461e+02 7.133e+02, threshold=4.301e+02, percent-clipped=1.0 +2023-04-26 23:04:43,316 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1824, 1.4761, 1.5088, 2.0162, 2.2333, 1.9077, 1.8307, 1.5426], + device='cuda:3'), covar=tensor([0.1873, 0.1861, 0.1741, 0.1839, 0.1296, 0.1799, 0.2186, 0.1880], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0320, 0.0351, 0.0295, 0.0331, 0.0319, 0.0304, 0.0354], + device='cuda:3'), out_proj_covar=tensor([6.4482e-05, 6.7865e-05, 7.5670e-05, 6.0831e-05, 6.9366e-05, 6.8483e-05, + 6.5442e-05, 7.5915e-05], device='cuda:3') +2023-04-26 23:04:53,792 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-26 23:04:56,004 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7895, 1.3103, 4.9023, 4.5434, 4.3112, 4.6145, 4.3182, 4.3925], + device='cuda:3'), covar=tensor([0.6984, 0.6267, 0.0982, 0.1919, 0.1030, 0.1287, 0.1662, 0.1513], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0309, 0.0405, 0.0411, 0.0351, 0.0405, 0.0316, 0.0370], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:05:01,570 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1179, 1.4893, 1.9218, 2.3422, 1.8781, 1.5326, 1.1709, 1.7011], + device='cuda:3'), covar=tensor([0.3909, 0.4262, 0.2037, 0.2905, 0.3267, 0.3266, 0.5111, 0.2745], + device='cuda:3'), in_proj_covar=tensor([0.0281, 0.0250, 0.0219, 0.0317, 0.0213, 0.0227, 0.0232, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 23:05:05,670 INFO [finetune.py:976] (3/7) Epoch 9, batch 4250, loss[loss=0.1696, simple_loss=0.2246, pruned_loss=0.0573, over 4308.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.26, pruned_loss=0.06356, over 950093.50 frames. ], batch size: 18, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:05:09,941 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50078.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:16,026 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50088.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:05:34,628 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8869, 4.0190, 1.1117, 2.1054, 2.2300, 2.7358, 2.3902, 0.9589], + device='cuda:3'), covar=tensor([0.1264, 0.1181, 0.1908, 0.1300, 0.1050, 0.1027, 0.1377, 0.2100], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0251, 0.0141, 0.0123, 0.0136, 0.0154, 0.0118, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:05:37,583 INFO [finetune.py:976] (3/7) Epoch 9, batch 4300, loss[loss=0.2273, simple_loss=0.2847, pruned_loss=0.08489, over 4890.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2581, pruned_loss=0.0633, over 951762.04 frames. ], batch size: 32, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:05:37,652 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50122.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:40,103 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50126.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:41,359 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50128.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:46,442 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.729e+02 1.986e+02 2.496e+02 5.058e+02, threshold=3.971e+02, percent-clipped=3.0 +2023-04-26 23:05:47,121 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50136.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:06:10,817 INFO [finetune.py:976] (3/7) Epoch 9, batch 4350, loss[loss=0.2069, simple_loss=0.2694, pruned_loss=0.07222, over 4260.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2548, pruned_loss=0.0622, over 952166.69 frames. ], batch size: 65, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:06:20,526 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4201, 1.7790, 2.2543, 2.8487, 2.2621, 1.7955, 1.6692, 2.2757], + device='cuda:3'), covar=tensor([0.3690, 0.3786, 0.1678, 0.2867, 0.3269, 0.2955, 0.4633, 0.2518], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0250, 0.0219, 0.0318, 0.0213, 0.0228, 0.0233, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 23:06:22,173 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50189.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:06:27,957 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50197.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:06:41,721 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-26 23:06:44,561 INFO [finetune.py:976] (3/7) Epoch 9, batch 4400, loss[loss=0.2113, simple_loss=0.2774, pruned_loss=0.07267, over 4816.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2582, pruned_loss=0.06447, over 952546.53 frames. ], batch size: 40, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:06:52,519 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.620e+02 1.881e+02 2.281e+02 3.883e+02, threshold=3.762e+02, percent-clipped=0.0 +2023-04-26 23:07:07,379 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-04-26 23:07:14,005 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50258.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:07:27,674 INFO [finetune.py:976] (3/7) Epoch 9, batch 4450, loss[loss=0.1926, simple_loss=0.2576, pruned_loss=0.06378, over 4822.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2624, pruned_loss=0.06607, over 953116.45 frames. ], batch size: 30, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:07:38,429 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 23:08:33,132 INFO [finetune.py:976] (3/7) Epoch 9, batch 4500, loss[loss=0.2047, simple_loss=0.254, pruned_loss=0.07773, over 4398.00 frames. ], tot_loss[loss=0.1988, simple_loss=0.264, pruned_loss=0.06679, over 951491.95 frames. ], batch size: 19, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:08:41,059 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-04-26 23:08:46,790 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.697e+02 2.110e+02 2.509e+02 5.255e+02, threshold=4.219e+02, percent-clipped=3.0 +2023-04-26 23:08:50,593 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5698, 1.8322, 0.8254, 1.3262, 1.9368, 1.4776, 1.4216, 1.4532], + device='cuda:3'), covar=tensor([0.0512, 0.0341, 0.0382, 0.0545, 0.0265, 0.0524, 0.0491, 0.0561], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:09:02,440 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 23:09:12,328 INFO [finetune.py:976] (3/7) Epoch 9, batch 4550, loss[loss=0.1886, simple_loss=0.2597, pruned_loss=0.05876, over 4887.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2651, pruned_loss=0.06667, over 953592.53 frames. ], batch size: 35, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:09:43,720 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0623, 2.5035, 0.8496, 1.5040, 1.4955, 1.9146, 1.5820, 0.8714], + device='cuda:3'), covar=tensor([0.1493, 0.1023, 0.1765, 0.1212, 0.1172, 0.0873, 0.1490, 0.1710], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0252, 0.0142, 0.0123, 0.0136, 0.0155, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:10:20,890 INFO [finetune.py:976] (3/7) Epoch 9, batch 4600, loss[loss=0.1774, simple_loss=0.2447, pruned_loss=0.05508, over 4796.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2639, pruned_loss=0.06599, over 953951.25 frames. ], batch size: 25, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:10:20,988 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50422.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:10:39,536 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.674e+02 1.954e+02 2.272e+02 3.604e+02, threshold=3.908e+02, percent-clipped=0.0 +2023-04-26 23:10:50,480 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0481, 1.5230, 1.8575, 2.1421, 1.8466, 1.4995, 1.0028, 1.5683], + device='cuda:3'), covar=tensor([0.3848, 0.4130, 0.2097, 0.2792, 0.3145, 0.3365, 0.5189, 0.2702], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0252, 0.0220, 0.0319, 0.0214, 0.0229, 0.0233, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 23:11:14,981 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50470.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:11:16,115 INFO [finetune.py:976] (3/7) Epoch 9, batch 4650, loss[loss=0.1586, simple_loss=0.2285, pruned_loss=0.04436, over 4820.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2604, pruned_loss=0.06443, over 954035.24 frames. ], batch size: 30, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:11:23,419 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50484.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:11:36,104 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.92 vs. limit=5.0 +2023-04-26 23:11:49,306 INFO [finetune.py:976] (3/7) Epoch 9, batch 4700, loss[loss=0.2042, simple_loss=0.2593, pruned_loss=0.07451, over 4905.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2572, pruned_loss=0.06346, over 955646.75 frames. ], batch size: 36, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:11:57,186 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.640e+02 1.914e+02 2.343e+02 4.102e+02, threshold=3.828e+02, percent-clipped=1.0 +2023-04-26 23:12:04,088 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50546.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:12:08,323 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50553.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 23:12:22,116 INFO [finetune.py:976] (3/7) Epoch 9, batch 4750, loss[loss=0.2254, simple_loss=0.2934, pruned_loss=0.07873, over 4869.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.256, pruned_loss=0.06309, over 956562.17 frames. ], batch size: 34, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:12:35,402 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 23:12:54,821 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50607.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:13:03,083 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-26 23:13:16,244 INFO [finetune.py:976] (3/7) Epoch 9, batch 4800, loss[loss=0.2232, simple_loss=0.2849, pruned_loss=0.08074, over 4912.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2583, pruned_loss=0.06445, over 955341.48 frames. ], batch size: 36, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:13:17,571 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5464, 1.3640, 0.5311, 1.2410, 1.5400, 1.4163, 1.2893, 1.3071], + device='cuda:3'), covar=tensor([0.0552, 0.0456, 0.0463, 0.0620, 0.0312, 0.0595, 0.0599, 0.0644], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:13:30,153 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.707e+02 2.037e+02 2.400e+02 5.618e+02, threshold=4.074e+02, percent-clipped=3.0 +2023-04-26 23:13:55,575 INFO [finetune.py:976] (3/7) Epoch 9, batch 4850, loss[loss=0.199, simple_loss=0.2629, pruned_loss=0.06756, over 4684.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2606, pruned_loss=0.06482, over 954208.89 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:13:55,754 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 23:14:04,043 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-26 23:14:28,054 INFO [finetune.py:976] (3/7) Epoch 9, batch 4900, loss[loss=0.1955, simple_loss=0.2637, pruned_loss=0.06369, over 4756.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2617, pruned_loss=0.06479, over 953962.96 frames. ], batch size: 26, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:14:36,442 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4815, 2.4818, 2.1272, 2.0859, 2.5235, 2.0172, 3.3708, 1.9249], + device='cuda:3'), covar=tensor([0.3881, 0.2376, 0.4464, 0.3540, 0.1937, 0.2722, 0.1613, 0.4135], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0349, 0.0430, 0.0360, 0.0387, 0.0381, 0.0379, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:14:36,899 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.673e+02 1.942e+02 2.428e+02 3.700e+02, threshold=3.884e+02, percent-clipped=0.0 +2023-04-26 23:14:38,277 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 23:14:41,452 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 23:15:14,091 INFO [finetune.py:976] (3/7) Epoch 9, batch 4950, loss[loss=0.1586, simple_loss=0.2292, pruned_loss=0.04398, over 4896.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2638, pruned_loss=0.06523, over 956349.22 frames. ], batch size: 32, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:15:33,283 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50782.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:15:34,467 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50784.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:15:40,660 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-26 23:15:40,862 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8359, 2.3215, 1.9792, 2.0982, 1.6369, 1.7771, 1.9845, 1.4180], + device='cuda:3'), covar=tensor([0.2462, 0.1608, 0.0930, 0.1478, 0.3571, 0.1525, 0.2096, 0.3001], + device='cuda:3'), in_proj_covar=tensor([0.0296, 0.0316, 0.0227, 0.0289, 0.0316, 0.0272, 0.0258, 0.0280], + device='cuda:3'), out_proj_covar=tensor([1.2010e-04, 1.2745e-04, 9.1249e-05, 1.1573e-04, 1.2931e-04, 1.0958e-04, + 1.0534e-04, 1.1218e-04], device='cuda:3') +2023-04-26 23:15:51,326 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 23:16:02,489 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50808.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:16:11,331 INFO [finetune.py:976] (3/7) Epoch 9, batch 5000, loss[loss=0.1769, simple_loss=0.232, pruned_loss=0.06093, over 4169.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2623, pruned_loss=0.06486, over 957358.14 frames. ], batch size: 18, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:16:19,879 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50832.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:16:21,616 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.711e+02 2.099e+02 2.479e+02 5.783e+02, threshold=4.198e+02, percent-clipped=3.0 +2023-04-26 23:16:26,687 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50843.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:16:31,909 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-26 23:16:32,749 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50853.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:16:42,419 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50869.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:16:44,514 INFO [finetune.py:976] (3/7) Epoch 9, batch 5050, loss[loss=0.18, simple_loss=0.2322, pruned_loss=0.06395, over 4674.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2597, pruned_loss=0.06439, over 958039.03 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:17:04,645 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50901.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:17:05,265 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50902.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:17:05,918 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6806, 1.6778, 0.7102, 1.3705, 1.7855, 1.5448, 1.4484, 1.4829], + device='cuda:3'), covar=tensor([0.0523, 0.0385, 0.0399, 0.0561, 0.0288, 0.0545, 0.0514, 0.0612], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:17:17,317 INFO [finetune.py:976] (3/7) Epoch 9, batch 5100, loss[loss=0.1918, simple_loss=0.2607, pruned_loss=0.06147, over 4935.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2568, pruned_loss=0.06326, over 958350.77 frames. ], batch size: 38, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:17:26,166 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.643e+02 1.891e+02 2.439e+02 4.473e+02, threshold=3.781e+02, percent-clipped=2.0 +2023-04-26 23:17:34,467 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50946.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:17:50,270 INFO [finetune.py:976] (3/7) Epoch 9, batch 5150, loss[loss=0.2135, simple_loss=0.2835, pruned_loss=0.07174, over 4874.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2574, pruned_loss=0.06361, over 958343.28 frames. ], batch size: 34, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:17:50,989 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6298, 1.8964, 0.7757, 1.2586, 1.9536, 1.5339, 1.4215, 1.4158], + device='cuda:3'), covar=tensor([0.0529, 0.0377, 0.0390, 0.0597, 0.0271, 0.0596, 0.0544, 0.0618], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:17:53,616 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-26 23:17:54,549 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7335, 2.4760, 1.8875, 1.6417, 1.2035, 1.2836, 1.9693, 1.2692], + device='cuda:3'), covar=tensor([0.1814, 0.1591, 0.1551, 0.2077, 0.2595, 0.2128, 0.1080, 0.2214], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0217, 0.0172, 0.0205, 0.0205, 0.0184, 0.0161, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 23:17:58,109 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0898, 2.5051, 0.8081, 1.4995, 1.5687, 1.8844, 1.5793, 0.8866], + device='cuda:3'), covar=tensor([0.1443, 0.1189, 0.1820, 0.1383, 0.1087, 0.0910, 0.1511, 0.1670], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0251, 0.0141, 0.0122, 0.0136, 0.0154, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:18:25,479 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51007.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:18:35,218 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8351, 1.3650, 1.6481, 1.6625, 1.6324, 1.3186, 0.7669, 1.3166], + device='cuda:3'), covar=tensor([0.3663, 0.3924, 0.1877, 0.2542, 0.2865, 0.2957, 0.4574, 0.2454], + device='cuda:3'), in_proj_covar=tensor([0.0280, 0.0248, 0.0218, 0.0315, 0.0212, 0.0227, 0.0231, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-26 23:18:40,595 INFO [finetune.py:976] (3/7) Epoch 9, batch 5200, loss[loss=0.1826, simple_loss=0.2516, pruned_loss=0.05678, over 4901.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2618, pruned_loss=0.06497, over 958347.74 frames. ], batch size: 35, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:18:49,053 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.700e+02 2.036e+02 2.415e+02 4.035e+02, threshold=4.072e+02, percent-clipped=2.0 +2023-04-26 23:18:52,933 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9712, 2.2937, 1.0073, 1.2953, 1.7609, 1.2425, 3.1404, 1.6588], + device='cuda:3'), covar=tensor([0.0689, 0.0751, 0.0803, 0.1268, 0.0520, 0.0986, 0.0202, 0.0614], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0078, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-26 23:19:08,578 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51062.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:19:14,485 INFO [finetune.py:976] (3/7) Epoch 9, batch 5250, loss[loss=0.2003, simple_loss=0.2734, pruned_loss=0.06358, over 4895.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2638, pruned_loss=0.06527, over 956633.93 frames. ], batch size: 43, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:19:47,750 INFO [finetune.py:976] (3/7) Epoch 9, batch 5300, loss[loss=0.1609, simple_loss=0.236, pruned_loss=0.04289, over 4747.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2649, pruned_loss=0.06563, over 956420.59 frames. ], batch size: 27, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:19:48,484 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51123.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:19:56,068 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.811e+02 2.045e+02 2.583e+02 4.950e+02, threshold=4.090e+02, percent-clipped=1.0 +2023-04-26 23:19:57,961 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51138.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:20:04,602 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1299, 2.6319, 2.2545, 2.4880, 1.7450, 2.1109, 2.2967, 1.7313], + device='cuda:3'), covar=tensor([0.1838, 0.1168, 0.0850, 0.1166, 0.3426, 0.1304, 0.1726, 0.2535], + device='cuda:3'), in_proj_covar=tensor([0.0296, 0.0316, 0.0226, 0.0288, 0.0316, 0.0272, 0.0258, 0.0279], + device='cuda:3'), out_proj_covar=tensor([1.1979e-04, 1.2758e-04, 9.0916e-05, 1.1576e-04, 1.2952e-04, 1.0972e-04, + 1.0521e-04, 1.1185e-04], device='cuda:3') +2023-04-26 23:20:15,867 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51164.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:20:20,663 INFO [finetune.py:976] (3/7) Epoch 9, batch 5350, loss[loss=0.169, simple_loss=0.2335, pruned_loss=0.05227, over 4921.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2645, pruned_loss=0.06535, over 954171.64 frames. ], batch size: 33, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:20:20,771 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51172.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:20:21,942 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1692, 1.5747, 1.4039, 1.6810, 1.6091, 1.9069, 1.3607, 3.3924], + device='cuda:3'), covar=tensor([0.0606, 0.0711, 0.0726, 0.1120, 0.0573, 0.0549, 0.0732, 0.0133], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 23:20:22,858 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-04-26 23:20:46,714 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51202.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:21:10,259 INFO [finetune.py:976] (3/7) Epoch 9, batch 5400, loss[loss=0.2037, simple_loss=0.2611, pruned_loss=0.07316, over 4816.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2601, pruned_loss=0.06394, over 953537.73 frames. ], batch size: 33, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:21:21,597 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51233.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:21:22,681 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.244e+02 1.662e+02 1.922e+02 2.270e+02 4.708e+02, threshold=3.844e+02, percent-clipped=3.0 +2023-04-26 23:21:44,190 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51250.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:22:14,587 INFO [finetune.py:976] (3/7) Epoch 9, batch 5450, loss[loss=0.1306, simple_loss=0.2067, pruned_loss=0.02721, over 4753.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2564, pruned_loss=0.0624, over 955179.17 frames. ], batch size: 27, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:22:47,580 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51302.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:22:59,654 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-04-26 23:23:19,200 INFO [finetune.py:976] (3/7) Epoch 9, batch 5500, loss[loss=0.1516, simple_loss=0.2237, pruned_loss=0.03978, over 4795.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.253, pruned_loss=0.06106, over 956656.56 frames. ], batch size: 29, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:23:32,910 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.202e+01 1.588e+02 1.902e+02 2.243e+02 3.887e+02, threshold=3.804e+02, percent-clipped=1.0 +2023-04-26 23:24:10,592 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9449, 1.6197, 1.9022, 2.3152, 2.3474, 1.8207, 1.5602, 1.8576], + device='cuda:3'), covar=tensor([0.0809, 0.1071, 0.0599, 0.0526, 0.0487, 0.0794, 0.0795, 0.0640], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0204, 0.0181, 0.0175, 0.0176, 0.0187, 0.0158, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:24:14,746 INFO [finetune.py:976] (3/7) Epoch 9, batch 5550, loss[loss=0.1827, simple_loss=0.2656, pruned_loss=0.04983, over 4823.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2567, pruned_loss=0.06298, over 953689.37 frames. ], batch size: 40, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:24:20,977 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51382.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:24:26,982 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51392.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:24:30,704 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3604, 1.7021, 1.7183, 1.8241, 1.6196, 1.7286, 1.8485, 1.7284], + device='cuda:3'), covar=tensor([0.5270, 0.7647, 0.6089, 0.5712, 0.7304, 0.9912, 0.7586, 0.6642], + device='cuda:3'), in_proj_covar=tensor([0.0323, 0.0382, 0.0315, 0.0326, 0.0337, 0.0399, 0.0359, 0.0322], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 23:24:40,605 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8726, 4.3703, 0.9285, 2.2623, 2.3035, 2.9015, 2.4944, 0.9505], + device='cuda:3'), covar=tensor([0.1419, 0.0881, 0.2197, 0.1281, 0.1137, 0.1107, 0.1472, 0.2304], + device='cuda:3'), in_proj_covar=tensor([0.0119, 0.0253, 0.0142, 0.0123, 0.0138, 0.0156, 0.0120, 0.0124], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:24:42,928 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51418.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:24:45,249 INFO [finetune.py:976] (3/7) Epoch 9, batch 5600, loss[loss=0.2269, simple_loss=0.2986, pruned_loss=0.07757, over 4797.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2604, pruned_loss=0.06402, over 953710.84 frames. ], batch size: 51, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:24:52,213 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6393, 1.4549, 0.5930, 1.3143, 1.6032, 1.4665, 1.3724, 1.4030], + device='cuda:3'), covar=tensor([0.0536, 0.0403, 0.0459, 0.0595, 0.0305, 0.0575, 0.0592, 0.0591], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:3') +2023-04-26 23:24:52,680 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.747e+02 2.120e+02 2.551e+02 6.497e+02, threshold=4.239e+02, percent-clipped=3.0 +2023-04-26 23:24:54,539 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51438.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:24:57,454 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51443.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:03,230 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51453.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:10,023 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51464.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:25:15,160 INFO [finetune.py:976] (3/7) Epoch 9, batch 5650, loss[loss=0.1928, simple_loss=0.2589, pruned_loss=0.06338, over 4869.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2624, pruned_loss=0.0641, over 952321.91 frames. ], batch size: 44, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:25:23,728 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51486.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:29,032 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51495.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:39,095 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51512.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:25:45,276 INFO [finetune.py:976] (3/7) Epoch 9, batch 5700, loss[loss=0.1456, simple_loss=0.2173, pruned_loss=0.03694, over 4244.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2579, pruned_loss=0.0638, over 932628.73 frames. ], batch size: 18, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:25:47,124 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0393, 3.8606, 1.2477, 2.1903, 2.4309, 2.8387, 2.2362, 1.2673], + device='cuda:3'), covar=tensor([0.1242, 0.0983, 0.1892, 0.1261, 0.0969, 0.0932, 0.1472, 0.1826], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0251, 0.0142, 0.0122, 0.0136, 0.0155, 0.0119, 0.0123], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:25:48,922 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51528.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:25:53,107 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.631e+02 1.982e+02 2.330e+02 4.156e+02, threshold=3.963e+02, percent-clipped=0.0 +2023-04-26 23:26:16,105 INFO [finetune.py:976] (3/7) Epoch 10, batch 0, loss[loss=0.2, simple_loss=0.274, pruned_loss=0.06303, over 4878.00 frames. ], tot_loss[loss=0.2, simple_loss=0.274, pruned_loss=0.06303, over 4878.00 frames. ], batch size: 43, lr: 3.76e-03, grad_scale: 32.0 +2023-04-26 23:26:16,105 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-26 23:26:31,821 INFO [finetune.py:1010] (3/7) Epoch 10, validation: loss=0.1558, simple_loss=0.2282, pruned_loss=0.04164, over 2265189.00 frames. +2023-04-26 23:26:31,821 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-26 23:26:37,637 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51556.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:26:41,615 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-26 23:27:05,702 INFO [finetune.py:976] (3/7) Epoch 10, batch 50, loss[loss=0.1819, simple_loss=0.2515, pruned_loss=0.05612, over 4846.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.26, pruned_loss=0.06157, over 216091.00 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 32.0 +2023-04-26 23:27:08,575 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51602.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:27:23,689 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6454, 2.2188, 2.5710, 3.0464, 2.8128, 2.1548, 1.9613, 2.4688], + device='cuda:3'), covar=tensor([0.0824, 0.0975, 0.0527, 0.0558, 0.0621, 0.0945, 0.0890, 0.0644], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0202, 0.0181, 0.0173, 0.0175, 0.0186, 0.0158, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:27:31,934 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.709e+02 2.017e+02 2.493e+02 1.011e+03, threshold=4.035e+02, percent-clipped=6.0 +2023-04-26 23:27:45,647 INFO [finetune.py:976] (3/7) Epoch 10, batch 100, loss[loss=0.1738, simple_loss=0.2397, pruned_loss=0.054, over 4727.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2559, pruned_loss=0.06222, over 379427.96 frames. ], batch size: 23, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:27:47,277 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51650.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:27:57,611 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4221, 3.1921, 0.9450, 1.8555, 1.9633, 2.3950, 1.9392, 1.0090], + device='cuda:3'), covar=tensor([0.1419, 0.0918, 0.2049, 0.1217, 0.1005, 0.0932, 0.1316, 0.2000], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0248, 0.0140, 0.0121, 0.0135, 0.0153, 0.0118, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:28:21,328 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6277, 3.6032, 1.1363, 2.0853, 2.0150, 2.5410, 2.0241, 0.9668], + device='cuda:3'), covar=tensor([0.1302, 0.0846, 0.1878, 0.1172, 0.1022, 0.1030, 0.1451, 0.2132], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0249, 0.0141, 0.0122, 0.0135, 0.0154, 0.0118, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:28:38,560 INFO [finetune.py:976] (3/7) Epoch 10, batch 150, loss[loss=0.2128, simple_loss=0.2596, pruned_loss=0.08296, over 4802.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2535, pruned_loss=0.06256, over 509191.50 frames. ], batch size: 45, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:29:02,032 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8991, 2.7604, 2.1784, 3.3295, 2.8578, 2.8528, 1.1576, 2.7686], + device='cuda:3'), covar=tensor([0.2032, 0.1945, 0.3505, 0.2750, 0.3244, 0.2311, 0.6058, 0.3038], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0215, 0.0248, 0.0301, 0.0298, 0.0248, 0.0268, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 23:29:04,459 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51718.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:12,322 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-04-26 23:29:16,235 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0847, 2.8291, 2.0700, 2.0988, 1.5026, 1.4492, 2.1661, 1.4800], + device='cuda:3'), covar=tensor([0.1932, 0.1745, 0.1686, 0.2117, 0.2864, 0.2406, 0.1175, 0.2373], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0216, 0.0171, 0.0205, 0.0205, 0.0184, 0.0160, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 23:29:20,965 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.275e+02 1.676e+02 2.023e+02 2.458e+02 4.768e+02, threshold=4.046e+02, percent-clipped=1.0 +2023-04-26 23:29:22,219 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51738.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:29,233 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51747.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:29,800 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51748.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:30,341 INFO [finetune.py:976] (3/7) Epoch 10, batch 200, loss[loss=0.1797, simple_loss=0.238, pruned_loss=0.06065, over 4693.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2528, pruned_loss=0.06336, over 608863.72 frames. ], batch size: 23, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:29:42,681 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51766.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:30:04,063 INFO [finetune.py:976] (3/7) Epoch 10, batch 250, loss[loss=0.1941, simple_loss=0.2545, pruned_loss=0.06682, over 4890.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2544, pruned_loss=0.06333, over 686788.66 frames. ], batch size: 35, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:30:11,168 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51808.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:30:23,870 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51828.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:30:28,669 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.252e+02 1.711e+02 1.990e+02 2.532e+02 5.377e+02, threshold=3.981e+02, percent-clipped=2.0 +2023-04-26 23:30:37,590 INFO [finetune.py:976] (3/7) Epoch 10, batch 300, loss[loss=0.2242, simple_loss=0.2887, pruned_loss=0.07983, over 4859.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2589, pruned_loss=0.06448, over 746030.84 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:30:39,369 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51851.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:30:44,661 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7225, 1.1865, 1.7542, 2.1549, 1.8383, 1.7369, 1.7302, 1.7853], + device='cuda:3'), covar=tensor([0.5664, 0.7903, 0.8205, 0.8041, 0.7087, 0.9338, 0.9481, 0.7852], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0416, 0.0502, 0.0520, 0.0437, 0.0456, 0.0467, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.9536e-05, 1.0319e-04, 1.1320e-04, 1.2373e-04, 1.0616e-04, 1.1038e-04, + 1.1219e-04, 1.1223e-04], device='cuda:3') +2023-04-26 23:30:56,480 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51876.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:31:10,832 INFO [finetune.py:976] (3/7) Epoch 10, batch 350, loss[loss=0.2362, simple_loss=0.2936, pruned_loss=0.08941, over 4790.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2614, pruned_loss=0.06496, over 794516.23 frames. ], batch size: 51, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:31:26,186 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-04-26 23:31:41,425 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.954e+01 1.688e+02 1.992e+02 2.453e+02 5.822e+02, threshold=3.984e+02, percent-clipped=3.0 +2023-04-26 23:32:00,718 INFO [finetune.py:976] (3/7) Epoch 10, batch 400, loss[loss=0.1818, simple_loss=0.248, pruned_loss=0.0578, over 4747.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2613, pruned_loss=0.06382, over 830387.96 frames. ], batch size: 26, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:32:15,700 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-26 23:32:17,899 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6166, 3.6935, 0.8544, 1.8818, 2.0617, 2.5788, 2.0088, 0.9191], + device='cuda:3'), covar=tensor([0.1417, 0.0949, 0.2301, 0.1365, 0.1109, 0.0998, 0.1520, 0.2114], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0250, 0.0141, 0.0122, 0.0136, 0.0155, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:32:40,865 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-04-26 23:32:51,604 INFO [finetune.py:976] (3/7) Epoch 10, batch 450, loss[loss=0.1834, simple_loss=0.2433, pruned_loss=0.06168, over 4929.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2604, pruned_loss=0.06393, over 859746.12 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:33:31,183 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-04-26 23:33:45,356 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.213e+02 1.582e+02 1.952e+02 2.304e+02 4.077e+02, threshold=3.905e+02, percent-clipped=1.0 +2023-04-26 23:33:46,677 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52038.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:33:53,302 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52040.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:33:58,100 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52048.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:34:04,124 INFO [finetune.py:976] (3/7) Epoch 10, batch 500, loss[loss=0.1887, simple_loss=0.2625, pruned_loss=0.05743, over 4827.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2573, pruned_loss=0.0629, over 878394.58 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:34:51,736 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52086.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:09,317 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52096.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:11,182 INFO [finetune.py:976] (3/7) Epoch 10, batch 550, loss[loss=0.1677, simple_loss=0.2352, pruned_loss=0.05006, over 4909.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2541, pruned_loss=0.06204, over 894528.84 frames. ], batch size: 36, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:35:12,502 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52101.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:13,634 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52103.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:22,471 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7964, 2.0880, 1.9933, 2.1653, 1.8780, 2.0068, 2.0264, 1.9875], + device='cuda:3'), covar=tensor([0.5100, 0.7807, 0.6315, 0.5503, 0.7156, 0.9550, 0.7984, 0.6879], + device='cuda:3'), in_proj_covar=tensor([0.0327, 0.0385, 0.0318, 0.0328, 0.0342, 0.0405, 0.0362, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 23:36:05,087 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.236e+02 1.854e+02 2.167e+02 2.503e+02 5.657e+02, threshold=4.334e+02, percent-clipped=4.0 +2023-04-26 23:36:18,686 INFO [finetune.py:976] (3/7) Epoch 10, batch 600, loss[loss=0.1711, simple_loss=0.2394, pruned_loss=0.05145, over 4906.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2547, pruned_loss=0.06221, over 909068.03 frames. ], batch size: 36, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:36:19,993 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52151.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:37:10,897 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7652, 1.9373, 1.8726, 2.0931, 1.8314, 2.0296, 1.9940, 1.8743], + device='cuda:3'), covar=tensor([0.4658, 0.7711, 0.6536, 0.5249, 0.7148, 0.9175, 0.7683, 0.7028], + device='cuda:3'), in_proj_covar=tensor([0.0326, 0.0384, 0.0317, 0.0327, 0.0340, 0.0404, 0.0361, 0.0325], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-26 23:37:12,639 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5600, 1.4507, 0.6446, 1.2461, 1.6084, 1.4518, 1.3336, 1.3586], + device='cuda:3'), covar=tensor([0.0514, 0.0391, 0.0398, 0.0593, 0.0288, 0.0553, 0.0507, 0.0577], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:37:16,286 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52193.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:37:25,227 INFO [finetune.py:976] (3/7) Epoch 10, batch 650, loss[loss=0.2093, simple_loss=0.2903, pruned_loss=0.06412, over 4900.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2585, pruned_loss=0.06315, over 917578.65 frames. ], batch size: 37, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:37:25,288 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52199.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:38:17,627 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.623e+02 1.914e+02 2.334e+02 8.066e+02, threshold=3.828e+02, percent-clipped=2.0 +2023-04-26 23:38:18,947 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3168, 1.8019, 1.5429, 2.1772, 1.9917, 2.0838, 1.6600, 4.6360], + device='cuda:3'), covar=tensor([0.0614, 0.0778, 0.0849, 0.1220, 0.0654, 0.0560, 0.0797, 0.0096], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 23:38:30,842 INFO [finetune.py:976] (3/7) Epoch 10, batch 700, loss[loss=0.1954, simple_loss=0.2707, pruned_loss=0.0601, over 4834.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2595, pruned_loss=0.06254, over 927387.09 frames. ], batch size: 49, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:38:38,225 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 23:38:39,792 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52254.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:39:44,471 INFO [finetune.py:976] (3/7) Epoch 10, batch 750, loss[loss=0.2249, simple_loss=0.2864, pruned_loss=0.08172, over 4755.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.262, pruned_loss=0.06369, over 935806.05 frames. ], batch size: 27, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:39:56,800 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52312.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:40:13,564 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.731e+02 1.996e+02 2.532e+02 4.524e+02, threshold=3.992e+02, percent-clipped=2.0 +2023-04-26 23:40:22,484 INFO [finetune.py:976] (3/7) Epoch 10, batch 800, loss[loss=0.1784, simple_loss=0.2494, pruned_loss=0.05374, over 4671.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2617, pruned_loss=0.06325, over 938561.10 frames. ], batch size: 23, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:40:29,474 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8735, 2.6002, 2.0257, 1.7483, 1.3463, 1.4173, 1.9325, 1.2601], + device='cuda:3'), covar=tensor([0.1819, 0.1371, 0.1497, 0.1962, 0.2648, 0.2128, 0.1178, 0.2290], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0218, 0.0172, 0.0206, 0.0207, 0.0185, 0.0162, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 23:40:31,340 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9097, 1.6970, 1.9709, 2.3863, 2.3589, 1.8601, 1.5657, 1.9704], + device='cuda:3'), covar=tensor([0.0986, 0.1194, 0.0684, 0.0555, 0.0670, 0.0900, 0.0914, 0.0684], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0203, 0.0180, 0.0173, 0.0175, 0.0186, 0.0158, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:40:37,974 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52373.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:40:50,969 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-26 23:40:54,333 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52396.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:40:54,470 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-26 23:40:56,135 INFO [finetune.py:976] (3/7) Epoch 10, batch 850, loss[loss=0.1838, simple_loss=0.2502, pruned_loss=0.05873, over 4739.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2589, pruned_loss=0.06222, over 942976.21 frames. ], batch size: 54, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:40:58,777 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52403.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:41:30,360 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5923, 1.3794, 4.3601, 4.1022, 3.7621, 4.1104, 4.0503, 3.7636], + device='cuda:3'), covar=tensor([0.6878, 0.6043, 0.0973, 0.1701, 0.1154, 0.1491, 0.1338, 0.1619], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0308, 0.0407, 0.0410, 0.0349, 0.0405, 0.0314, 0.0371], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:41:30,882 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.587e+02 1.831e+02 2.127e+02 6.660e+02, threshold=3.662e+02, percent-clipped=3.0 +2023-04-26 23:41:33,555 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 23:41:33,603 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-26 23:41:41,223 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 23:41:44,733 INFO [finetune.py:976] (3/7) Epoch 10, batch 900, loss[loss=0.172, simple_loss=0.2385, pruned_loss=0.05277, over 4851.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2562, pruned_loss=0.06132, over 946650.19 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:41:51,264 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52451.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:42:46,263 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8142, 2.5765, 1.9015, 1.7904, 1.3136, 1.3479, 1.8955, 1.2628], + device='cuda:3'), covar=tensor([0.1705, 0.1309, 0.1407, 0.1784, 0.2477, 0.2005, 0.1072, 0.2105], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0216, 0.0171, 0.0205, 0.0206, 0.0184, 0.0161, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 23:42:49,242 INFO [finetune.py:976] (3/7) Epoch 10, batch 950, loss[loss=0.1581, simple_loss=0.2279, pruned_loss=0.04417, over 4828.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2547, pruned_loss=0.06143, over 948220.12 frames. ], batch size: 30, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:43:41,296 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.114e+02 1.638e+02 1.935e+02 2.436e+02 4.596e+02, threshold=3.871e+02, percent-clipped=3.0 +2023-04-26 23:43:55,594 INFO [finetune.py:976] (3/7) Epoch 10, batch 1000, loss[loss=0.2187, simple_loss=0.2884, pruned_loss=0.07449, over 4817.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2572, pruned_loss=0.06214, over 947445.13 frames. ], batch size: 47, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:44:01,403 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52549.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:44:05,058 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52555.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:45:07,087 INFO [finetune.py:976] (3/7) Epoch 10, batch 1050, loss[loss=0.204, simple_loss=0.2753, pruned_loss=0.06631, over 4919.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2603, pruned_loss=0.06297, over 949167.47 frames. ], batch size: 42, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:45:11,802 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.73 vs. limit=5.0 +2023-04-26 23:45:29,121 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52616.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:45:52,967 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.267e+02 1.739e+02 2.075e+02 2.346e+02 3.891e+02, threshold=4.149e+02, percent-clipped=1.0 +2023-04-26 23:46:13,535 INFO [finetune.py:976] (3/7) Epoch 10, batch 1100, loss[loss=0.2108, simple_loss=0.2722, pruned_loss=0.07474, over 4901.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2619, pruned_loss=0.06359, over 951892.10 frames. ], batch size: 37, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:46:36,554 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52668.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:46:56,132 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52696.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:46:57,845 INFO [finetune.py:976] (3/7) Epoch 10, batch 1150, loss[loss=0.173, simple_loss=0.2518, pruned_loss=0.04708, over 4905.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2609, pruned_loss=0.06322, over 951992.12 frames. ], batch size: 37, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:47:02,613 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9924, 2.5339, 2.0909, 2.3867, 1.8652, 2.0408, 2.0606, 1.6419], + device='cuda:3'), covar=tensor([0.2246, 0.1463, 0.0853, 0.1304, 0.3219, 0.1433, 0.2206, 0.2966], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0314, 0.0225, 0.0286, 0.0314, 0.0270, 0.0255, 0.0277], + device='cuda:3'), out_proj_covar=tensor([1.1825e-04, 1.2638e-04, 9.0611e-05, 1.1476e-04, 1.2870e-04, 1.0883e-04, + 1.0421e-04, 1.1129e-04], device='cuda:3') +2023-04-26 23:47:03,250 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5386, 1.6924, 1.4519, 1.0699, 1.1643, 1.1425, 1.4247, 1.0904], + device='cuda:3'), covar=tensor([0.1911, 0.1476, 0.1725, 0.1957, 0.2716, 0.2280, 0.1234, 0.2284], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0216, 0.0170, 0.0204, 0.0205, 0.0184, 0.0160, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 23:47:21,726 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.582e+02 1.906e+02 2.303e+02 4.186e+02, threshold=3.812e+02, percent-clipped=1.0 +2023-04-26 23:47:33,675 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52744.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:47:43,259 INFO [finetune.py:976] (3/7) Epoch 10, batch 1200, loss[loss=0.1523, simple_loss=0.2247, pruned_loss=0.04001, over 4641.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2591, pruned_loss=0.06246, over 951617.93 frames. ], batch size: 20, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:47:45,159 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52752.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:48:25,365 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 23:48:48,804 INFO [finetune.py:976] (3/7) Epoch 10, batch 1250, loss[loss=0.2198, simple_loss=0.265, pruned_loss=0.08733, over 4730.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2573, pruned_loss=0.06242, over 952440.21 frames. ], batch size: 59, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:49:09,940 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2907, 1.2571, 1.4210, 1.6500, 1.6995, 1.2994, 1.0386, 1.4185], + device='cuda:3'), covar=tensor([0.0827, 0.1299, 0.0811, 0.0553, 0.0641, 0.0864, 0.0846, 0.0670], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0204, 0.0182, 0.0176, 0.0178, 0.0189, 0.0160, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:49:09,962 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52813.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:49:30,457 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8590, 1.6082, 1.8944, 2.2239, 2.2423, 1.8014, 1.4559, 1.9030], + device='cuda:3'), covar=tensor([0.0800, 0.1171, 0.0782, 0.0648, 0.0587, 0.0811, 0.0896, 0.0645], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0205, 0.0183, 0.0177, 0.0179, 0.0190, 0.0161, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:49:34,633 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.609e+02 1.926e+02 2.217e+02 3.154e+02, threshold=3.851e+02, percent-clipped=0.0 +2023-04-26 23:49:48,758 INFO [finetune.py:976] (3/7) Epoch 10, batch 1300, loss[loss=0.1898, simple_loss=0.2568, pruned_loss=0.06139, over 4834.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2543, pruned_loss=0.06123, over 951034.34 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:49:48,861 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:50:20,541 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52897.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:50:21,745 INFO [finetune.py:976] (3/7) Epoch 10, batch 1350, loss[loss=0.202, simple_loss=0.2716, pruned_loss=0.06621, over 4849.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2557, pruned_loss=0.06213, over 952954.07 frames. ], batch size: 44, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:50:31,127 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52911.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:50:41,047 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-26 23:50:43,112 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9331, 1.6395, 3.9690, 3.7620, 3.5416, 3.5174, 3.4548, 3.5734], + device='cuda:3'), covar=tensor([0.6083, 0.4909, 0.1049, 0.1578, 0.1133, 0.2054, 0.4697, 0.1392], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0304, 0.0403, 0.0408, 0.0347, 0.0403, 0.0313, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:50:44,389 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6911, 2.2926, 1.6343, 1.4166, 1.2230, 1.2426, 1.7496, 1.1458], + device='cuda:3'), covar=tensor([0.1960, 0.1456, 0.1688, 0.2153, 0.2770, 0.2229, 0.1218, 0.2349], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0218, 0.0172, 0.0206, 0.0206, 0.0186, 0.0162, 0.0190], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-26 23:50:46,658 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.765e+02 2.062e+02 2.587e+02 4.153e+02, threshold=4.125e+02, percent-clipped=1.0 +2023-04-26 23:50:49,935 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 23:50:55,066 INFO [finetune.py:976] (3/7) Epoch 10, batch 1400, loss[loss=0.1546, simple_loss=0.2331, pruned_loss=0.03804, over 4850.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.259, pruned_loss=0.06306, over 951834.76 frames. ], batch size: 49, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:51:01,369 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6428, 1.7703, 0.8629, 1.3607, 2.0690, 1.5272, 1.5039, 1.5176], + device='cuda:3'), covar=tensor([0.0497, 0.0382, 0.0357, 0.0551, 0.0259, 0.0538, 0.0512, 0.0570], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:51:09,494 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52968.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:51:13,198 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0114, 1.3872, 1.2778, 1.6209, 1.4978, 1.6469, 1.3436, 2.4386], + device='cuda:3'), covar=tensor([0.0666, 0.0843, 0.0863, 0.1271, 0.0672, 0.0460, 0.0781, 0.0218], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 23:51:13,843 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4546, 1.0588, 0.4548, 1.1369, 1.1664, 1.3358, 1.2310, 1.1918], + device='cuda:3'), covar=tensor([0.0509, 0.0422, 0.0415, 0.0593, 0.0289, 0.0537, 0.0513, 0.0583], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-26 23:51:28,729 INFO [finetune.py:976] (3/7) Epoch 10, batch 1450, loss[loss=0.1485, simple_loss=0.2269, pruned_loss=0.035, over 4757.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2606, pruned_loss=0.06351, over 952272.79 frames. ], batch size: 26, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:51:41,184 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53016.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:52:04,980 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.250e+02 1.692e+02 2.005e+02 2.440e+02 4.945e+02, threshold=4.009e+02, percent-clipped=1.0 +2023-04-26 23:52:18,924 INFO [finetune.py:976] (3/7) Epoch 10, batch 1500, loss[loss=0.2336, simple_loss=0.2987, pruned_loss=0.0843, over 4874.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2614, pruned_loss=0.06404, over 953070.27 frames. ], batch size: 34, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:52:28,786 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53056.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:53:19,946 INFO [finetune.py:976] (3/7) Epoch 10, batch 1550, loss[loss=0.1586, simple_loss=0.2385, pruned_loss=0.03935, over 4813.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2609, pruned_loss=0.0637, over 950456.45 frames. ], batch size: 40, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:53:25,473 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53108.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:53:29,413 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 23:53:32,450 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53117.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:53:45,206 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.713e+02 2.006e+02 2.452e+02 5.414e+02, threshold=4.013e+02, percent-clipped=2.0 +2023-04-26 23:53:53,134 INFO [finetune.py:976] (3/7) Epoch 10, batch 1600, loss[loss=0.1543, simple_loss=0.2151, pruned_loss=0.04674, over 4905.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2576, pruned_loss=0.06287, over 950325.53 frames. ], batch size: 46, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:54:33,157 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.69 vs. limit=5.0 +2023-04-26 23:54:44,154 INFO [finetune.py:976] (3/7) Epoch 10, batch 1650, loss[loss=0.1841, simple_loss=0.2404, pruned_loss=0.06387, over 4922.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2552, pruned_loss=0.06182, over 952574.83 frames. ], batch size: 37, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:54:52,156 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53211.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:55:09,623 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.636e+02 1.864e+02 2.338e+02 5.656e+02, threshold=3.728e+02, percent-clipped=1.0 +2023-04-26 23:55:12,849 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53241.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:55:17,567 INFO [finetune.py:976] (3/7) Epoch 10, batch 1700, loss[loss=0.2177, simple_loss=0.2774, pruned_loss=0.07897, over 4854.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.254, pruned_loss=0.06152, over 953679.63 frames. ], batch size: 49, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:55:23,649 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 23:55:24,104 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53259.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:55:51,481 INFO [finetune.py:976] (3/7) Epoch 10, batch 1750, loss[loss=0.224, simple_loss=0.2827, pruned_loss=0.08269, over 4862.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2547, pruned_loss=0.06199, over 953812.48 frames. ], batch size: 44, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:55:53,417 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53302.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:55:54,614 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53304.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:01,784 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53315.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:16,877 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.803e+02 2.178e+02 2.575e+02 5.782e+02, threshold=4.356e+02, percent-clipped=6.0 +2023-04-26 23:56:19,407 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1128, 0.7565, 0.9494, 0.7149, 1.2268, 0.9850, 0.8571, 1.0197], + device='cuda:3'), covar=tensor([0.1373, 0.1470, 0.2057, 0.1584, 0.0961, 0.1466, 0.1708, 0.1999], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0321, 0.0355, 0.0298, 0.0336, 0.0321, 0.0307, 0.0359], + device='cuda:3'), out_proj_covar=tensor([6.4662e-05, 6.8055e-05, 7.6727e-05, 6.1456e-05, 7.0334e-05, 6.8882e-05, + 6.5813e-05, 7.7032e-05], device='cuda:3') +2023-04-26 23:56:25,433 INFO [finetune.py:976] (3/7) Epoch 10, batch 1800, loss[loss=0.2212, simple_loss=0.2952, pruned_loss=0.07363, over 4764.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2588, pruned_loss=0.06256, over 953229.54 frames. ], batch size: 54, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:56:29,739 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6047, 1.4298, 4.6084, 4.2633, 4.0667, 4.3852, 4.2980, 4.0560], + device='cuda:3'), covar=tensor([0.7021, 0.5949, 0.1018, 0.1956, 0.1103, 0.1794, 0.1052, 0.1626], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0307, 0.0406, 0.0410, 0.0349, 0.0404, 0.0314, 0.0371], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-26 23:56:35,685 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53365.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:36,917 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53367.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:43,376 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53376.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:59,051 INFO [finetune.py:976] (3/7) Epoch 10, batch 1850, loss[loss=0.2134, simple_loss=0.2819, pruned_loss=0.07243, over 4811.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.26, pruned_loss=0.06317, over 953619.28 frames. ], batch size: 40, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:57:10,977 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53408.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:57:12,821 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4698, 1.2905, 1.8408, 1.7889, 1.3442, 1.0851, 1.4805, 1.1331], + device='cuda:3'), covar=tensor([0.0667, 0.0818, 0.0503, 0.0738, 0.0809, 0.1299, 0.0685, 0.0826], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0072, 0.0070, 0.0067, 0.0075, 0.0095, 0.0077, 0.0073], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-26 23:57:13,867 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53412.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:57:35,489 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53428.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:57:45,405 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.670e+02 2.068e+02 2.536e+02 6.143e+02, threshold=4.136e+02, percent-clipped=4.0 +2023-04-26 23:58:05,921 INFO [finetune.py:976] (3/7) Epoch 10, batch 1900, loss[loss=0.1915, simple_loss=0.2594, pruned_loss=0.06184, over 4755.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2604, pruned_loss=0.06307, over 954134.52 frames. ], batch size: 27, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:58:15,434 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53456.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:58:48,691 INFO [finetune.py:976] (3/7) Epoch 10, batch 1950, loss[loss=0.1563, simple_loss=0.2169, pruned_loss=0.04785, over 4826.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2594, pruned_loss=0.06242, over 957073.08 frames. ], batch size: 38, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:58:55,499 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0546, 1.3489, 1.2782, 1.6345, 1.4164, 1.5032, 1.3242, 2.4977], + device='cuda:3'), covar=tensor([0.0634, 0.0861, 0.0815, 0.1249, 0.0683, 0.0486, 0.0763, 0.0219], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-26 23:58:58,681 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.18 vs. limit=5.0 +2023-04-26 23:59:12,237 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.213e+02 1.722e+02 1.968e+02 2.257e+02 3.746e+02, threshold=3.936e+02, percent-clipped=0.0 +2023-04-26 23:59:22,129 INFO [finetune.py:976] (3/7) Epoch 10, batch 2000, loss[loss=0.162, simple_loss=0.2216, pruned_loss=0.05118, over 4785.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2564, pruned_loss=0.06198, over 957165.44 frames. ], batch size: 25, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:59:55,769 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 00:00:04,637 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53597.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:00:05,763 INFO [finetune.py:976] (3/7) Epoch 10, batch 2050, loss[loss=0.186, simple_loss=0.2424, pruned_loss=0.06483, over 4852.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.253, pruned_loss=0.06074, over 956465.62 frames. ], batch size: 44, lr: 3.75e-03, grad_scale: 16.0 +2023-04-27 00:00:34,503 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.698e+02 1.922e+02 2.404e+02 4.326e+02, threshold=3.844e+02, percent-clipped=2.0 +2023-04-27 00:00:44,012 INFO [finetune.py:976] (3/7) Epoch 10, batch 2100, loss[loss=0.1464, simple_loss=0.219, pruned_loss=0.03685, over 4829.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2529, pruned_loss=0.06094, over 955676.84 frames. ], batch size: 25, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:00:51,821 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53660.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:00:58,522 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53671.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:01:14,821 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 00:01:16,444 INFO [finetune.py:976] (3/7) Epoch 10, batch 2150, loss[loss=0.2258, simple_loss=0.3008, pruned_loss=0.07538, over 4799.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2576, pruned_loss=0.06376, over 952460.64 frames. ], batch size: 45, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:01:26,028 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53712.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:01:32,728 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53723.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:01:38,805 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2009, 2.1684, 1.9925, 1.9729, 2.5357, 1.9258, 2.9896, 1.8210], + device='cuda:3'), covar=tensor([0.3661, 0.1840, 0.3734, 0.2690, 0.1450, 0.2416, 0.1107, 0.3775], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0345, 0.0429, 0.0360, 0.0386, 0.0381, 0.0377, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:01:41,148 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.745e+02 2.178e+02 2.516e+02 3.729e+02, threshold=4.356e+02, percent-clipped=0.0 +2023-04-27 00:01:49,681 INFO [finetune.py:976] (3/7) Epoch 10, batch 2200, loss[loss=0.2512, simple_loss=0.2974, pruned_loss=0.1025, over 4795.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2596, pruned_loss=0.06457, over 951368.94 frames. ], batch size: 45, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:01:56,951 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 00:01:57,879 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53760.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:02:22,060 INFO [finetune.py:976] (3/7) Epoch 10, batch 2250, loss[loss=0.2407, simple_loss=0.3005, pruned_loss=0.09049, over 4183.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2608, pruned_loss=0.06462, over 951643.16 frames. ], batch size: 65, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:02:46,631 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.602e+02 1.955e+02 2.365e+02 4.712e+02, threshold=3.910e+02, percent-clipped=2.0 +2023-04-27 00:03:06,366 INFO [finetune.py:976] (3/7) Epoch 10, batch 2300, loss[loss=0.1867, simple_loss=0.2506, pruned_loss=0.06143, over 4744.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2611, pruned_loss=0.06386, over 949436.64 frames. ], batch size: 23, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:03:08,169 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0853, 2.7256, 1.1618, 1.3478, 2.1889, 1.1993, 3.3484, 1.6452], + device='cuda:3'), covar=tensor([0.0671, 0.0763, 0.0916, 0.1264, 0.0456, 0.0975, 0.0199, 0.0618], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0067, 0.0049, 0.0047, 0.0051, 0.0052, 0.0078, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 00:03:18,376 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1734, 1.6687, 1.9741, 2.1336, 1.9551, 1.6028, 1.0812, 1.5990], + device='cuda:3'), covar=tensor([0.3765, 0.3838, 0.1965, 0.3003, 0.3177, 0.3200, 0.5091, 0.2654], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0250, 0.0220, 0.0315, 0.0212, 0.0227, 0.0234, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 00:04:00,278 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.46 vs. limit=5.0 +2023-04-27 00:04:10,840 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53897.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:04:11,998 INFO [finetune.py:976] (3/7) Epoch 10, batch 2350, loss[loss=0.1695, simple_loss=0.2343, pruned_loss=0.05237, over 4822.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2588, pruned_loss=0.06294, over 951801.49 frames. ], batch size: 30, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:04:57,666 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.640e+02 1.984e+02 2.431e+02 6.591e+02, threshold=3.969e+02, percent-clipped=4.0 +2023-04-27 00:05:08,689 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:05:16,063 INFO [finetune.py:976] (3/7) Epoch 10, batch 2400, loss[loss=0.2126, simple_loss=0.2726, pruned_loss=0.07627, over 4942.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2565, pruned_loss=0.06246, over 953808.12 frames. ], batch size: 33, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:05:23,354 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53960.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:05:32,002 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53971.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:05:47,718 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7222, 2.0276, 1.0087, 1.4329, 2.2610, 1.6221, 1.5696, 1.5621], + device='cuda:3'), covar=tensor([0.0536, 0.0361, 0.0335, 0.0558, 0.0230, 0.0515, 0.0513, 0.0588], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:05:49,429 INFO [finetune.py:976] (3/7) Epoch 10, batch 2450, loss[loss=0.1788, simple_loss=0.2343, pruned_loss=0.06167, over 4784.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2539, pruned_loss=0.06186, over 956056.05 frames. ], batch size: 29, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:05:56,751 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54008.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:06:04,941 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54019.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:06:07,939 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=54023.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:06:15,673 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 1.726e+02 2.083e+02 2.546e+02 5.806e+02, threshold=4.166e+02, percent-clipped=1.0 +2023-04-27 00:06:24,053 INFO [finetune.py:976] (3/7) Epoch 10, batch 2500, loss[loss=0.1675, simple_loss=0.2411, pruned_loss=0.04698, over 4763.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2555, pruned_loss=0.06346, over 956638.87 frames. ], batch size: 28, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:06:39,520 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54071.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:06:41,969 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5767, 1.2205, 1.6660, 2.0873, 1.6880, 1.5351, 1.6256, 1.6088], + device='cuda:3'), covar=tensor([0.5990, 0.8057, 0.8425, 0.8503, 0.7532, 1.0011, 0.9920, 0.9451], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0414, 0.0497, 0.0518, 0.0437, 0.0456, 0.0466, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.9526e-05, 1.0249e-04, 1.1214e-04, 1.2304e-04, 1.0623e-04, 1.1014e-04, + 1.1174e-04, 1.1201e-04], device='cuda:3') +2023-04-27 00:06:57,698 INFO [finetune.py:976] (3/7) Epoch 10, batch 2550, loss[loss=0.1738, simple_loss=0.2513, pruned_loss=0.04813, over 4898.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2591, pruned_loss=0.06431, over 955353.10 frames. ], batch size: 36, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:07:19,921 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-04-27 00:07:22,641 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.598e+02 1.963e+02 2.419e+02 4.837e+02, threshold=3.926e+02, percent-clipped=1.0 +2023-04-27 00:07:30,643 INFO [finetune.py:976] (3/7) Epoch 10, batch 2600, loss[loss=0.1997, simple_loss=0.2744, pruned_loss=0.06244, over 4843.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2603, pruned_loss=0.06413, over 955139.02 frames. ], batch size: 47, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:08:04,396 INFO [finetune.py:976] (3/7) Epoch 10, batch 2650, loss[loss=0.1738, simple_loss=0.252, pruned_loss=0.04783, over 4687.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2618, pruned_loss=0.06389, over 957564.21 frames. ], batch size: 59, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:08:06,180 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 00:08:39,254 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0764, 1.9604, 2.4815, 2.6999, 1.9703, 1.7041, 2.0851, 1.1362], + device='cuda:3'), covar=tensor([0.0756, 0.1178, 0.0513, 0.0735, 0.0928, 0.1370, 0.0996, 0.1065], + device='cuda:3'), in_proj_covar=tensor([0.0065, 0.0072, 0.0070, 0.0066, 0.0075, 0.0095, 0.0076, 0.0072], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:08:39,723 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.638e+02 1.968e+02 2.302e+02 4.272e+02, threshold=3.936e+02, percent-clipped=1.0 +2023-04-27 00:08:52,973 INFO [finetune.py:976] (3/7) Epoch 10, batch 2700, loss[loss=0.1891, simple_loss=0.2559, pruned_loss=0.06117, over 4936.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.261, pruned_loss=0.06316, over 956729.44 frames. ], batch size: 33, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:09:00,221 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0666, 1.5177, 1.4711, 1.7801, 1.5982, 1.9073, 1.4548, 3.3375], + device='cuda:3'), covar=tensor([0.0731, 0.0831, 0.0819, 0.1173, 0.0646, 0.0507, 0.0761, 0.0141], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-27 00:10:03,552 INFO [finetune.py:976] (3/7) Epoch 10, batch 2750, loss[loss=0.1214, simple_loss=0.1986, pruned_loss=0.02205, over 4793.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2577, pruned_loss=0.06225, over 957301.51 frames. ], batch size: 29, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:10:49,710 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.579e+02 1.855e+02 2.381e+02 3.900e+02, threshold=3.711e+02, percent-clipped=0.0 +2023-04-27 00:11:09,555 INFO [finetune.py:976] (3/7) Epoch 10, batch 2800, loss[loss=0.1973, simple_loss=0.2582, pruned_loss=0.06819, over 4905.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2546, pruned_loss=0.06121, over 959116.14 frames. ], batch size: 35, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:11:10,270 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1810, 2.7095, 1.0803, 1.3917, 2.0702, 1.2430, 3.5120, 1.8364], + device='cuda:3'), covar=tensor([0.0651, 0.0572, 0.0754, 0.1289, 0.0529, 0.1058, 0.0327, 0.0628], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 00:11:21,348 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-27 00:11:27,933 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 00:11:31,542 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8450, 1.3594, 1.5999, 2.1245, 2.3750, 1.5611, 1.3696, 1.8906], + device='cuda:3'), covar=tensor([0.0902, 0.1618, 0.1053, 0.0623, 0.0506, 0.1071, 0.0988, 0.0657], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0203, 0.0182, 0.0175, 0.0178, 0.0188, 0.0159, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:11:33,685 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 00:11:48,127 INFO [finetune.py:976] (3/7) Epoch 10, batch 2850, loss[loss=0.2462, simple_loss=0.3079, pruned_loss=0.09226, over 4755.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2528, pruned_loss=0.06044, over 956431.74 frames. ], batch size: 54, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:11:54,527 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9300, 2.5733, 1.8900, 1.7240, 1.3988, 1.4285, 1.9329, 1.3694], + device='cuda:3'), covar=tensor([0.1795, 0.1358, 0.1625, 0.1978, 0.2529, 0.2029, 0.1119, 0.2172], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0215, 0.0170, 0.0205, 0.0204, 0.0184, 0.0160, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 00:12:08,303 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8230, 4.2501, 0.8666, 2.1434, 2.4355, 2.9183, 2.4397, 0.9718], + device='cuda:3'), covar=tensor([0.1564, 0.1367, 0.2421, 0.1537, 0.1095, 0.1121, 0.1564, 0.2370], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0248, 0.0139, 0.0122, 0.0134, 0.0154, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:12:11,420 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7262, 1.9765, 1.8923, 1.9984, 1.8259, 1.9710, 2.0236, 1.8400], + device='cuda:3'), covar=tensor([0.5772, 0.7872, 0.6964, 0.6336, 0.7300, 0.9633, 0.8053, 0.7653], + device='cuda:3'), in_proj_covar=tensor([0.0322, 0.0377, 0.0311, 0.0323, 0.0335, 0.0397, 0.0355, 0.0320], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 00:12:11,871 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.761e+02 2.016e+02 2.397e+02 4.207e+02, threshold=4.033e+02, percent-clipped=3.0 +2023-04-27 00:12:21,809 INFO [finetune.py:976] (3/7) Epoch 10, batch 2900, loss[loss=0.2111, simple_loss=0.2808, pruned_loss=0.07069, over 4825.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2574, pruned_loss=0.06255, over 954695.71 frames. ], batch size: 40, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:12:23,246 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-27 00:12:38,948 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1926, 1.6880, 1.5414, 1.9089, 1.7434, 2.0195, 1.5178, 3.7642], + device='cuda:3'), covar=tensor([0.0664, 0.0793, 0.0776, 0.1195, 0.0645, 0.0507, 0.0769, 0.0139], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:3') +2023-04-27 00:12:55,792 INFO [finetune.py:976] (3/7) Epoch 10, batch 2950, loss[loss=0.2041, simple_loss=0.2801, pruned_loss=0.06408, over 4755.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.26, pruned_loss=0.0627, over 953624.05 frames. ], batch size: 54, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:13:12,780 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6647, 2.6163, 2.8400, 3.1651, 2.9954, 2.4638, 2.1938, 2.7837], + device='cuda:3'), covar=tensor([0.0869, 0.0879, 0.0564, 0.0594, 0.0635, 0.0866, 0.0807, 0.0589], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0201, 0.0181, 0.0173, 0.0176, 0.0186, 0.0158, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:13:19,226 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.620e+02 2.110e+02 2.463e+02 5.874e+02, threshold=4.221e+02, percent-clipped=2.0 +2023-04-27 00:13:29,120 INFO [finetune.py:976] (3/7) Epoch 10, batch 3000, loss[loss=0.1867, simple_loss=0.2645, pruned_loss=0.05447, over 4802.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2608, pruned_loss=0.06278, over 953916.64 frames. ], batch size: 51, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:13:29,120 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 00:13:37,590 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8550, 1.1360, 1.7286, 2.2734, 2.0053, 1.7884, 1.7314, 1.7833], + device='cuda:3'), covar=tensor([0.6202, 0.8281, 0.7906, 0.8120, 0.7421, 0.9719, 1.0136, 0.8616], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0414, 0.0498, 0.0517, 0.0438, 0.0456, 0.0467, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.9470e-05, 1.0275e-04, 1.1241e-04, 1.2299e-04, 1.0628e-04, 1.1019e-04, + 1.1196e-04, 1.1194e-04], device='cuda:3') +2023-04-27 00:13:45,198 INFO [finetune.py:1010] (3/7) Epoch 10, validation: loss=0.1531, simple_loss=0.2257, pruned_loss=0.04026, over 2265189.00 frames. +2023-04-27 00:13:45,198 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 00:14:32,347 INFO [finetune.py:976] (3/7) Epoch 10, batch 3050, loss[loss=0.1924, simple_loss=0.2666, pruned_loss=0.05908, over 4825.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2613, pruned_loss=0.06274, over 954157.44 frames. ], batch size: 39, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:14:50,009 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.87 vs. limit=5.0 +2023-04-27 00:14:57,002 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.192e+02 1.553e+02 1.895e+02 2.251e+02 3.622e+02, threshold=3.789e+02, percent-clipped=0.0 +2023-04-27 00:15:05,017 INFO [finetune.py:976] (3/7) Epoch 10, batch 3100, loss[loss=0.2309, simple_loss=0.3005, pruned_loss=0.08068, over 4861.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.26, pruned_loss=0.06276, over 953717.81 frames. ], batch size: 44, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:15:13,057 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4406, 1.6668, 1.7059, 1.8933, 1.7244, 1.8365, 1.8732, 1.7975], + device='cuda:3'), covar=tensor([0.4314, 0.5791, 0.5104, 0.4693, 0.6129, 0.8414, 0.5619, 0.5550], + device='cuda:3'), in_proj_covar=tensor([0.0322, 0.0377, 0.0312, 0.0322, 0.0335, 0.0397, 0.0356, 0.0320], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 00:15:27,248 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54664.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:16:11,741 INFO [finetune.py:976] (3/7) Epoch 10, batch 3150, loss[loss=0.154, simple_loss=0.2158, pruned_loss=0.04612, over 4795.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2573, pruned_loss=0.06244, over 954008.20 frames. ], batch size: 25, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:16:53,623 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54725.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 00:16:55,447 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5178, 1.1595, 1.2603, 1.2108, 1.6847, 1.3818, 1.1205, 1.2293], + device='cuda:3'), covar=tensor([0.1340, 0.1280, 0.1920, 0.1452, 0.0828, 0.1724, 0.1884, 0.2085], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0323, 0.0355, 0.0298, 0.0337, 0.0322, 0.0308, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.4655e-05, 6.8365e-05, 7.6536e-05, 6.1341e-05, 7.0399e-05, 6.9056e-05, + 6.6150e-05, 7.7461e-05], device='cuda:3') +2023-04-27 00:17:04,897 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.175e+02 1.617e+02 1.915e+02 2.289e+02 4.542e+02, threshold=3.830e+02, percent-clipped=1.0 +2023-04-27 00:17:25,056 INFO [finetune.py:976] (3/7) Epoch 10, batch 3200, loss[loss=0.1848, simple_loss=0.2463, pruned_loss=0.06168, over 4912.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2536, pruned_loss=0.06135, over 954469.38 frames. ], batch size: 36, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:17:25,971 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 00:17:26,623 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.29 vs. limit=5.0 +2023-04-27 00:18:08,051 INFO [finetune.py:976] (3/7) Epoch 10, batch 3250, loss[loss=0.1959, simple_loss=0.2709, pruned_loss=0.06045, over 4855.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2541, pruned_loss=0.06168, over 956314.16 frames. ], batch size: 44, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:18:33,642 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.703e+02 2.055e+02 2.409e+02 4.051e+02, threshold=4.111e+02, percent-clipped=1.0 +2023-04-27 00:18:34,405 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4191, 2.5447, 2.1352, 2.0122, 2.5436, 2.1316, 3.1650, 1.6897], + device='cuda:3'), covar=tensor([0.3191, 0.1471, 0.3185, 0.2447, 0.1384, 0.2166, 0.1017, 0.4269], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0346, 0.0431, 0.0359, 0.0385, 0.0381, 0.0377, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:18:42,048 INFO [finetune.py:976] (3/7) Epoch 10, batch 3300, loss[loss=0.2031, simple_loss=0.2727, pruned_loss=0.06669, over 4809.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2579, pruned_loss=0.06289, over 956784.23 frames. ], batch size: 51, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:18:47,241 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-04-27 00:18:48,275 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7469, 1.3907, 1.3358, 1.5459, 1.9393, 1.5425, 1.3259, 1.2857], + device='cuda:3'), covar=tensor([0.1381, 0.1377, 0.2204, 0.1460, 0.0842, 0.1582, 0.2225, 0.2167], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0323, 0.0355, 0.0298, 0.0337, 0.0322, 0.0309, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.4452e-05, 6.8419e-05, 7.6614e-05, 6.1344e-05, 7.0530e-05, 6.8977e-05, + 6.6217e-05, 7.7475e-05], device='cuda:3') +2023-04-27 00:18:50,593 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54862.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:19:11,461 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-04-27 00:19:15,379 INFO [finetune.py:976] (3/7) Epoch 10, batch 3350, loss[loss=0.2039, simple_loss=0.2763, pruned_loss=0.0658, over 4824.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2603, pruned_loss=0.06395, over 955483.49 frames. ], batch size: 39, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:19:30,574 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54923.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 00:19:39,814 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.272e+02 1.802e+02 2.053e+02 2.510e+02 4.625e+02, threshold=4.107e+02, percent-clipped=1.0 +2023-04-27 00:19:47,704 INFO [finetune.py:976] (3/7) Epoch 10, batch 3400, loss[loss=0.2232, simple_loss=0.3036, pruned_loss=0.07141, over 4891.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2621, pruned_loss=0.06416, over 954880.63 frames. ], batch size: 37, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:20:12,759 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2054, 1.3068, 1.6329, 1.7768, 1.6139, 1.7951, 1.7514, 1.7181], + device='cuda:3'), covar=tensor([0.4654, 0.6746, 0.5989, 0.5455, 0.6609, 0.8431, 0.6184, 0.5767], + device='cuda:3'), in_proj_covar=tensor([0.0324, 0.0378, 0.0314, 0.0324, 0.0338, 0.0398, 0.0357, 0.0322], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 00:20:18,846 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-04-27 00:20:20,498 INFO [finetune.py:976] (3/7) Epoch 10, batch 3450, loss[loss=0.1634, simple_loss=0.2437, pruned_loss=0.04157, over 4764.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2612, pruned_loss=0.06332, over 953914.06 frames. ], batch size: 28, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:20:25,801 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.08 vs. limit=5.0 +2023-04-27 00:20:34,367 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55020.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 00:20:42,579 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5623, 1.3548, 1.2398, 1.4507, 1.7847, 1.5052, 1.3002, 1.1470], + device='cuda:3'), covar=tensor([0.1217, 0.1108, 0.1512, 0.1119, 0.0623, 0.1173, 0.1634, 0.1773], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0323, 0.0355, 0.0298, 0.0337, 0.0321, 0.0308, 0.0360], + device='cuda:3'), out_proj_covar=tensor([6.4486e-05, 6.8370e-05, 7.6580e-05, 6.1342e-05, 7.0626e-05, 6.8712e-05, + 6.5921e-05, 7.7353e-05], device='cuda:3') +2023-04-27 00:20:45,490 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.637e+02 2.017e+02 2.373e+02 5.913e+02, threshold=4.034e+02, percent-clipped=2.0 +2023-04-27 00:20:53,438 INFO [finetune.py:976] (3/7) Epoch 10, batch 3500, loss[loss=0.1762, simple_loss=0.2385, pruned_loss=0.05691, over 4817.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2586, pruned_loss=0.06223, over 954203.19 frames. ], batch size: 41, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:21:45,549 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-27 00:21:46,852 INFO [finetune.py:976] (3/7) Epoch 10, batch 3550, loss[loss=0.1661, simple_loss=0.2454, pruned_loss=0.04339, over 4838.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2563, pruned_loss=0.06166, over 955867.25 frames. ], batch size: 30, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:21:59,860 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-27 00:22:19,493 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6667, 1.4114, 0.7524, 1.3248, 1.4213, 1.5399, 1.4127, 1.3869], + device='cuda:3'), covar=tensor([0.0521, 0.0424, 0.0406, 0.0583, 0.0300, 0.0539, 0.0578, 0.0623], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:22:27,074 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.344e+01 1.599e+02 1.898e+02 2.251e+02 4.553e+02, threshold=3.796e+02, percent-clipped=2.0 +2023-04-27 00:22:35,900 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55148.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:22:36,396 INFO [finetune.py:976] (3/7) Epoch 10, batch 3600, loss[loss=0.2032, simple_loss=0.2676, pruned_loss=0.06946, over 4860.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2541, pruned_loss=0.06098, over 955489.27 frames. ], batch size: 34, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:23:26,775 INFO [finetune.py:976] (3/7) Epoch 10, batch 3650, loss[loss=0.251, simple_loss=0.3153, pruned_loss=0.09337, over 4803.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2568, pruned_loss=0.062, over 955831.82 frames. ], batch size: 40, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:23:33,117 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55209.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:23:34,360 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55211.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:23:39,096 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55218.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:23:50,724 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.786e+02 2.133e+02 2.599e+02 5.918e+02, threshold=4.267e+02, percent-clipped=5.0 +2023-04-27 00:24:00,573 INFO [finetune.py:976] (3/7) Epoch 10, batch 3700, loss[loss=0.2242, simple_loss=0.3042, pruned_loss=0.07208, over 4839.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2605, pruned_loss=0.06293, over 956157.36 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:24:15,034 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:24:33,932 INFO [finetune.py:976] (3/7) Epoch 10, batch 3750, loss[loss=0.2405, simple_loss=0.2977, pruned_loss=0.09163, over 4914.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2608, pruned_loss=0.06263, over 954667.72 frames. ], batch size: 37, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:24:47,356 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55320.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 00:24:57,141 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.679e+02 1.930e+02 2.224e+02 3.496e+02, threshold=3.860e+02, percent-clipped=0.0 +2023-04-27 00:25:07,102 INFO [finetune.py:976] (3/7) Epoch 10, batch 3800, loss[loss=0.1918, simple_loss=0.2606, pruned_loss=0.06155, over 4767.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2627, pruned_loss=0.06347, over 954911.44 frames. ], batch size: 28, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:25:19,728 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55368.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:25:40,048 INFO [finetune.py:976] (3/7) Epoch 10, batch 3850, loss[loss=0.1983, simple_loss=0.2674, pruned_loss=0.06463, over 4917.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2596, pruned_loss=0.0621, over 954770.12 frames. ], batch size: 37, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:25:52,109 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7119, 1.8927, 1.8706, 2.5260, 2.7427, 2.3221, 2.2791, 2.0451], + device='cuda:3'), covar=tensor([0.2005, 0.1955, 0.2279, 0.1734, 0.1224, 0.1888, 0.2247, 0.2247], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0321, 0.0354, 0.0298, 0.0335, 0.0319, 0.0306, 0.0359], + device='cuda:3'), out_proj_covar=tensor([6.4171e-05, 6.8032e-05, 7.6431e-05, 6.1364e-05, 7.0077e-05, 6.8283e-05, + 6.5568e-05, 7.7135e-05], device='cuda:3') +2023-04-27 00:26:04,614 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.602e+02 1.915e+02 2.260e+02 3.579e+02, threshold=3.830e+02, percent-clipped=0.0 +2023-04-27 00:26:12,956 INFO [finetune.py:976] (3/7) Epoch 10, batch 3900, loss[loss=0.1751, simple_loss=0.2277, pruned_loss=0.06128, over 4078.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2554, pruned_loss=0.0606, over 953846.10 frames. ], batch size: 17, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:26:24,649 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7323, 1.3126, 1.8316, 2.1848, 1.8429, 1.6915, 1.7774, 1.7437], + device='cuda:3'), covar=tensor([0.5550, 0.7645, 0.7376, 0.7598, 0.6757, 0.9160, 0.9085, 0.9302], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0414, 0.0497, 0.0517, 0.0437, 0.0456, 0.0467, 0.0464], + device='cuda:3'), out_proj_covar=tensor([9.9496e-05, 1.0251e-04, 1.1228e-04, 1.2272e-04, 1.0596e-04, 1.1016e-04, + 1.1187e-04, 1.1171e-04], device='cuda:3') +2023-04-27 00:26:31,221 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0798, 1.8137, 2.0991, 2.5379, 2.6177, 2.0302, 1.6148, 2.2085], + device='cuda:3'), covar=tensor([0.0934, 0.1255, 0.0795, 0.0595, 0.0565, 0.0942, 0.0914, 0.0618], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0205, 0.0183, 0.0176, 0.0179, 0.0189, 0.0161, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:27:07,758 INFO [finetune.py:976] (3/7) Epoch 10, batch 3950, loss[loss=0.1536, simple_loss=0.2328, pruned_loss=0.03725, over 4882.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.253, pruned_loss=0.06021, over 954735.11 frames. ], batch size: 35, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:27:15,671 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55504.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:27:36,357 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55518.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:27:47,671 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.461e+01 1.659e+02 2.039e+02 2.496e+02 7.335e+02, threshold=4.077e+02, percent-clipped=3.0 +2023-04-27 00:27:56,583 INFO [finetune.py:976] (3/7) Epoch 10, batch 4000, loss[loss=0.1313, simple_loss=0.1987, pruned_loss=0.03191, over 4757.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2538, pruned_loss=0.06086, over 955404.80 frames. ], batch size: 26, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:28:09,058 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55566.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:28:12,138 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55567.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:28:43,372 INFO [finetune.py:976] (3/7) Epoch 10, batch 4050, loss[loss=0.2355, simple_loss=0.2997, pruned_loss=0.08562, over 4902.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2581, pruned_loss=0.06271, over 955034.95 frames. ], batch size: 36, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:28:48,874 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4198, 1.2016, 1.7419, 1.6322, 1.3166, 1.1222, 1.3340, 1.0035], + device='cuda:3'), covar=tensor([0.0657, 0.0921, 0.0535, 0.0768, 0.0836, 0.1364, 0.0734, 0.0816], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0073, 0.0072, 0.0068, 0.0076, 0.0096, 0.0078, 0.0073], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:29:09,493 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.669e+02 1.991e+02 2.505e+02 4.319e+02, threshold=3.981e+02, percent-clipped=1.0 +2023-04-27 00:29:16,837 INFO [finetune.py:976] (3/7) Epoch 10, batch 4100, loss[loss=0.2038, simple_loss=0.2762, pruned_loss=0.06567, over 4828.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2612, pruned_loss=0.06381, over 953548.48 frames. ], batch size: 47, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:29:27,122 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 00:29:42,498 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1469, 2.1380, 2.1078, 1.7777, 2.2099, 1.8880, 2.8691, 1.7798], + device='cuda:3'), covar=tensor([0.3893, 0.1724, 0.3690, 0.2795, 0.1634, 0.2404, 0.1366, 0.4524], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0348, 0.0431, 0.0360, 0.0387, 0.0384, 0.0381, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:29:50,673 INFO [finetune.py:976] (3/7) Epoch 10, batch 4150, loss[loss=0.1623, simple_loss=0.2382, pruned_loss=0.04323, over 4745.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2619, pruned_loss=0.06435, over 953536.08 frames. ], batch size: 27, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:30:16,063 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.755e+02 2.031e+02 2.318e+02 3.746e+02, threshold=4.063e+02, percent-clipped=0.0 +2023-04-27 00:30:23,272 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55748.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:30:23,768 INFO [finetune.py:976] (3/7) Epoch 10, batch 4200, loss[loss=0.1688, simple_loss=0.234, pruned_loss=0.05177, over 4744.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2615, pruned_loss=0.06364, over 952646.35 frames. ], batch size: 59, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:30:43,231 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1799, 1.7048, 2.0930, 2.4889, 2.0140, 1.5965, 1.3154, 1.8141], + device='cuda:3'), covar=tensor([0.3415, 0.3626, 0.1714, 0.2491, 0.3138, 0.2718, 0.4880, 0.2521], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0250, 0.0220, 0.0317, 0.0214, 0.0228, 0.0234, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 00:30:57,481 INFO [finetune.py:976] (3/7) Epoch 10, batch 4250, loss[loss=0.2094, simple_loss=0.2748, pruned_loss=0.07196, over 4700.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2602, pruned_loss=0.06422, over 952000.26 frames. ], batch size: 54, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:31:00,763 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55804.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:31:04,341 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55809.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:31:23,549 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.518e+01 1.564e+02 1.887e+02 2.474e+02 5.983e+02, threshold=3.773e+02, percent-clipped=1.0 +2023-04-27 00:31:30,790 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:31:31,266 INFO [finetune.py:976] (3/7) Epoch 10, batch 4300, loss[loss=0.1471, simple_loss=0.2218, pruned_loss=0.03615, over 4772.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2564, pruned_loss=0.06244, over 953883.64 frames. ], batch size: 27, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:31:33,044 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:31:48,280 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55867.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:31:59,626 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 00:32:31,697 INFO [finetune.py:976] (3/7) Epoch 10, batch 4350, loss[loss=0.1563, simple_loss=0.2349, pruned_loss=0.03888, over 4743.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2523, pruned_loss=0.06086, over 951474.20 frames. ], batch size: 27, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:32:41,546 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0700, 1.8401, 2.1324, 2.4437, 2.5044, 1.8749, 1.5940, 2.2186], + device='cuda:3'), covar=tensor([0.0860, 0.1162, 0.0719, 0.0635, 0.0597, 0.1027, 0.0919, 0.0574], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0205, 0.0184, 0.0177, 0.0179, 0.0190, 0.0161, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:32:43,369 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:32:43,963 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2815, 2.1923, 2.4498, 2.7904, 2.7429, 2.0694, 1.8145, 2.4142], + device='cuda:3'), covar=tensor([0.0892, 0.1022, 0.0602, 0.0598, 0.0699, 0.0928, 0.0965, 0.0586], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0205, 0.0183, 0.0177, 0.0179, 0.0190, 0.0161, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:32:45,533 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 00:32:46,920 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55915.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:33:02,707 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.769e+02 2.030e+02 2.376e+02 3.891e+02, threshold=4.060e+02, percent-clipped=2.0 +2023-04-27 00:33:10,570 INFO [finetune.py:976] (3/7) Epoch 10, batch 4400, loss[loss=0.1374, simple_loss=0.1909, pruned_loss=0.042, over 4390.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2528, pruned_loss=0.06111, over 949794.23 frames. ], batch size: 18, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:33:24,942 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 00:34:06,193 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0955, 2.0618, 1.7040, 1.6767, 2.0686, 1.7568, 2.5364, 1.4577], + device='cuda:3'), covar=tensor([0.3703, 0.1970, 0.4798, 0.3260, 0.1951, 0.2404, 0.1548, 0.4956], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0350, 0.0433, 0.0365, 0.0390, 0.0386, 0.0383, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:34:08,060 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6267, 1.3468, 1.7773, 2.1114, 1.7639, 1.6008, 1.6996, 1.6806], + device='cuda:3'), covar=tensor([0.6221, 0.8405, 0.9022, 0.8103, 0.7838, 1.0832, 1.0912, 1.0403], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0415, 0.0500, 0.0519, 0.0439, 0.0460, 0.0470, 0.0468], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:34:18,723 INFO [finetune.py:976] (3/7) Epoch 10, batch 4450, loss[loss=0.1822, simple_loss=0.2594, pruned_loss=0.05249, over 4871.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2566, pruned_loss=0.0618, over 948458.65 frames. ], batch size: 34, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:34:57,405 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.646e+02 1.926e+02 2.335e+02 3.268e+02, threshold=3.851e+02, percent-clipped=0.0 +2023-04-27 00:34:58,362 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-04-27 00:35:05,273 INFO [finetune.py:976] (3/7) Epoch 10, batch 4500, loss[loss=0.1805, simple_loss=0.2482, pruned_loss=0.05642, over 4783.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2579, pruned_loss=0.06216, over 950333.46 frames. ], batch size: 29, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:35:38,736 INFO [finetune.py:976] (3/7) Epoch 10, batch 4550, loss[loss=0.213, simple_loss=0.2843, pruned_loss=0.07084, over 4889.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2596, pruned_loss=0.06276, over 952039.13 frames. ], batch size: 43, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:35:41,832 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56104.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:35:44,523 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 00:35:48,463 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2320, 2.0206, 2.3025, 2.6196, 2.6220, 1.9958, 1.7092, 2.3860], + device='cuda:3'), covar=tensor([0.0791, 0.1119, 0.0613, 0.0555, 0.0559, 0.0938, 0.0845, 0.0523], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0204, 0.0183, 0.0176, 0.0178, 0.0189, 0.0160, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:36:03,337 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.622e+01 1.743e+02 2.073e+02 2.397e+02 5.279e+02, threshold=4.145e+02, percent-clipped=1.0 +2023-04-27 00:36:09,758 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56145.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:36:12,138 INFO [finetune.py:976] (3/7) Epoch 10, batch 4600, loss[loss=0.1544, simple_loss=0.2298, pruned_loss=0.0395, over 4775.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.26, pruned_loss=0.06336, over 951107.41 frames. ], batch size: 28, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:36:15,258 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-27 00:36:17,528 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3395, 3.1524, 0.8124, 1.7465, 1.6657, 2.2558, 1.7898, 0.9892], + device='cuda:3'), covar=tensor([0.1540, 0.1130, 0.2132, 0.1390, 0.1179, 0.1047, 0.1480, 0.2199], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0251, 0.0142, 0.0123, 0.0136, 0.0155, 0.0119, 0.0123], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:36:17,618 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 00:36:45,851 INFO [finetune.py:976] (3/7) Epoch 10, batch 4650, loss[loss=0.1885, simple_loss=0.2568, pruned_loss=0.06016, over 4922.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2584, pruned_loss=0.06337, over 950951.21 frames. ], batch size: 36, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:36:45,925 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.9722, 4.6352, 3.2178, 5.5976, 4.9216, 4.8744, 2.2181, 4.9300], + device='cuda:3'), covar=tensor([0.1387, 0.0824, 0.3051, 0.0885, 0.2666, 0.1751, 0.5297, 0.1712], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0215, 0.0248, 0.0304, 0.0298, 0.0250, 0.0267, 0.0268], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 00:36:49,602 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56204.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:36:50,869 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:36:55,674 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.7424, 4.5528, 3.0852, 5.3495, 4.7552, 4.6324, 2.1498, 4.6301], + device='cuda:3'), covar=tensor([0.1538, 0.0875, 0.3261, 0.0997, 0.2679, 0.1547, 0.5442, 0.1804], + device='cuda:3'), in_proj_covar=tensor([0.0241, 0.0214, 0.0246, 0.0302, 0.0297, 0.0249, 0.0266, 0.0267], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 00:36:58,059 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-27 00:37:16,889 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.791e+01 1.513e+02 1.802e+02 2.222e+02 6.871e+02, threshold=3.603e+02, percent-clipped=2.0 +2023-04-27 00:37:24,883 INFO [finetune.py:976] (3/7) Epoch 10, batch 4700, loss[loss=0.1667, simple_loss=0.2398, pruned_loss=0.0468, over 4942.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2539, pruned_loss=0.06098, over 954138.40 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:37:54,342 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1679, 1.5546, 2.1006, 2.3842, 2.0231, 1.5649, 1.3077, 1.8526], + device='cuda:3'), covar=tensor([0.3486, 0.3682, 0.1735, 0.2811, 0.2976, 0.3055, 0.4683, 0.2508], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0250, 0.0220, 0.0317, 0.0214, 0.0228, 0.0233, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 00:37:57,893 INFO [finetune.py:976] (3/7) Epoch 10, batch 4750, loss[loss=0.1958, simple_loss=0.2634, pruned_loss=0.06409, over 4823.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2521, pruned_loss=0.06025, over 955540.93 frames. ], batch size: 39, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:38:04,953 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4417, 3.1926, 0.7925, 1.6864, 1.6862, 2.2764, 1.7711, 0.9792], + device='cuda:3'), covar=tensor([0.1395, 0.1114, 0.2021, 0.1406, 0.1187, 0.1106, 0.1654, 0.2064], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0250, 0.0141, 0.0122, 0.0135, 0.0154, 0.0118, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:38:06,742 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6798, 1.6597, 0.7044, 1.3105, 1.6786, 1.5545, 1.4418, 1.4062], + device='cuda:3'), covar=tensor([0.0531, 0.0395, 0.0428, 0.0601, 0.0299, 0.0550, 0.0561, 0.0631], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:38:19,548 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5178, 1.3690, 1.7513, 1.7827, 1.3619, 1.0137, 1.3875, 0.8787], + device='cuda:3'), covar=tensor([0.0657, 0.0683, 0.0531, 0.0554, 0.0840, 0.1809, 0.0824, 0.1071], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0073, 0.0071, 0.0067, 0.0076, 0.0096, 0.0077, 0.0073], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:38:23,668 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.579e+02 1.869e+02 2.281e+02 6.054e+02, threshold=3.738e+02, percent-clipped=4.0 +2023-04-27 00:38:31,956 INFO [finetune.py:976] (3/7) Epoch 10, batch 4800, loss[loss=0.1736, simple_loss=0.226, pruned_loss=0.06061, over 2971.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2545, pruned_loss=0.06101, over 953017.09 frames. ], batch size: 12, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:39:29,245 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7353, 2.0802, 0.9887, 1.3699, 2.0353, 1.6060, 1.5210, 1.5410], + device='cuda:3'), covar=tensor([0.0533, 0.0371, 0.0378, 0.0610, 0.0273, 0.0594, 0.0603, 0.0620], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:39:33,342 INFO [finetune.py:976] (3/7) Epoch 10, batch 4850, loss[loss=0.1908, simple_loss=0.2636, pruned_loss=0.05896, over 4855.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2591, pruned_loss=0.06293, over 952486.79 frames. ], batch size: 44, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:39:43,307 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56404.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:39:52,028 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2528, 2.8207, 1.6070, 1.9551, 2.7564, 2.1680, 2.1301, 2.2880], + device='cuda:3'), covar=tensor([0.0463, 0.0295, 0.0290, 0.0495, 0.0195, 0.0465, 0.0477, 0.0508], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0030], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:40:10,801 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-04-27 00:40:14,075 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.686e+02 2.043e+02 2.423e+02 5.639e+02, threshold=4.086e+02, percent-clipped=3.0 +2023-04-27 00:40:22,378 INFO [finetune.py:976] (3/7) Epoch 10, batch 4900, loss[loss=0.2097, simple_loss=0.2725, pruned_loss=0.07342, over 4849.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2605, pruned_loss=0.06327, over 951114.78 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:40:24,262 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56452.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:40:40,333 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 00:40:56,301 INFO [finetune.py:976] (3/7) Epoch 10, batch 4950, loss[loss=0.1734, simple_loss=0.2449, pruned_loss=0.05101, over 4897.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2616, pruned_loss=0.06347, over 952344.87 frames. ], batch size: 36, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:40:57,605 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56501.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:40:59,475 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56504.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:41:12,963 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56523.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:41:21,844 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.631e+02 2.001e+02 2.395e+02 7.277e+02, threshold=4.002e+02, percent-clipped=1.0 +2023-04-27 00:41:29,729 INFO [finetune.py:976] (3/7) Epoch 10, batch 5000, loss[loss=0.1969, simple_loss=0.2648, pruned_loss=0.06448, over 4900.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2598, pruned_loss=0.06258, over 954339.81 frames. ], batch size: 32, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:41:32,065 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56552.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:41:44,571 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5049, 1.4495, 4.0190, 3.7961, 3.5453, 3.7789, 3.7661, 3.5686], + device='cuda:3'), covar=tensor([0.6279, 0.5175, 0.0967, 0.1560, 0.1002, 0.1467, 0.1707, 0.1427], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0308, 0.0407, 0.0409, 0.0350, 0.0406, 0.0315, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:41:53,715 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56584.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 00:42:03,261 INFO [finetune.py:976] (3/7) Epoch 10, batch 5050, loss[loss=0.1962, simple_loss=0.2587, pruned_loss=0.06686, over 4875.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2573, pruned_loss=0.06225, over 955264.56 frames. ], batch size: 34, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:42:56,298 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.627e+02 1.925e+02 2.295e+02 4.200e+02, threshold=3.850e+02, percent-clipped=1.0 +2023-04-27 00:43:09,522 INFO [finetune.py:976] (3/7) Epoch 10, batch 5100, loss[loss=0.1729, simple_loss=0.2454, pruned_loss=0.05022, over 4753.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2539, pruned_loss=0.06085, over 955877.91 frames. ], batch size: 27, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:43:59,161 INFO [finetune.py:976] (3/7) Epoch 10, batch 5150, loss[loss=0.1995, simple_loss=0.2678, pruned_loss=0.06557, over 4876.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2544, pruned_loss=0.06163, over 955678.91 frames. ], batch size: 32, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:44:25,420 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.696e+02 2.008e+02 2.313e+02 3.981e+02, threshold=4.015e+02, percent-clipped=1.0 +2023-04-27 00:44:33,209 INFO [finetune.py:976] (3/7) Epoch 10, batch 5200, loss[loss=0.1885, simple_loss=0.2587, pruned_loss=0.05915, over 4904.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2584, pruned_loss=0.06308, over 955394.85 frames. ], batch size: 37, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:45:03,098 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3225, 1.6920, 2.1536, 2.7144, 2.2053, 1.6903, 1.4742, 1.9916], + device='cuda:3'), covar=tensor([0.3861, 0.3900, 0.1755, 0.2694, 0.3163, 0.3015, 0.4806, 0.2615], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0248, 0.0219, 0.0314, 0.0213, 0.0227, 0.0232, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 00:45:18,108 INFO [finetune.py:976] (3/7) Epoch 10, batch 5250, loss[loss=0.2485, simple_loss=0.3095, pruned_loss=0.09372, over 4915.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2589, pruned_loss=0.06297, over 954510.84 frames. ], batch size: 42, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:45:20,019 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56801.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:45:21,229 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56803.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:45:43,652 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.715e+02 1.995e+02 2.714e+02 4.040e+02, threshold=3.989e+02, percent-clipped=1.0 +2023-04-27 00:45:44,554 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-04-27 00:45:51,395 INFO [finetune.py:976] (3/7) Epoch 10, batch 5300, loss[loss=0.1759, simple_loss=0.2331, pruned_loss=0.05932, over 4401.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2616, pruned_loss=0.06399, over 954105.64 frames. ], batch size: 19, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:45:51,455 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56849.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:45:52,158 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6086, 1.3458, 1.7714, 2.0916, 1.7864, 1.6331, 1.6877, 1.7245], + device='cuda:3'), covar=tensor([0.5598, 0.7581, 0.7634, 0.7812, 0.6796, 0.9432, 0.9866, 0.9197], + device='cuda:3'), in_proj_covar=tensor([0.0408, 0.0413, 0.0496, 0.0517, 0.0438, 0.0457, 0.0468, 0.0466], + device='cuda:3'), out_proj_covar=tensor([9.9449e-05, 1.0225e-04, 1.1201e-04, 1.2276e-04, 1.0593e-04, 1.1046e-04, + 1.1213e-04, 1.1203e-04], device='cuda:3') +2023-04-27 00:45:58,052 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9982, 2.4341, 0.9955, 1.3145, 1.8390, 1.1556, 3.2931, 1.7919], + device='cuda:3'), covar=tensor([0.0668, 0.0571, 0.0722, 0.1353, 0.0535, 0.1039, 0.0243, 0.0634], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0067, 0.0050, 0.0047, 0.0051, 0.0053, 0.0079, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-27 00:46:01,123 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56864.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:46:11,153 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56879.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 00:46:11,208 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5848, 1.3020, 1.9173, 1.8249, 1.3618, 1.1763, 1.5707, 1.0865], + device='cuda:3'), covar=tensor([0.0640, 0.0873, 0.0471, 0.0731, 0.0905, 0.1529, 0.0757, 0.0887], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0073, 0.0071, 0.0068, 0.0076, 0.0097, 0.0077, 0.0073], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:46:25,220 INFO [finetune.py:976] (3/7) Epoch 10, batch 5350, loss[loss=0.1709, simple_loss=0.2407, pruned_loss=0.05054, over 4745.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2616, pruned_loss=0.06302, over 954772.36 frames. ], batch size: 27, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:46:31,365 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56908.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:46:43,974 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-27 00:46:44,417 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5052, 1.5621, 1.6397, 2.3712, 2.4695, 2.0644, 1.9055, 1.8069], + device='cuda:3'), covar=tensor([0.2360, 0.2587, 0.2595, 0.1939, 0.1554, 0.2809, 0.3219, 0.2529], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0321, 0.0355, 0.0299, 0.0337, 0.0320, 0.0307, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.4271e-05, 6.8020e-05, 7.6459e-05, 6.1668e-05, 7.0466e-05, 6.8483e-05, + 6.5688e-05, 7.7484e-05], device='cuda:3') +2023-04-27 00:46:50,604 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.631e+02 1.870e+02 2.324e+02 4.447e+02, threshold=3.741e+02, percent-clipped=2.0 +2023-04-27 00:46:58,345 INFO [finetune.py:976] (3/7) Epoch 10, batch 5400, loss[loss=0.1772, simple_loss=0.2387, pruned_loss=0.05788, over 4831.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2585, pruned_loss=0.06196, over 952235.25 frames. ], batch size: 30, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:47:11,571 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56969.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:47:28,552 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56993.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:47:29,771 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56995.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:47:32,094 INFO [finetune.py:976] (3/7) Epoch 10, batch 5450, loss[loss=0.2053, simple_loss=0.2691, pruned_loss=0.07078, over 4870.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2552, pruned_loss=0.06121, over 950514.03 frames. ], batch size: 31, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:48:18,181 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.695e+02 1.974e+02 2.503e+02 6.150e+02, threshold=3.947e+02, percent-clipped=1.0 +2023-04-27 00:48:38,383 INFO [finetune.py:976] (3/7) Epoch 10, batch 5500, loss[loss=0.1864, simple_loss=0.2549, pruned_loss=0.05895, over 4778.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2524, pruned_loss=0.06049, over 951999.55 frames. ], batch size: 29, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:48:47,134 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57054.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:48:48,366 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57056.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:49:09,200 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 00:49:35,989 INFO [finetune.py:976] (3/7) Epoch 10, batch 5550, loss[loss=0.2251, simple_loss=0.2948, pruned_loss=0.0777, over 4843.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2544, pruned_loss=0.06085, over 952484.16 frames. ], batch size: 49, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:49:47,138 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-27 00:49:59,961 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.628e+02 1.818e+02 2.115e+02 5.040e+02, threshold=3.636e+02, percent-clipped=1.0 +2023-04-27 00:50:06,926 INFO [finetune.py:976] (3/7) Epoch 10, batch 5600, loss[loss=0.2493, simple_loss=0.3083, pruned_loss=0.09516, over 4746.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2588, pruned_loss=0.062, over 952404.82 frames. ], batch size: 54, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:50:12,763 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57159.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:50:20,018 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 00:50:24,871 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57179.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:50:29,414 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57186.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:50:35,751 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57197.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:50:36,870 INFO [finetune.py:976] (3/7) Epoch 10, batch 5650, loss[loss=0.2109, simple_loss=0.2753, pruned_loss=0.07325, over 4761.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2605, pruned_loss=0.06167, over 954379.09 frames. ], batch size: 59, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:50:42,548 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7379, 1.1464, 1.6713, 2.1446, 1.8429, 1.6731, 1.7686, 1.7388], + device='cuda:3'), covar=tensor([0.6048, 0.7289, 0.7974, 0.8441, 0.8190, 0.8661, 0.9152, 0.7423], + device='cuda:3'), in_proj_covar=tensor([0.0407, 0.0411, 0.0496, 0.0515, 0.0436, 0.0456, 0.0466, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.8993e-05, 1.0188e-04, 1.1195e-04, 1.2235e-04, 1.0561e-04, 1.1023e-04, + 1.1180e-04, 1.1192e-04], device='cuda:3') +2023-04-27 00:50:53,686 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57227.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:50:56,035 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8462, 3.6670, 1.1988, 1.9907, 2.1948, 2.6667, 2.2055, 1.2716], + device='cuda:3'), covar=tensor([0.1324, 0.1036, 0.1975, 0.1309, 0.1069, 0.0978, 0.1411, 0.1818], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0249, 0.0142, 0.0122, 0.0134, 0.0153, 0.0118, 0.0123], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:50:57,290 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-27 00:50:59,541 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.673e+02 1.923e+02 2.288e+02 6.250e+02, threshold=3.847e+02, percent-clipped=4.0 +2023-04-27 00:51:00,283 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4106, 2.3320, 2.6357, 2.9277, 2.4933, 2.2889, 2.4976, 2.3095], + device='cuda:3'), covar=tensor([0.5519, 0.7053, 0.8125, 0.6896, 0.6872, 0.9445, 0.9409, 0.8676], + device='cuda:3'), in_proj_covar=tensor([0.0406, 0.0411, 0.0496, 0.0514, 0.0436, 0.0456, 0.0466, 0.0465], + device='cuda:3'), out_proj_covar=tensor([9.8904e-05, 1.0182e-04, 1.1197e-04, 1.2228e-04, 1.0550e-04, 1.1011e-04, + 1.1161e-04, 1.1185e-04], device='cuda:3') +2023-04-27 00:51:03,825 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57244.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:51:05,594 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57247.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:51:06,699 INFO [finetune.py:976] (3/7) Epoch 10, batch 5700, loss[loss=0.1805, simple_loss=0.2512, pruned_loss=0.05485, over 4407.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2561, pruned_loss=0.06139, over 933849.78 frames. ], batch size: 19, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:51:09,427 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-27 00:51:12,266 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57258.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:51:15,768 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57264.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:51:18,164 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([5.0051, 5.0585, 3.6822, 5.7715, 5.1068, 5.0139, 2.9817, 4.9644], + device='cuda:3'), covar=tensor([0.1547, 0.0917, 0.2321, 0.0866, 0.2337, 0.1583, 0.4848, 0.1906], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0214, 0.0248, 0.0303, 0.0298, 0.0248, 0.0267, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 00:51:38,656 INFO [finetune.py:976] (3/7) Epoch 11, batch 0, loss[loss=0.2153, simple_loss=0.2727, pruned_loss=0.07893, over 4906.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2727, pruned_loss=0.07893, over 4906.00 frames. ], batch size: 36, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:51:38,657 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 00:51:55,310 INFO [finetune.py:1010] (3/7) Epoch 11, validation: loss=0.1558, simple_loss=0.2272, pruned_loss=0.04225, over 2265189.00 frames. +2023-04-27 00:51:55,311 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 00:52:28,814 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57305.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:52:42,959 INFO [finetune.py:976] (3/7) Epoch 11, batch 50, loss[loss=0.1564, simple_loss=0.2286, pruned_loss=0.04205, over 4810.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2597, pruned_loss=0.0629, over 217523.55 frames. ], batch size: 39, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:52:49,892 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.652e+02 2.063e+02 2.429e+02 4.586e+02, threshold=4.127e+02, percent-clipped=3.0 +2023-04-27 00:52:57,809 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57349.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:52:59,007 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57351.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:53:20,601 INFO [finetune.py:976] (3/7) Epoch 11, batch 100, loss[loss=0.1383, simple_loss=0.2122, pruned_loss=0.03223, over 4756.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2522, pruned_loss=0.0597, over 380830.30 frames. ], batch size: 26, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:53:38,529 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5115, 2.4444, 2.6598, 2.9216, 2.9691, 2.3339, 2.0554, 2.6872], + device='cuda:3'), covar=tensor([0.0840, 0.0906, 0.0508, 0.0572, 0.0496, 0.0894, 0.0858, 0.0530], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0201, 0.0180, 0.0173, 0.0176, 0.0187, 0.0158, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 00:53:40,241 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57398.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:53:59,651 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5485, 1.1282, 0.4063, 1.2321, 1.0173, 1.4169, 1.2765, 1.3085], + device='cuda:3'), covar=tensor([0.0546, 0.0437, 0.0465, 0.0570, 0.0343, 0.0551, 0.0551, 0.0631], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:54:10,095 INFO [finetune.py:976] (3/7) Epoch 11, batch 150, loss[loss=0.1748, simple_loss=0.2365, pruned_loss=0.05655, over 4873.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2498, pruned_loss=0.05988, over 510207.45 frames. ], batch size: 34, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:54:27,173 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.639e+02 1.968e+02 2.285e+02 5.137e+02, threshold=3.937e+02, percent-clipped=1.0 +2023-04-27 00:54:52,715 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:54:52,738 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:00,045 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:09,555 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6635, 1.4422, 0.7093, 1.2937, 1.3187, 1.5125, 1.3728, 1.3576], + device='cuda:3'), covar=tensor([0.0513, 0.0398, 0.0397, 0.0571, 0.0318, 0.0541, 0.0531, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:55:10,526 INFO [finetune.py:976] (3/7) Epoch 11, batch 200, loss[loss=0.1637, simple_loss=0.2387, pruned_loss=0.04435, over 4766.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2485, pruned_loss=0.05989, over 610070.38 frames. ], batch size: 28, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:55:22,548 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:30,376 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57507.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:40,658 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57523.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:42,940 INFO [finetune.py:976] (3/7) Epoch 11, batch 250, loss[loss=0.1413, simple_loss=0.2132, pruned_loss=0.03468, over 4830.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2505, pruned_loss=0.06005, over 685986.52 frames. ], batch size: 25, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 00:55:50,073 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.267e+02 1.758e+02 2.095e+02 2.670e+02 5.121e+02, threshold=4.190e+02, percent-clipped=8.0 +2023-04-27 00:55:54,618 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57542.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:56:01,286 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57553.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:56:03,141 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57556.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:56:08,526 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57564.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:56:16,735 INFO [finetune.py:976] (3/7) Epoch 11, batch 300, loss[loss=0.1802, simple_loss=0.2432, pruned_loss=0.05863, over 4902.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2549, pruned_loss=0.06106, over 746733.62 frames. ], batch size: 32, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 00:56:18,066 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9947, 1.9847, 1.2987, 1.7267, 2.0128, 1.8626, 1.7711, 1.7725], + device='cuda:3'), covar=tensor([0.0475, 0.0331, 0.0344, 0.0534, 0.0270, 0.0456, 0.0464, 0.0529], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 00:56:32,871 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57600.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:56:40,095 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57612.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:56:50,090 INFO [finetune.py:976] (3/7) Epoch 11, batch 350, loss[loss=0.2257, simple_loss=0.2828, pruned_loss=0.08428, over 4741.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2563, pruned_loss=0.06154, over 792649.69 frames. ], batch size: 59, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 00:56:56,646 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.613e+02 1.965e+02 2.419e+02 4.062e+02, threshold=3.929e+02, percent-clipped=0.0 +2023-04-27 00:57:11,879 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57649.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:57:13,099 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57651.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:57:24,134 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7894, 2.3460, 1.9755, 2.2250, 1.7770, 2.0109, 1.8888, 1.6591], + device='cuda:3'), covar=tensor([0.1960, 0.1515, 0.0886, 0.1216, 0.3245, 0.1255, 0.1827, 0.2254], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0314, 0.0227, 0.0284, 0.0312, 0.0266, 0.0252, 0.0276], + device='cuda:3'), out_proj_covar=tensor([1.1807e-04, 1.2620e-04, 9.0994e-05, 1.1370e-04, 1.2766e-04, 1.0706e-04, + 1.0280e-04, 1.1046e-04], device='cuda:3') +2023-04-27 00:57:35,892 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 00:57:46,786 INFO [finetune.py:976] (3/7) Epoch 11, batch 400, loss[loss=0.1811, simple_loss=0.248, pruned_loss=0.05712, over 4829.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2582, pruned_loss=0.06141, over 830170.64 frames. ], batch size: 47, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 00:58:10,659 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57697.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:58:12,862 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57699.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:58:27,844 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2716, 2.1129, 2.5404, 2.7571, 2.0448, 1.7724, 2.0841, 1.2778], + device='cuda:3'), covar=tensor([0.0656, 0.1053, 0.0461, 0.0895, 0.0808, 0.1157, 0.0946, 0.0966], + device='cuda:3'), in_proj_covar=tensor([0.0065, 0.0071, 0.0069, 0.0066, 0.0074, 0.0094, 0.0075, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 00:58:30,753 INFO [finetune.py:976] (3/7) Epoch 11, batch 450, loss[loss=0.1627, simple_loss=0.229, pruned_loss=0.04826, over 4787.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2578, pruned_loss=0.0614, over 858952.14 frames. ], batch size: 29, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 00:58:35,005 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57733.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:58:37,318 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.604e+02 1.948e+02 2.335e+02 4.408e+02, threshold=3.896e+02, percent-clipped=1.0 +2023-04-27 00:59:00,416 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57754.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:59:32,063 INFO [finetune.py:976] (3/7) Epoch 11, batch 500, loss[loss=0.1542, simple_loss=0.2184, pruned_loss=0.04498, over 4916.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2548, pruned_loss=0.06021, over 880399.37 frames. ], batch size: 37, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 00:59:48,247 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 00:59:54,251 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57794.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:00:25,877 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57818.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:00:31,728 INFO [finetune.py:976] (3/7) Epoch 11, batch 550, loss[loss=0.2016, simple_loss=0.2652, pruned_loss=0.06894, over 4838.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2519, pruned_loss=0.05952, over 895240.20 frames. ], batch size: 33, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:00:38,240 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.623e+02 1.943e+02 2.373e+02 5.716e+02, threshold=3.887e+02, percent-clipped=4.0 +2023-04-27 01:00:41,362 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57842.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:00:47,291 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57851.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:00:49,025 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57853.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 01:01:04,583 INFO [finetune.py:976] (3/7) Epoch 11, batch 600, loss[loss=0.2017, simple_loss=0.2711, pruned_loss=0.06615, over 4833.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2509, pruned_loss=0.05947, over 909286.56 frames. ], batch size: 33, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:01:08,164 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1307, 2.0740, 1.6803, 1.8575, 2.1534, 1.6691, 2.7115, 1.4592], + device='cuda:3'), covar=tensor([0.4080, 0.2128, 0.5040, 0.3356, 0.2181, 0.2932, 0.1519, 0.4880], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0345, 0.0425, 0.0359, 0.0384, 0.0381, 0.0379, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:01:13,413 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57890.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:01:20,168 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57900.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:01:20,740 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57901.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:01:38,047 INFO [finetune.py:976] (3/7) Epoch 11, batch 650, loss[loss=0.1975, simple_loss=0.267, pruned_loss=0.064, over 4847.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.253, pruned_loss=0.05978, over 919947.02 frames. ], batch size: 44, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:01:44,600 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4395, 1.2588, 1.6021, 1.5436, 1.2890, 1.2181, 1.2885, 0.6798], + device='cuda:3'), covar=tensor([0.0591, 0.0767, 0.0622, 0.0562, 0.0734, 0.1230, 0.0655, 0.1056], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0072, 0.0070, 0.0067, 0.0075, 0.0095, 0.0076, 0.0072], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:01:45,070 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.753e+02 2.161e+02 2.818e+02 6.431e+02, threshold=4.321e+02, percent-clipped=6.0 +2023-04-27 01:01:47,531 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9624, 2.4639, 1.0756, 1.2428, 1.8398, 1.1427, 3.2473, 1.7088], + device='cuda:3'), covar=tensor([0.0729, 0.0613, 0.0767, 0.1328, 0.0509, 0.1112, 0.0249, 0.0630], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0078, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 01:01:51,798 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57948.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:02:11,760 INFO [finetune.py:976] (3/7) Epoch 11, batch 700, loss[loss=0.1995, simple_loss=0.2618, pruned_loss=0.06859, over 4820.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2564, pruned_loss=0.06123, over 927551.81 frames. ], batch size: 39, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:02:33,859 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2678, 3.2146, 2.5025, 3.8375, 3.2389, 3.2336, 1.6897, 3.3086], + device='cuda:3'), covar=tensor([0.1850, 0.1420, 0.3304, 0.2215, 0.2608, 0.1911, 0.5082, 0.2546], + device='cuda:3'), in_proj_covar=tensor([0.0240, 0.0214, 0.0247, 0.0300, 0.0295, 0.0248, 0.0266, 0.0268], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 01:02:57,541 INFO [finetune.py:976] (3/7) Epoch 11, batch 750, loss[loss=0.1522, simple_loss=0.2299, pruned_loss=0.03729, over 4768.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.259, pruned_loss=0.06276, over 932084.40 frames. ], batch size: 28, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:03:04,227 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.211e+02 1.766e+02 1.999e+02 2.458e+02 5.140e+02, threshold=3.998e+02, percent-clipped=2.0 +2023-04-27 01:03:06,092 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.49 vs. limit=5.0 +2023-04-27 01:03:14,989 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58054.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:03:31,279 INFO [finetune.py:976] (3/7) Epoch 11, batch 800, loss[loss=0.1608, simple_loss=0.2384, pruned_loss=0.04164, over 4760.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2571, pruned_loss=0.06106, over 935971.95 frames. ], batch size: 27, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:03:34,648 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 01:03:38,603 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58089.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:03:47,436 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58102.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:03:58,175 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58118.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:04,038 INFO [finetune.py:976] (3/7) Epoch 11, batch 850, loss[loss=0.2221, simple_loss=0.2794, pruned_loss=0.08246, over 4722.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2561, pruned_loss=0.06058, over 941246.28 frames. ], batch size: 23, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:04:10,688 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.590e+02 1.973e+02 2.582e+02 4.894e+02, threshold=3.945e+02, percent-clipped=3.0 +2023-04-27 01:04:12,610 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58141.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:14,916 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:16,069 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9512, 4.0359, 0.6704, 2.2006, 2.2701, 2.6315, 2.3569, 1.0962], + device='cuda:3'), covar=tensor([0.1202, 0.0783, 0.2220, 0.1146, 0.1008, 0.1064, 0.1487, 0.2027], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0250, 0.0141, 0.0123, 0.0134, 0.0154, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:04:19,581 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58151.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:39,246 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:52,833 INFO [finetune.py:976] (3/7) Epoch 11, batch 900, loss[loss=0.1632, simple_loss=0.2264, pruned_loss=0.05004, over 4841.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2541, pruned_loss=0.06053, over 943869.54 frames. ], batch size: 44, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:05:19,788 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58199.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:05:21,027 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3508, 3.3599, 2.6503, 3.8873, 3.3066, 3.3797, 1.7401, 3.2986], + device='cuda:3'), covar=tensor([0.1953, 0.1414, 0.3020, 0.2190, 0.2502, 0.1898, 0.5324, 0.2396], + device='cuda:3'), in_proj_covar=tensor([0.0241, 0.0214, 0.0248, 0.0301, 0.0296, 0.0250, 0.0267, 0.0268], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 01:05:22,289 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58202.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:05:28,583 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58205.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:05:54,758 INFO [finetune.py:976] (3/7) Epoch 11, batch 950, loss[loss=0.1632, simple_loss=0.2169, pruned_loss=0.05479, over 4156.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2521, pruned_loss=0.05979, over 946398.99 frames. ], batch size: 18, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:06:10,243 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.752e+02 2.086e+02 2.428e+02 7.149e+02, threshold=4.172e+02, percent-clipped=3.0 +2023-04-27 01:06:25,289 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1434, 4.3847, 0.9498, 2.2434, 2.5875, 2.9673, 2.6532, 0.9723], + device='cuda:3'), covar=tensor([0.1292, 0.1000, 0.2174, 0.1450, 0.1007, 0.1104, 0.1439, 0.2142], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0249, 0.0140, 0.0123, 0.0134, 0.0154, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:06:36,120 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-27 01:06:47,387 INFO [finetune.py:976] (3/7) Epoch 11, batch 1000, loss[loss=0.2007, simple_loss=0.2734, pruned_loss=0.06403, over 4867.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2553, pruned_loss=0.06144, over 949062.31 frames. ], batch size: 34, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:07:20,925 INFO [finetune.py:976] (3/7) Epoch 11, batch 1050, loss[loss=0.1911, simple_loss=0.2533, pruned_loss=0.06448, over 4808.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2582, pruned_loss=0.0624, over 947855.30 frames. ], batch size: 25, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:07:28,201 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.762e+02 2.120e+02 2.435e+02 5.017e+02, threshold=4.240e+02, percent-clipped=2.0 +2023-04-27 01:07:38,122 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4647, 3.0424, 0.9628, 1.7370, 1.7811, 2.2350, 1.7861, 0.9414], + device='cuda:3'), covar=tensor([0.1238, 0.1060, 0.1703, 0.1206, 0.1010, 0.0978, 0.1606, 0.1856], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0247, 0.0139, 0.0121, 0.0133, 0.0152, 0.0117, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:07:44,744 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4441, 1.8263, 1.7622, 2.2790, 2.0469, 2.0924, 1.7186, 4.5266], + device='cuda:3'), covar=tensor([0.0591, 0.0793, 0.0780, 0.1116, 0.0631, 0.0534, 0.0748, 0.0115], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 01:07:52,945 INFO [finetune.py:976] (3/7) Epoch 11, batch 1100, loss[loss=0.2104, simple_loss=0.284, pruned_loss=0.06842, over 4842.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2577, pruned_loss=0.06149, over 948917.09 frames. ], batch size: 49, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:08:01,690 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58389.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:08:26,953 INFO [finetune.py:976] (3/7) Epoch 11, batch 1150, loss[loss=0.1863, simple_loss=0.2535, pruned_loss=0.05961, over 4725.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2589, pruned_loss=0.06162, over 950206.66 frames. ], batch size: 59, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:08:34,545 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58437.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:08:35,086 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.681e+02 2.024e+02 2.460e+02 8.000e+02, threshold=4.047e+02, percent-clipped=3.0 +2023-04-27 01:08:59,899 INFO [finetune.py:976] (3/7) Epoch 11, batch 1200, loss[loss=0.1491, simple_loss=0.2161, pruned_loss=0.04105, over 4809.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2571, pruned_loss=0.0605, over 950868.38 frames. ], batch size: 38, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:09:13,883 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58497.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:09:15,683 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58500.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:09:19,674 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 01:09:26,041 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 01:09:32,953 INFO [finetune.py:976] (3/7) Epoch 11, batch 1250, loss[loss=0.1692, simple_loss=0.2406, pruned_loss=0.04888, over 4903.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2552, pruned_loss=0.06062, over 951837.23 frames. ], batch size: 46, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:09:41,127 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.214e+02 1.641e+02 1.959e+02 2.303e+02 4.487e+02, threshold=3.918e+02, percent-clipped=1.0 +2023-04-27 01:09:42,362 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8324, 1.5969, 1.8901, 2.2100, 2.2487, 1.8488, 1.5053, 2.0271], + device='cuda:3'), covar=tensor([0.0928, 0.1228, 0.0778, 0.0645, 0.0626, 0.0918, 0.0886, 0.0570], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0200, 0.0180, 0.0173, 0.0176, 0.0185, 0.0158, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:10:19,344 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8182, 2.4643, 1.8266, 1.5902, 1.3520, 1.3691, 1.8884, 1.2757], + device='cuda:3'), covar=tensor([0.1807, 0.1314, 0.1645, 0.1947, 0.2517, 0.2065, 0.1152, 0.2224], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0214, 0.0168, 0.0202, 0.0202, 0.0185, 0.0158, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 01:10:22,917 INFO [finetune.py:976] (3/7) Epoch 11, batch 1300, loss[loss=0.1641, simple_loss=0.2251, pruned_loss=0.05153, over 4913.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2533, pruned_loss=0.06019, over 951086.73 frames. ], batch size: 43, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:11:28,545 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-04-27 01:11:29,574 INFO [finetune.py:976] (3/7) Epoch 11, batch 1350, loss[loss=0.1851, simple_loss=0.2411, pruned_loss=0.06455, over 4051.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2553, pruned_loss=0.06175, over 951075.62 frames. ], batch size: 17, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:11:38,047 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4328, 3.5232, 0.8848, 1.7604, 1.9352, 2.5018, 1.9477, 1.1134], + device='cuda:3'), covar=tensor([0.1428, 0.0893, 0.2124, 0.1374, 0.1097, 0.1084, 0.1570, 0.2068], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0246, 0.0139, 0.0121, 0.0132, 0.0152, 0.0117, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:11:46,674 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.586e+02 1.956e+02 2.328e+02 4.668e+02, threshold=3.913e+02, percent-clipped=1.0 +2023-04-27 01:11:50,362 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7732, 1.4477, 1.4201, 1.4829, 1.9826, 1.5983, 1.2631, 1.3366], + device='cuda:3'), covar=tensor([0.1601, 0.1342, 0.2279, 0.1467, 0.0764, 0.1574, 0.2328, 0.2064], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0321, 0.0352, 0.0299, 0.0335, 0.0320, 0.0306, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.4343e-05, 6.7854e-05, 7.5664e-05, 6.1626e-05, 7.0058e-05, 6.8373e-05, + 6.5497e-05, 7.7468e-05], device='cuda:3') +2023-04-27 01:12:34,429 INFO [finetune.py:976] (3/7) Epoch 11, batch 1400, loss[loss=0.2309, simple_loss=0.2969, pruned_loss=0.08242, over 4857.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2595, pruned_loss=0.06312, over 952201.15 frames. ], batch size: 44, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:12:35,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3616, 1.2735, 3.9565, 3.6848, 3.3993, 3.6762, 3.6526, 3.4675], + device='cuda:3'), covar=tensor([0.7362, 0.5952, 0.1159, 0.1880, 0.1435, 0.2078, 0.2682, 0.1701], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0306, 0.0403, 0.0407, 0.0349, 0.0406, 0.0313, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:12:42,420 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58681.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:12:43,060 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7762, 1.7594, 1.5693, 1.2637, 1.7390, 1.4230, 2.1766, 1.2223], + device='cuda:3'), covar=tensor([0.3739, 0.1597, 0.4607, 0.2810, 0.1818, 0.2518, 0.1534, 0.5099], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0347, 0.0428, 0.0359, 0.0386, 0.0383, 0.0379, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:13:04,201 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 01:13:42,641 INFO [finetune.py:976] (3/7) Epoch 11, batch 1450, loss[loss=0.1975, simple_loss=0.2721, pruned_loss=0.06142, over 4827.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2606, pruned_loss=0.06323, over 951602.42 frames. ], batch size: 47, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:14:01,639 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.919e+01 1.615e+02 1.986e+02 2.296e+02 3.609e+02, threshold=3.972e+02, percent-clipped=0.0 +2023-04-27 01:14:05,183 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58742.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:14:13,561 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 01:14:16,804 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0988, 0.8274, 0.9296, 0.8507, 1.2713, 0.9926, 0.8626, 0.9564], + device='cuda:3'), covar=tensor([0.1485, 0.1416, 0.1925, 0.1540, 0.1026, 0.1359, 0.1832, 0.2183], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0323, 0.0354, 0.0301, 0.0337, 0.0322, 0.0308, 0.0363], + device='cuda:3'), out_proj_covar=tensor([6.4702e-05, 6.8319e-05, 7.6065e-05, 6.2037e-05, 7.0654e-05, 6.8868e-05, + 6.5918e-05, 7.7985e-05], device='cuda:3') +2023-04-27 01:14:37,125 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58775.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:14:38,276 INFO [finetune.py:976] (3/7) Epoch 11, batch 1500, loss[loss=0.216, simple_loss=0.2879, pruned_loss=0.07206, over 4895.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2603, pruned_loss=0.06274, over 951450.11 frames. ], batch size: 43, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:14:46,667 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2573, 2.7148, 1.2912, 1.5651, 2.1345, 1.4588, 3.3132, 1.8354], + device='cuda:3'), covar=tensor([0.0565, 0.0760, 0.0738, 0.1082, 0.0427, 0.0848, 0.0190, 0.0527], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0067, 0.0050, 0.0047, 0.0051, 0.0053, 0.0078, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 01:14:52,522 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:14:54,836 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4445, 3.4302, 0.9732, 1.9289, 1.9945, 2.4317, 1.9717, 0.9781], + device='cuda:3'), covar=tensor([0.1451, 0.0858, 0.2021, 0.1274, 0.1073, 0.1088, 0.1540, 0.2278], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0247, 0.0139, 0.0122, 0.0133, 0.0152, 0.0118, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:14:54,851 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58800.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:14:56,665 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1146, 2.5989, 1.0497, 1.4480, 2.0546, 1.2391, 3.5763, 1.8032], + device='cuda:3'), covar=tensor([0.0663, 0.0833, 0.0862, 0.1241, 0.0502, 0.1021, 0.0273, 0.0614], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0067, 0.0050, 0.0047, 0.0051, 0.0053, 0.0078, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:3') +2023-04-27 01:15:08,105 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7412, 1.9245, 1.3704, 1.5425, 2.0752, 1.6521, 1.6097, 1.6409], + device='cuda:3'), covar=tensor([0.0436, 0.0308, 0.0328, 0.0455, 0.0268, 0.0452, 0.0408, 0.0470], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:3') +2023-04-27 01:15:11,710 INFO [finetune.py:976] (3/7) Epoch 11, batch 1550, loss[loss=0.2027, simple_loss=0.273, pruned_loss=0.0662, over 4851.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2611, pruned_loss=0.06347, over 951773.69 frames. ], batch size: 44, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:15:18,196 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58836.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:15:19,338 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.373e+01 1.631e+02 1.941e+02 2.305e+02 6.407e+02, threshold=3.882e+02, percent-clipped=1.0 +2023-04-27 01:15:24,176 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58845.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:15:26,504 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:15:40,820 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5005, 1.3631, 1.7942, 1.7596, 1.3840, 1.2121, 1.4602, 0.9168], + device='cuda:3'), covar=tensor([0.0587, 0.0731, 0.0501, 0.0617, 0.0849, 0.1399, 0.0707, 0.0898], + device='cuda:3'), in_proj_covar=tensor([0.0066, 0.0071, 0.0070, 0.0067, 0.0075, 0.0095, 0.0076, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:15:44,972 INFO [finetune.py:976] (3/7) Epoch 11, batch 1600, loss[loss=0.2011, simple_loss=0.2711, pruned_loss=0.0655, over 4900.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.259, pruned_loss=0.06302, over 953467.60 frames. ], batch size: 35, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:16:05,011 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58906.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:16:18,089 INFO [finetune.py:976] (3/7) Epoch 11, batch 1650, loss[loss=0.16, simple_loss=0.2271, pruned_loss=0.04651, over 4813.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2561, pruned_loss=0.06192, over 953443.90 frames. ], batch size: 51, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:16:18,897 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-27 01:16:25,722 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.662e+02 1.988e+02 2.360e+02 4.922e+02, threshold=3.975e+02, percent-clipped=3.0 +2023-04-27 01:16:44,576 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2352, 2.9950, 2.3096, 2.6664, 2.0457, 2.4370, 2.5972, 1.8990], + device='cuda:3'), covar=tensor([0.1947, 0.0919, 0.0792, 0.1231, 0.2960, 0.0968, 0.1959, 0.2321], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0313, 0.0227, 0.0282, 0.0311, 0.0267, 0.0251, 0.0273], + device='cuda:3'), out_proj_covar=tensor([1.1812e-04, 1.2568e-04, 9.1132e-05, 1.1304e-04, 1.2722e-04, 1.0753e-04, + 1.0248e-04, 1.0953e-04], device='cuda:3') +2023-04-27 01:17:06,257 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7144, 1.3440, 1.9099, 2.2205, 1.8571, 1.6785, 1.7526, 1.8369], + device='cuda:3'), covar=tensor([0.5751, 0.7918, 0.8168, 0.7466, 0.7005, 1.0125, 0.9500, 0.8813], + device='cuda:3'), in_proj_covar=tensor([0.0408, 0.0412, 0.0498, 0.0517, 0.0439, 0.0457, 0.0468, 0.0467], + device='cuda:3'), out_proj_covar=tensor([9.9374e-05, 1.0213e-04, 1.1239e-04, 1.2277e-04, 1.0623e-04, 1.1054e-04, + 1.1222e-04, 1.1217e-04], device='cuda:3') +2023-04-27 01:17:07,484 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58967.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:17:13,434 INFO [finetune.py:976] (3/7) Epoch 11, batch 1700, loss[loss=0.1857, simple_loss=0.2556, pruned_loss=0.05791, over 4817.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2527, pruned_loss=0.06027, over 954718.44 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:17:41,000 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4902, 1.1618, 4.3171, 4.0618, 3.7248, 4.0240, 3.9833, 3.7586], + device='cuda:3'), covar=tensor([0.7418, 0.6462, 0.0986, 0.1474, 0.1249, 0.1515, 0.1796, 0.1563], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0305, 0.0402, 0.0406, 0.0349, 0.0405, 0.0312, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:17:41,011 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2618, 3.4720, 0.7534, 1.8667, 1.7893, 2.3352, 1.9070, 0.9041], + device='cuda:3'), covar=tensor([0.1473, 0.0779, 0.2123, 0.1263, 0.1087, 0.1111, 0.1476, 0.2236], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0249, 0.0141, 0.0122, 0.0134, 0.0154, 0.0119, 0.0123], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:17:47,002 INFO [finetune.py:976] (3/7) Epoch 11, batch 1750, loss[loss=0.1812, simple_loss=0.2488, pruned_loss=0.05679, over 4766.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2531, pruned_loss=0.06035, over 955010.63 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:17:53,226 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59037.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:17:53,742 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.792e+02 2.172e+02 2.715e+02 5.550e+02, threshold=4.344e+02, percent-clipped=4.0 +2023-04-27 01:18:18,317 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3146, 2.2047, 2.3773, 2.8181, 2.8451, 2.1729, 1.7681, 2.5492], + device='cuda:3'), covar=tensor([0.0830, 0.0908, 0.0591, 0.0523, 0.0532, 0.0833, 0.0911, 0.0471], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0204, 0.0183, 0.0176, 0.0179, 0.0189, 0.0160, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:18:30,830 INFO [finetune.py:976] (3/7) Epoch 11, batch 1800, loss[loss=0.2004, simple_loss=0.2654, pruned_loss=0.06768, over 4772.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2578, pruned_loss=0.06141, over 956088.11 frames. ], batch size: 28, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:19:09,377 INFO [finetune.py:976] (3/7) Epoch 11, batch 1850, loss[loss=0.158, simple_loss=0.2459, pruned_loss=0.03508, over 4816.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2576, pruned_loss=0.06101, over 956035.68 frames. ], batch size: 38, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:19:11,121 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 01:19:11,900 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59131.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:19:21,281 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.699e+02 2.019e+02 2.381e+02 7.878e+02, threshold=4.039e+02, percent-clipped=1.0 +2023-04-27 01:19:52,235 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59164.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:20:10,990 INFO [finetune.py:976] (3/7) Epoch 11, batch 1900, loss[loss=0.2668, simple_loss=0.3152, pruned_loss=0.1093, over 4814.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2595, pruned_loss=0.06164, over 957727.82 frames. ], batch size: 45, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:20:11,350 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 01:21:00,119 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:21:01,213 INFO [finetune.py:976] (3/7) Epoch 11, batch 1950, loss[loss=0.2212, simple_loss=0.2719, pruned_loss=0.0853, over 4703.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2572, pruned_loss=0.06036, over 957159.82 frames. ], batch size: 59, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:21:08,358 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.114e+02 1.600e+02 1.889e+02 2.365e+02 4.581e+02, threshold=3.778e+02, percent-clipped=1.0 +2023-04-27 01:21:20,313 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 01:21:24,550 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59262.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:21:35,025 INFO [finetune.py:976] (3/7) Epoch 11, batch 2000, loss[loss=0.1571, simple_loss=0.2267, pruned_loss=0.04373, over 4767.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2543, pruned_loss=0.05949, over 957642.20 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:22:04,048 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-27 01:22:08,863 INFO [finetune.py:976] (3/7) Epoch 11, batch 2050, loss[loss=0.1652, simple_loss=0.2378, pruned_loss=0.04633, over 4756.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2512, pruned_loss=0.05919, over 954877.11 frames. ], batch size: 27, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:22:15,379 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:22:15,893 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.589e+02 2.008e+02 2.324e+02 3.763e+02, threshold=4.015e+02, percent-clipped=0.0 +2023-04-27 01:22:48,351 INFO [finetune.py:976] (3/7) Epoch 11, batch 2100, loss[loss=0.2554, simple_loss=0.309, pruned_loss=0.1009, over 4913.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2512, pruned_loss=0.05943, over 956099.86 frames. ], batch size: 36, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:22:53,374 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:22:58,192 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3612, 1.6329, 1.6677, 1.7817, 1.6304, 1.7323, 1.7929, 1.7048], + device='cuda:3'), covar=tensor([0.4515, 0.6289, 0.5280, 0.5056, 0.6231, 0.8396, 0.6286, 0.6404], + device='cuda:3'), in_proj_covar=tensor([0.0326, 0.0379, 0.0314, 0.0325, 0.0338, 0.0400, 0.0357, 0.0322], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 01:23:06,288 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2320, 1.5992, 1.3515, 1.7810, 1.5898, 1.9575, 1.3942, 3.6286], + device='cuda:3'), covar=tensor([0.0610, 0.0771, 0.0767, 0.1122, 0.0624, 0.0537, 0.0704, 0.0122], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 01:23:12,699 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6241, 1.5159, 4.4473, 4.1634, 3.8734, 4.2323, 4.1493, 3.8889], + device='cuda:3'), covar=tensor([0.6854, 0.5958, 0.1119, 0.1687, 0.1232, 0.2096, 0.1371, 0.1478], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0303, 0.0397, 0.0403, 0.0344, 0.0401, 0.0310, 0.0363], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:23:32,963 INFO [finetune.py:976] (3/7) Epoch 11, batch 2150, loss[loss=0.2052, simple_loss=0.2838, pruned_loss=0.06325, over 4899.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2537, pruned_loss=0.06063, over 955045.02 frames. ], batch size: 43, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:23:35,511 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59431.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:23:44,713 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.256e+02 1.698e+02 2.142e+02 2.512e+02 4.265e+02, threshold=4.284e+02, percent-clipped=3.0 +2023-04-27 01:24:04,852 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6301, 1.7033, 1.8346, 1.2410, 1.7694, 1.5690, 2.3599, 1.5692], + device='cuda:3'), covar=tensor([0.3362, 0.1657, 0.3892, 0.2759, 0.1466, 0.2125, 0.1366, 0.4346], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0344, 0.0425, 0.0357, 0.0383, 0.0382, 0.0378, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:24:11,431 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5537, 1.3245, 1.7853, 1.7443, 1.3970, 1.3176, 1.4165, 0.9345], + device='cuda:3'), covar=tensor([0.0669, 0.0821, 0.0491, 0.0751, 0.0822, 0.1321, 0.0734, 0.0827], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0072, 0.0070, 0.0067, 0.0075, 0.0096, 0.0076, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:24:15,852 INFO [finetune.py:976] (3/7) Epoch 11, batch 2200, loss[loss=0.1564, simple_loss=0.2257, pruned_loss=0.04352, over 4802.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2569, pruned_loss=0.06122, over 955856.46 frames. ], batch size: 25, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:24:18,112 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59479.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:24:37,651 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4945, 3.4455, 2.6719, 4.0907, 3.4862, 3.4920, 1.6786, 3.4881], + device='cuda:3'), covar=tensor([0.1819, 0.1473, 0.3386, 0.2043, 0.3120, 0.2213, 0.5407, 0.2576], + device='cuda:3'), in_proj_covar=tensor([0.0241, 0.0216, 0.0248, 0.0302, 0.0296, 0.0249, 0.0267, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 01:24:37,686 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8703, 1.5675, 1.8391, 2.2181, 2.2194, 1.7084, 1.4011, 1.9610], + device='cuda:3'), covar=tensor([0.0849, 0.1176, 0.0746, 0.0557, 0.0541, 0.0917, 0.0860, 0.0570], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0204, 0.0184, 0.0176, 0.0179, 0.0189, 0.0160, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:24:37,698 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59502.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:24:47,058 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7225, 1.7217, 1.5904, 1.3106, 1.7533, 1.5367, 2.2158, 1.4276], + device='cuda:3'), covar=tensor([0.3342, 0.1633, 0.4649, 0.2729, 0.1549, 0.2011, 0.1416, 0.4246], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0347, 0.0427, 0.0359, 0.0384, 0.0384, 0.0380, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:24:54,709 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:24:56,713 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-27 01:24:59,364 INFO [finetune.py:976] (3/7) Epoch 11, batch 2250, loss[loss=0.2168, simple_loss=0.2765, pruned_loss=0.07852, over 4896.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2594, pruned_loss=0.06282, over 955274.91 frames. ], batch size: 32, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:25:13,213 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.641e+02 2.091e+02 2.357e+02 4.911e+02, threshold=4.183e+02, percent-clipped=2.0 +2023-04-27 01:25:13,348 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6620, 1.5581, 1.9830, 2.0169, 1.5220, 1.4257, 1.6238, 1.1352], + device='cuda:3'), covar=tensor([0.0596, 0.0881, 0.0427, 0.0850, 0.0817, 0.1274, 0.0752, 0.0796], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0072, 0.0070, 0.0067, 0.0075, 0.0096, 0.0076, 0.0072], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:25:45,854 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59562.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:25:46,507 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59563.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:25:52,822 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0635, 1.0568, 1.2428, 1.1741, 0.9646, 0.9215, 1.0629, 0.5539], + device='cuda:3'), covar=tensor([0.0554, 0.0590, 0.0542, 0.0526, 0.0729, 0.1346, 0.0491, 0.0848], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0072, 0.0070, 0.0067, 0.0075, 0.0096, 0.0076, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:25:55,720 INFO [finetune.py:976] (3/7) Epoch 11, batch 2300, loss[loss=0.1522, simple_loss=0.2225, pruned_loss=0.04097, over 4814.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2598, pruned_loss=0.06247, over 956069.49 frames. ], batch size: 39, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:25:55,952 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 01:26:06,395 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0969, 1.4955, 1.2708, 1.6879, 1.5517, 1.7456, 1.3072, 3.0467], + device='cuda:3'), covar=tensor([0.0681, 0.0823, 0.0862, 0.1260, 0.0674, 0.0498, 0.0813, 0.0167], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 01:26:17,899 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59610.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:26:29,150 INFO [finetune.py:976] (3/7) Epoch 11, batch 2350, loss[loss=0.2235, simple_loss=0.289, pruned_loss=0.07904, over 4810.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2581, pruned_loss=0.06149, over 955785.15 frames. ], batch size: 40, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:26:37,789 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.642e+02 1.943e+02 2.312e+02 3.854e+02, threshold=3.885e+02, percent-clipped=0.0 +2023-04-27 01:26:45,514 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 01:27:02,576 INFO [finetune.py:976] (3/7) Epoch 11, batch 2400, loss[loss=0.1841, simple_loss=0.2557, pruned_loss=0.05625, over 4723.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2555, pruned_loss=0.06084, over 957161.92 frames. ], batch size: 54, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:27:36,062 INFO [finetune.py:976] (3/7) Epoch 11, batch 2450, loss[loss=0.137, simple_loss=0.2077, pruned_loss=0.03311, over 4822.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2518, pruned_loss=0.05911, over 958294.84 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:27:43,829 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.580e+02 1.900e+02 2.443e+02 4.290e+02, threshold=3.800e+02, percent-clipped=1.0 +2023-04-27 01:28:09,958 INFO [finetune.py:976] (3/7) Epoch 11, batch 2500, loss[loss=0.1395, simple_loss=0.212, pruned_loss=0.03355, over 4823.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.255, pruned_loss=0.06129, over 960239.85 frames. ], batch size: 30, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:28:44,933 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:28:54,834 INFO [finetune.py:976] (3/7) Epoch 11, batch 2550, loss[loss=0.2153, simple_loss=0.2842, pruned_loss=0.07322, over 4870.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2582, pruned_loss=0.06144, over 958942.50 frames. ], batch size: 44, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:29:12,510 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.688e+02 2.032e+02 2.494e+02 8.629e+02, threshold=4.065e+02, percent-clipped=5.0 +2023-04-27 01:29:21,645 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 01:29:23,391 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:29:35,316 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59858.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:29:46,963 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59868.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:29:57,640 INFO [finetune.py:976] (3/7) Epoch 11, batch 2600, loss[loss=0.1823, simple_loss=0.246, pruned_loss=0.05923, over 4815.00 frames. ], tot_loss[loss=0.192, simple_loss=0.26, pruned_loss=0.06204, over 958703.11 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:30:10,803 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59889.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:30:29,481 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:30:51,535 INFO [finetune.py:976] (3/7) Epoch 11, batch 2650, loss[loss=0.1965, simple_loss=0.2695, pruned_loss=0.06178, over 4810.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2613, pruned_loss=0.06273, over 956013.90 frames. ], batch size: 41, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:31:03,339 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2912, 3.3386, 2.5382, 3.8540, 3.3991, 3.3062, 1.5433, 3.2389], + device='cuda:3'), covar=tensor([0.2010, 0.1353, 0.3187, 0.2262, 0.3108, 0.2148, 0.5609, 0.2869], + device='cuda:3'), in_proj_covar=tensor([0.0241, 0.0215, 0.0249, 0.0303, 0.0296, 0.0249, 0.0268, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 01:31:04,460 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.693e+02 1.978e+02 2.480e+02 4.060e+02, threshold=3.956e+02, percent-clipped=0.0 +2023-04-27 01:31:21,150 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:31:43,484 INFO [finetune.py:976] (3/7) Epoch 11, batch 2700, loss[loss=0.2043, simple_loss=0.2716, pruned_loss=0.06854, over 4859.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2589, pruned_loss=0.06198, over 952984.72 frames. ], batch size: 34, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:31:44,285 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 01:32:17,130 INFO [finetune.py:976] (3/7) Epoch 11, batch 2750, loss[loss=0.167, simple_loss=0.2246, pruned_loss=0.05474, over 4816.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2551, pruned_loss=0.06077, over 953338.38 frames. ], batch size: 38, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:32:24,354 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.196e+02 1.624e+02 1.944e+02 2.490e+02 4.678e+02, threshold=3.889e+02, percent-clipped=1.0 +2023-04-27 01:32:25,703 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:32:33,992 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60052.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:32:42,800 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60065.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:32:50,410 INFO [finetune.py:976] (3/7) Epoch 11, batch 2800, loss[loss=0.1314, simple_loss=0.2018, pruned_loss=0.03055, over 4777.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2525, pruned_loss=0.06001, over 954257.29 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:32:51,090 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:33:05,643 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60101.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 01:33:14,371 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60113.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 01:33:22,792 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60126.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:33:23,286 INFO [finetune.py:976] (3/7) Epoch 11, batch 2850, loss[loss=0.185, simple_loss=0.2533, pruned_loss=0.05837, over 4796.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2514, pruned_loss=0.05985, over 954367.80 frames. ], batch size: 29, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:33:29,986 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.671e+02 1.998e+02 2.360e+02 5.267e+02, threshold=3.997e+02, percent-clipped=5.0 +2023-04-27 01:33:30,724 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60139.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:33:43,075 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:33:55,930 INFO [finetune.py:976] (3/7) Epoch 11, batch 2900, loss[loss=0.2404, simple_loss=0.3036, pruned_loss=0.08859, over 4807.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2546, pruned_loss=0.06081, over 955059.09 frames. ], batch size: 45, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:33:58,073 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 01:34:15,775 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60198.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:34:25,150 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60204.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:34:26,338 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:34:59,279 INFO [finetune.py:976] (3/7) Epoch 11, batch 2950, loss[loss=0.174, simple_loss=0.2374, pruned_loss=0.05534, over 4232.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2572, pruned_loss=0.06097, over 954489.48 frames. ], batch size: 18, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:35:03,058 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2167, 1.6985, 2.1140, 2.4135, 2.0121, 1.6679, 1.1009, 1.8079], + device='cuda:3'), covar=tensor([0.3715, 0.3603, 0.1810, 0.2531, 0.3265, 0.2932, 0.4916, 0.2453], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0249, 0.0221, 0.0316, 0.0215, 0.0228, 0.0231, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 01:35:12,692 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.759e+01 1.634e+02 1.896e+02 2.531e+02 4.222e+02, threshold=3.792e+02, percent-clipped=1.0 +2023-04-27 01:35:21,698 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60245.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:35:36,388 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60259.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:36:05,412 INFO [finetune.py:976] (3/7) Epoch 11, batch 3000, loss[loss=0.1816, simple_loss=0.2602, pruned_loss=0.05153, over 4917.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.259, pruned_loss=0.06159, over 956634.18 frames. ], batch size: 42, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:36:05,412 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 01:36:27,793 INFO [finetune.py:1010] (3/7) Epoch 11, validation: loss=0.1531, simple_loss=0.2255, pruned_loss=0.04032, over 2265189.00 frames. +2023-04-27 01:36:27,793 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 01:37:31,378 INFO [finetune.py:976] (3/7) Epoch 11, batch 3050, loss[loss=0.2063, simple_loss=0.2715, pruned_loss=0.07051, over 4855.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2611, pruned_loss=0.0621, over 958302.00 frames. ], batch size: 31, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:37:43,036 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 01:37:45,692 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.734e+02 2.169e+02 2.489e+02 4.339e+02, threshold=4.338e+02, percent-clipped=2.0 +2023-04-27 01:38:26,535 INFO [finetune.py:976] (3/7) Epoch 11, batch 3100, loss[loss=0.1831, simple_loss=0.2589, pruned_loss=0.05366, over 4769.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2582, pruned_loss=0.06105, over 957670.95 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:38:27,807 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8509, 1.3973, 1.4647, 1.5648, 1.9935, 1.6854, 1.3936, 1.3450], + device='cuda:3'), covar=tensor([0.1647, 0.1572, 0.2058, 0.1452, 0.1025, 0.1586, 0.2360, 0.2325], + device='cuda:3'), in_proj_covar=tensor([0.0304, 0.0319, 0.0350, 0.0297, 0.0334, 0.0318, 0.0304, 0.0360], + device='cuda:3'), out_proj_covar=tensor([6.3760e-05, 6.7513e-05, 7.5091e-05, 6.1023e-05, 6.9832e-05, 6.7815e-05, + 6.5012e-05, 7.7292e-05], device='cuda:3') +2023-04-27 01:38:40,717 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60396.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 01:38:48,251 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60408.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 01:38:50,811 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 01:38:56,656 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60421.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:39:00,267 INFO [finetune.py:976] (3/7) Epoch 11, batch 3150, loss[loss=0.205, simple_loss=0.2613, pruned_loss=0.07431, over 4805.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2555, pruned_loss=0.06088, over 957146.27 frames. ], batch size: 25, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:39:06,446 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60434.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:39:09,789 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.640e+02 1.939e+02 2.354e+02 4.857e+02, threshold=3.879e+02, percent-clipped=2.0 +2023-04-27 01:39:22,333 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 01:39:26,223 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0605, 4.3751, 0.9236, 2.0679, 2.5556, 2.7669, 2.5102, 0.9886], + device='cuda:3'), covar=tensor([0.1311, 0.0829, 0.2309, 0.1327, 0.0924, 0.1075, 0.1413, 0.2059], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0249, 0.0142, 0.0122, 0.0134, 0.0154, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:39:28,658 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1601, 1.4615, 1.3455, 1.7505, 1.5879, 1.6334, 1.3383, 3.0927], + device='cuda:3'), covar=tensor([0.0774, 0.1075, 0.1060, 0.1333, 0.0828, 0.0635, 0.0977, 0.0231], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 01:39:33,961 INFO [finetune.py:976] (3/7) Epoch 11, batch 3200, loss[loss=0.1803, simple_loss=0.2428, pruned_loss=0.05889, over 4900.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2517, pruned_loss=0.05931, over 958261.17 frames. ], batch size: 32, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:39:53,425 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60504.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:40:07,940 INFO [finetune.py:976] (3/7) Epoch 11, batch 3250, loss[loss=0.2457, simple_loss=0.3158, pruned_loss=0.08782, over 4844.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2532, pruned_loss=0.06061, over 954054.17 frames. ], batch size: 47, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:40:11,837 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 01:40:15,719 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.675e+02 1.982e+02 2.288e+02 4.621e+02, threshold=3.964e+02, percent-clipped=1.0 +2023-04-27 01:40:20,967 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60545.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:40:25,630 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60552.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:40:26,841 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60554.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:40:52,022 INFO [finetune.py:976] (3/7) Epoch 11, batch 3300, loss[loss=0.1839, simple_loss=0.2561, pruned_loss=0.05586, over 4179.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2564, pruned_loss=0.06192, over 951135.04 frames. ], batch size: 65, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:41:14,046 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60593.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:41:58,104 INFO [finetune.py:976] (3/7) Epoch 11, batch 3350, loss[loss=0.1472, simple_loss=0.2131, pruned_loss=0.04062, over 4720.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2576, pruned_loss=0.06202, over 953192.04 frames. ], batch size: 23, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:42:11,849 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.646e+02 2.104e+02 2.646e+02 5.419e+02, threshold=4.209e+02, percent-clipped=4.0 +2023-04-27 01:42:52,676 INFO [finetune.py:976] (3/7) Epoch 11, batch 3400, loss[loss=0.1962, simple_loss=0.2747, pruned_loss=0.05884, over 4913.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2594, pruned_loss=0.06262, over 953252.51 frames. ], batch size: 37, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:43:05,429 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60696.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:14,639 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60708.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:20,709 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:22,445 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60721.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:26,001 INFO [finetune.py:976] (3/7) Epoch 11, batch 3450, loss[loss=0.1873, simple_loss=0.2512, pruned_loss=0.06171, over 4759.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2596, pruned_loss=0.06226, over 953851.96 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:43:30,797 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60734.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:33,700 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.685e+02 2.029e+02 2.444e+02 3.873e+02, threshold=4.057e+02, percent-clipped=0.0 +2023-04-27 01:43:36,747 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60744.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:44,952 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1266, 2.0331, 2.3760, 2.6205, 1.9310, 1.6544, 1.9425, 1.1008], + device='cuda:3'), covar=tensor([0.0664, 0.0927, 0.0633, 0.0917, 0.0948, 0.1478, 0.1090, 0.1153], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0072, 0.0071, 0.0068, 0.0076, 0.0098, 0.0077, 0.0072], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:43:45,936 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60756.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:54,319 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60769.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:59,108 INFO [finetune.py:976] (3/7) Epoch 11, batch 3500, loss[loss=0.1603, simple_loss=0.228, pruned_loss=0.04635, over 4938.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2575, pruned_loss=0.06166, over 955400.49 frames. ], batch size: 38, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:44:00,470 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60779.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:44:02,191 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60782.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:44:32,809 INFO [finetune.py:976] (3/7) Epoch 11, batch 3550, loss[loss=0.1437, simple_loss=0.2165, pruned_loss=0.03543, over 4826.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2548, pruned_loss=0.06115, over 955626.52 frames. ], batch size: 41, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:44:40,090 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.561e+02 1.887e+02 2.297e+02 4.767e+02, threshold=3.774e+02, percent-clipped=1.0 +2023-04-27 01:44:50,249 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60854.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:45:03,850 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.90 vs. limit=5.0 +2023-04-27 01:45:06,040 INFO [finetune.py:976] (3/7) Epoch 11, batch 3600, loss[loss=0.2636, simple_loss=0.2985, pruned_loss=0.1143, over 4193.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2528, pruned_loss=0.06099, over 953200.58 frames. ], batch size: 65, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:45:21,747 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60902.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:45:39,841 INFO [finetune.py:976] (3/7) Epoch 11, batch 3650, loss[loss=0.2116, simple_loss=0.2881, pruned_loss=0.06755, over 4924.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2548, pruned_loss=0.06177, over 952966.98 frames. ], batch size: 42, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:45:47,183 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.725e+02 2.086e+02 2.393e+02 4.773e+02, threshold=4.172e+02, percent-clipped=3.0 +2023-04-27 01:46:05,992 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8625, 2.5731, 1.9428, 1.9345, 1.4581, 1.4812, 2.0336, 1.4710], + device='cuda:3'), covar=tensor([0.1478, 0.1367, 0.1407, 0.1663, 0.2199, 0.1711, 0.0901, 0.1800], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0205, 0.0203, 0.0184, 0.0158, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 01:46:25,827 INFO [finetune.py:976] (3/7) Epoch 11, batch 3700, loss[loss=0.2007, simple_loss=0.2768, pruned_loss=0.06229, over 4140.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2577, pruned_loss=0.06219, over 952084.20 frames. ], batch size: 65, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:46:28,428 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.44 vs. limit=5.0 +2023-04-27 01:46:52,270 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-27 01:47:05,592 INFO [finetune.py:976] (3/7) Epoch 11, batch 3750, loss[loss=0.2198, simple_loss=0.2908, pruned_loss=0.07433, over 4820.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2583, pruned_loss=0.06202, over 952392.42 frames. ], batch size: 39, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:47:17,586 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5594, 1.3036, 4.2575, 3.9618, 3.7410, 3.9711, 3.9270, 3.8066], + device='cuda:3'), covar=tensor([0.7228, 0.5610, 0.1046, 0.1709, 0.1034, 0.1430, 0.1920, 0.1471], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0311, 0.0406, 0.0410, 0.0353, 0.0410, 0.0317, 0.0371], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:47:18,726 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.566e+02 1.879e+02 2.369e+02 3.505e+02, threshold=3.758e+02, percent-clipped=0.0 +2023-04-27 01:47:40,522 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61061.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:48:01,152 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61074.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:48:05,284 INFO [finetune.py:976] (3/7) Epoch 11, batch 3800, loss[loss=0.1918, simple_loss=0.2493, pruned_loss=0.06712, over 4862.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2593, pruned_loss=0.06239, over 951473.92 frames. ], batch size: 34, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:48:56,484 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61122.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:48:59,870 INFO [finetune.py:976] (3/7) Epoch 11, batch 3850, loss[loss=0.1628, simple_loss=0.2347, pruned_loss=0.04541, over 4848.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2573, pruned_loss=0.06127, over 952701.17 frames. ], batch size: 44, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:49:08,085 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.671e+02 1.909e+02 2.247e+02 3.528e+02, threshold=3.817e+02, percent-clipped=0.0 +2023-04-27 01:49:27,490 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-04-27 01:49:33,126 INFO [finetune.py:976] (3/7) Epoch 11, batch 3900, loss[loss=0.1674, simple_loss=0.2397, pruned_loss=0.04759, over 4779.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2555, pruned_loss=0.06112, over 953788.36 frames. ], batch size: 29, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:49:42,532 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5781, 3.4471, 2.5863, 4.1418, 3.6087, 3.5997, 1.4645, 3.5554], + device='cuda:3'), covar=tensor([0.1885, 0.1457, 0.3340, 0.1969, 0.3243, 0.1796, 0.5876, 0.2367], + device='cuda:3'), in_proj_covar=tensor([0.0238, 0.0213, 0.0245, 0.0300, 0.0292, 0.0245, 0.0265, 0.0264], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 01:50:04,196 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5239, 1.9031, 2.3587, 2.9790, 2.2501, 1.9138, 1.7193, 2.1085], + device='cuda:3'), covar=tensor([0.3209, 0.3595, 0.1654, 0.2405, 0.3107, 0.2791, 0.4259, 0.2535], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0246, 0.0219, 0.0313, 0.0212, 0.0225, 0.0229, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 01:50:05,882 INFO [finetune.py:976] (3/7) Epoch 11, batch 3950, loss[loss=0.1768, simple_loss=0.2479, pruned_loss=0.05283, over 4711.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2519, pruned_loss=0.05971, over 954515.99 frames. ], batch size: 59, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:50:15,547 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.581e+02 1.829e+02 2.221e+02 4.088e+02, threshold=3.658e+02, percent-clipped=2.0 +2023-04-27 01:50:39,622 INFO [finetune.py:976] (3/7) Epoch 11, batch 4000, loss[loss=0.1505, simple_loss=0.2015, pruned_loss=0.04977, over 4030.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2516, pruned_loss=0.06019, over 949541.84 frames. ], batch size: 17, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:50:48,467 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1298, 2.7458, 2.1252, 2.4597, 1.8610, 2.1721, 2.1934, 1.6129], + device='cuda:3'), covar=tensor([0.1807, 0.1079, 0.0849, 0.1141, 0.2984, 0.1239, 0.1938, 0.2792], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0313, 0.0226, 0.0283, 0.0312, 0.0270, 0.0253, 0.0274], + device='cuda:3'), out_proj_covar=tensor([1.1767e-04, 1.2575e-04, 9.0340e-05, 1.1311e-04, 1.2760e-04, 1.0838e-04, + 1.0324e-04, 1.0989e-04], device='cuda:3') +2023-04-27 01:50:55,010 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5000, 1.3565, 0.6538, 1.2294, 1.4046, 1.3875, 1.2936, 1.2865], + device='cuda:3'), covar=tensor([0.0572, 0.0419, 0.0410, 0.0611, 0.0304, 0.0569, 0.0545, 0.0646], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0037, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 01:51:18,202 INFO [finetune.py:976] (3/7) Epoch 11, batch 4050, loss[loss=0.1856, simple_loss=0.2621, pruned_loss=0.05459, over 4779.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2544, pruned_loss=0.06038, over 949322.30 frames. ], batch size: 29, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:51:37,047 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.723e+02 1.975e+02 2.601e+02 5.201e+02, threshold=3.950e+02, percent-clipped=3.0 +2023-04-27 01:51:59,610 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6116, 1.5068, 1.8767, 1.8555, 1.4110, 1.3275, 1.5506, 1.0241], + device='cuda:3'), covar=tensor([0.0602, 0.0729, 0.0484, 0.0726, 0.0881, 0.1275, 0.0646, 0.0768], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0071, 0.0071, 0.0067, 0.0075, 0.0097, 0.0076, 0.0072], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:52:11,032 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4915, 0.6602, 1.3126, 1.7649, 1.5863, 1.4212, 1.3990, 1.4497], + device='cuda:3'), covar=tensor([0.4884, 0.6173, 0.6017, 0.6764, 0.6142, 0.7435, 0.7458, 0.6758], + device='cuda:3'), in_proj_covar=tensor([0.0412, 0.0412, 0.0499, 0.0518, 0.0441, 0.0461, 0.0471, 0.0471], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 01:52:21,696 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:52:23,439 INFO [finetune.py:976] (3/7) Epoch 11, batch 4100, loss[loss=0.1605, simple_loss=0.2348, pruned_loss=0.04305, over 4807.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2566, pruned_loss=0.06111, over 947623.12 frames. ], batch size: 45, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:53:16,502 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61417.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:53:25,490 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=61422.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:53:28,494 INFO [finetune.py:976] (3/7) Epoch 11, batch 4150, loss[loss=0.189, simple_loss=0.2604, pruned_loss=0.0588, over 4833.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2579, pruned_loss=0.06154, over 948414.50 frames. ], batch size: 49, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:53:48,240 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.660e+02 1.903e+02 2.358e+02 4.640e+02, threshold=3.807e+02, percent-clipped=2.0 +2023-04-27 01:54:37,122 INFO [finetune.py:976] (3/7) Epoch 11, batch 4200, loss[loss=0.1842, simple_loss=0.2538, pruned_loss=0.05726, over 4804.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2592, pruned_loss=0.06168, over 951511.06 frames. ], batch size: 40, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:55:45,102 INFO [finetune.py:976] (3/7) Epoch 11, batch 4250, loss[loss=0.1477, simple_loss=0.2163, pruned_loss=0.0396, over 4693.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2559, pruned_loss=0.06056, over 952590.99 frames. ], batch size: 59, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:55:45,853 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8226, 2.4197, 1.8820, 1.6882, 1.3599, 1.3804, 1.8412, 1.2827], + device='cuda:3'), covar=tensor([0.1877, 0.1480, 0.1616, 0.2050, 0.2652, 0.2166, 0.1190, 0.2274], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0204, 0.0203, 0.0184, 0.0158, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 01:55:57,237 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.633e+02 1.870e+02 2.370e+02 4.371e+02, threshold=3.739e+02, percent-clipped=2.0 +2023-04-27 01:56:05,328 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 01:56:49,823 INFO [finetune.py:976] (3/7) Epoch 11, batch 4300, loss[loss=0.2182, simple_loss=0.2845, pruned_loss=0.0759, over 4824.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2523, pruned_loss=0.05926, over 951745.89 frames. ], batch size: 39, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:57:48,688 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3366, 2.8419, 1.0973, 1.8173, 1.8251, 2.1961, 1.8741, 1.1796], + device='cuda:3'), covar=tensor([0.1191, 0.0896, 0.1690, 0.1108, 0.0941, 0.0831, 0.1266, 0.1713], + device='cuda:3'), in_proj_covar=tensor([0.0119, 0.0250, 0.0141, 0.0123, 0.0135, 0.0154, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 01:57:57,465 INFO [finetune.py:976] (3/7) Epoch 11, batch 4350, loss[loss=0.1747, simple_loss=0.2484, pruned_loss=0.05048, over 4914.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2501, pruned_loss=0.0583, over 950673.47 frames. ], batch size: 36, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:58:10,462 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.619e+02 1.912e+02 2.174e+02 4.082e+02, threshold=3.823e+02, percent-clipped=2.0 +2023-04-27 01:59:02,124 INFO [finetune.py:976] (3/7) Epoch 11, batch 4400, loss[loss=0.2134, simple_loss=0.2615, pruned_loss=0.08268, over 4422.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2515, pruned_loss=0.05937, over 951379.57 frames. ], batch size: 19, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:59:11,959 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7841, 3.7652, 2.7438, 4.4879, 3.9342, 3.7819, 1.5972, 3.7983], + device='cuda:3'), covar=tensor([0.1722, 0.1251, 0.3355, 0.1364, 0.3590, 0.1735, 0.6021, 0.2454], + device='cuda:3'), in_proj_covar=tensor([0.0238, 0.0213, 0.0246, 0.0300, 0.0292, 0.0246, 0.0265, 0.0265], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 01:59:23,249 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7247, 2.1470, 0.9828, 1.0396, 1.5287, 1.0446, 2.4582, 1.1607], + device='cuda:3'), covar=tensor([0.0784, 0.0708, 0.0661, 0.1339, 0.0487, 0.1115, 0.0353, 0.0802], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0077, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 01:59:57,239 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61717.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:00:05,324 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 02:00:08,835 INFO [finetune.py:976] (3/7) Epoch 11, batch 4450, loss[loss=0.1873, simple_loss=0.2586, pruned_loss=0.05798, over 4857.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2555, pruned_loss=0.06046, over 952447.80 frames. ], batch size: 49, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:00:17,364 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61732.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:00:21,986 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.778e+02 2.036e+02 2.485e+02 5.642e+02, threshold=4.071e+02, percent-clipped=3.0 +2023-04-27 02:00:24,577 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61743.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:00:29,424 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.44 vs. limit=5.0 +2023-04-27 02:00:43,729 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8109, 2.2682, 1.1444, 1.6122, 2.2468, 1.6885, 1.6159, 1.6967], + device='cuda:3'), covar=tensor([0.0633, 0.0306, 0.0329, 0.0589, 0.0249, 0.0633, 0.0621, 0.0620], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 02:00:45,396 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=61765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:01:03,256 INFO [finetune.py:976] (3/7) Epoch 11, batch 4500, loss[loss=0.199, simple_loss=0.2694, pruned_loss=0.06437, over 4878.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2568, pruned_loss=0.0609, over 951158.87 frames. ], batch size: 34, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:01:18,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7489, 3.8522, 2.6699, 4.4141, 3.8875, 3.7806, 1.5451, 3.8289], + device='cuda:3'), covar=tensor([0.1733, 0.1093, 0.3138, 0.1620, 0.2906, 0.1686, 0.5919, 0.2347], + device='cuda:3'), in_proj_covar=tensor([0.0240, 0.0215, 0.0249, 0.0303, 0.0295, 0.0247, 0.0266, 0.0267], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:01:19,121 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61793.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:01:26,322 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61804.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:01:42,634 INFO [finetune.py:976] (3/7) Epoch 11, batch 4550, loss[loss=0.1607, simple_loss=0.2274, pruned_loss=0.04698, over 4738.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2579, pruned_loss=0.06178, over 951425.28 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:01:49,952 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.694e+02 2.002e+02 2.452e+02 5.471e+02, threshold=4.003e+02, percent-clipped=1.0 +2023-04-27 02:01:58,484 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6094, 2.1943, 2.6148, 3.1314, 2.5067, 2.1257, 2.0305, 2.3806], + device='cuda:3'), covar=tensor([0.3464, 0.3215, 0.1557, 0.2088, 0.2766, 0.2634, 0.3764, 0.2134], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0249, 0.0221, 0.0316, 0.0214, 0.0227, 0.0232, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 02:02:04,865 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-04-27 02:02:16,335 INFO [finetune.py:976] (3/7) Epoch 11, batch 4600, loss[loss=0.1628, simple_loss=0.2428, pruned_loss=0.04144, over 4900.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2573, pruned_loss=0.06103, over 951351.02 frames. ], batch size: 37, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:02:37,779 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0971, 1.6501, 1.5640, 1.7792, 1.7719, 1.9130, 1.4426, 3.4851], + device='cuda:3'), covar=tensor([0.0612, 0.0764, 0.0723, 0.1160, 0.0591, 0.0527, 0.0725, 0.0154], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:02:40,209 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61914.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:02:45,144 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-27 02:02:49,393 INFO [finetune.py:976] (3/7) Epoch 11, batch 4650, loss[loss=0.214, simple_loss=0.2784, pruned_loss=0.07478, over 4848.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2554, pruned_loss=0.06057, over 953367.41 frames. ], batch size: 44, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:02:56,645 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.191e+02 1.561e+02 1.984e+02 2.276e+02 4.966e+02, threshold=3.968e+02, percent-clipped=2.0 +2023-04-27 02:03:20,497 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61975.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:03:22,014 INFO [finetune.py:976] (3/7) Epoch 11, batch 4700, loss[loss=0.1897, simple_loss=0.2489, pruned_loss=0.06525, over 4817.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2524, pruned_loss=0.05942, over 954561.23 frames. ], batch size: 40, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:03:33,462 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8136, 2.3155, 1.8153, 1.5887, 1.3250, 1.3306, 1.8986, 1.2673], + device='cuda:3'), covar=tensor([0.1623, 0.1489, 0.1449, 0.1977, 0.2342, 0.1946, 0.1016, 0.2097], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0212, 0.0167, 0.0202, 0.0200, 0.0182, 0.0157, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 02:04:06,619 INFO [finetune.py:976] (3/7) Epoch 11, batch 4750, loss[loss=0.1975, simple_loss=0.2644, pruned_loss=0.06528, over 4809.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2505, pruned_loss=0.05906, over 956173.29 frames. ], batch size: 41, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:04:20,381 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.587e+02 1.879e+02 2.301e+02 4.862e+02, threshold=3.757e+02, percent-clipped=3.0 +2023-04-27 02:04:56,948 INFO [finetune.py:976] (3/7) Epoch 11, batch 4800, loss[loss=0.1976, simple_loss=0.2477, pruned_loss=0.07378, over 4735.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2528, pruned_loss=0.06008, over 957362.10 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:05:05,198 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:05:11,911 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62099.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:05:25,713 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 02:05:27,323 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62122.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:05:30,801 INFO [finetune.py:976] (3/7) Epoch 11, batch 4850, loss[loss=0.1884, simple_loss=0.2653, pruned_loss=0.05571, over 4827.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2549, pruned_loss=0.06065, over 954074.59 frames. ], batch size: 39, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:05:39,088 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.826e+02 2.171e+02 2.650e+02 4.437e+02, threshold=4.341e+02, percent-clipped=4.0 +2023-04-27 02:05:57,610 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 02:06:08,686 INFO [finetune.py:976] (3/7) Epoch 11, batch 4900, loss[loss=0.1762, simple_loss=0.2638, pruned_loss=0.04428, over 4833.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2567, pruned_loss=0.06116, over 954139.31 frames. ], batch size: 47, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:06:15,827 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 02:06:18,148 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62183.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:06:52,813 INFO [finetune.py:976] (3/7) Epoch 11, batch 4950, loss[loss=0.1865, simple_loss=0.2421, pruned_loss=0.06547, over 4705.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2586, pruned_loss=0.06162, over 956356.73 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 64.0 +2023-04-27 02:07:01,571 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.624e+02 1.966e+02 2.483e+02 3.537e+02, threshold=3.932e+02, percent-clipped=0.0 +2023-04-27 02:07:21,452 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62270.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:07:23,328 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:07:26,147 INFO [finetune.py:976] (3/7) Epoch 11, batch 5000, loss[loss=0.1742, simple_loss=0.2333, pruned_loss=0.05759, over 4781.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2567, pruned_loss=0.06107, over 957304.47 frames. ], batch size: 26, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:07:56,979 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 02:07:57,192 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7782, 1.7703, 1.7944, 1.3265, 1.8361, 1.6906, 2.3577, 1.6424], + device='cuda:3'), covar=tensor([0.3678, 0.1871, 0.5122, 0.2775, 0.1667, 0.2175, 0.1690, 0.4858], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0348, 0.0426, 0.0358, 0.0384, 0.0380, 0.0377, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:07:58,883 INFO [finetune.py:976] (3/7) Epoch 11, batch 5050, loss[loss=0.1934, simple_loss=0.2599, pruned_loss=0.06341, over 4780.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2533, pruned_loss=0.05977, over 956744.12 frames. ], batch size: 51, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:08:04,228 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62334.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:08:08,228 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.670e+02 2.040e+02 2.398e+02 4.126e+02, threshold=4.080e+02, percent-clipped=2.0 +2023-04-27 02:08:32,082 INFO [finetune.py:976] (3/7) Epoch 11, batch 5100, loss[loss=0.1655, simple_loss=0.2314, pruned_loss=0.04981, over 3948.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2509, pruned_loss=0.0592, over 954991.62 frames. ], batch size: 17, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:08:37,448 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3410, 3.2041, 1.1064, 1.7742, 1.7254, 2.4492, 1.9275, 1.0460], + device='cuda:3'), covar=tensor([0.1368, 0.0809, 0.1836, 0.1200, 0.1067, 0.0846, 0.1330, 0.2004], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0249, 0.0141, 0.0121, 0.0133, 0.0153, 0.0118, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 02:08:40,382 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62388.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:08:42,183 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62391.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:08:44,823 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.11 vs. limit=5.0 +2023-04-27 02:08:47,993 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62399.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:08:55,581 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 02:09:06,027 INFO [finetune.py:976] (3/7) Epoch 11, batch 5150, loss[loss=0.1803, simple_loss=0.2644, pruned_loss=0.04813, over 4926.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2509, pruned_loss=0.05905, over 951898.87 frames. ], batch size: 38, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:09:12,053 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62436.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:09:14,848 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.955e+01 1.646e+02 2.010e+02 2.557e+02 5.535e+02, threshold=4.020e+02, percent-clipped=1.0 +2023-04-27 02:09:25,278 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62447.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:09:27,125 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8672, 2.8300, 2.3141, 3.2942, 2.8402, 2.8587, 1.3728, 2.7917], + device='cuda:3'), covar=tensor([0.2111, 0.1799, 0.3452, 0.2769, 0.3847, 0.2489, 0.5513, 0.3004], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0218, 0.0252, 0.0306, 0.0299, 0.0249, 0.0270, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:09:33,232 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62452.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:10:01,359 INFO [finetune.py:976] (3/7) Epoch 11, batch 5200, loss[loss=0.2759, simple_loss=0.3301, pruned_loss=0.1108, over 4271.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2564, pruned_loss=0.06113, over 953158.32 frames. ], batch size: 65, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:10:02,042 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62478.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:10:15,654 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-27 02:10:58,657 INFO [finetune.py:976] (3/7) Epoch 11, batch 5250, loss[loss=0.1644, simple_loss=0.2384, pruned_loss=0.0452, over 4931.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2586, pruned_loss=0.06216, over 952757.08 frames. ], batch size: 38, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:10:59,410 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8829, 2.4626, 1.8505, 1.7029, 1.3641, 1.3901, 1.9580, 1.2995], + device='cuda:3'), covar=tensor([0.1770, 0.1515, 0.1614, 0.2020, 0.2674, 0.2039, 0.1189, 0.2169], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0215, 0.0169, 0.0204, 0.0202, 0.0184, 0.0159, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 02:11:07,067 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.221e+02 1.615e+02 2.039e+02 2.344e+02 5.619e+02, threshold=4.078e+02, percent-clipped=3.0 +2023-04-27 02:11:10,699 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62545.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:11:28,236 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62570.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:11:32,451 INFO [finetune.py:976] (3/7) Epoch 11, batch 5300, loss[loss=0.2109, simple_loss=0.284, pruned_loss=0.06894, over 4741.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2598, pruned_loss=0.06176, over 954982.69 frames. ], batch size: 54, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:11:37,356 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62585.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:11:52,276 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62606.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:00,029 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62618.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:05,910 INFO [finetune.py:976] (3/7) Epoch 11, batch 5350, loss[loss=0.2104, simple_loss=0.2801, pruned_loss=0.07036, over 4910.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2583, pruned_loss=0.06124, over 953842.63 frames. ], batch size: 43, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:12:07,195 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62629.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:13,875 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 1.690e+02 2.017e+02 2.486e+02 5.260e+02, threshold=4.033e+02, percent-clipped=4.0 +2023-04-27 02:12:18,128 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62646.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:32,508 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62666.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:39,516 INFO [finetune.py:976] (3/7) Epoch 11, batch 5400, loss[loss=0.1675, simple_loss=0.2375, pruned_loss=0.04879, over 4898.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2564, pruned_loss=0.06053, over 954799.20 frames. ], batch size: 35, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:13:00,543 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 02:13:12,193 INFO [finetune.py:976] (3/7) Epoch 11, batch 5450, loss[loss=0.1398, simple_loss=0.2076, pruned_loss=0.03598, over 4827.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2541, pruned_loss=0.05988, over 953211.34 frames. ], batch size: 51, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:13:12,304 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62727.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:13:20,544 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.672e+02 1.919e+02 2.198e+02 3.769e+02, threshold=3.837e+02, percent-clipped=0.0 +2023-04-27 02:13:22,499 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1275, 1.4867, 1.3136, 1.7261, 1.6170, 1.7787, 1.3442, 3.0774], + device='cuda:3'), covar=tensor([0.0677, 0.0815, 0.0877, 0.1260, 0.0632, 0.0468, 0.0780, 0.0198], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:13:24,900 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62747.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:13:35,004 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7584, 2.2083, 1.6220, 1.3304, 1.2825, 1.3129, 1.6931, 1.2225], + device='cuda:3'), covar=tensor([0.1816, 0.1401, 0.1612, 0.2066, 0.2700, 0.2152, 0.1223, 0.2247], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0213, 0.0168, 0.0202, 0.0201, 0.0182, 0.0157, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 02:13:42,666 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62773.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:13:45,485 INFO [finetune.py:976] (3/7) Epoch 11, batch 5500, loss[loss=0.2025, simple_loss=0.2589, pruned_loss=0.07309, over 4829.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2518, pruned_loss=0.05931, over 955395.78 frames. ], batch size: 40, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:13:46,159 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62778.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:13:49,669 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3069, 1.4972, 1.4534, 1.8120, 1.6811, 2.0581, 1.3467, 3.6167], + device='cuda:3'), covar=tensor([0.0654, 0.0843, 0.0858, 0.1241, 0.0658, 0.0467, 0.0792, 0.0160], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:14:17,365 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6938, 1.2698, 1.9081, 2.1253, 1.7846, 1.6944, 1.8036, 1.7668], + device='cuda:3'), covar=tensor([0.6198, 0.8458, 0.8142, 0.8096, 0.7826, 0.9997, 1.0101, 0.9498], + device='cuda:3'), in_proj_covar=tensor([0.0408, 0.0408, 0.0495, 0.0513, 0.0439, 0.0457, 0.0466, 0.0467], + device='cuda:3'), out_proj_covar=tensor([9.9269e-05, 1.0101e-04, 1.1165e-04, 1.2177e-04, 1.0631e-04, 1.1043e-04, + 1.1162e-04, 1.1197e-04], device='cuda:3') +2023-04-27 02:14:18,370 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62826.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:14:18,931 INFO [finetune.py:976] (3/7) Epoch 11, batch 5550, loss[loss=0.2173, simple_loss=0.2936, pruned_loss=0.0705, over 4854.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2535, pruned_loss=0.0598, over 956515.95 frames. ], batch size: 49, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:14:23,723 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62834.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:14:27,256 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.295e+02 1.658e+02 2.068e+02 2.664e+02 6.166e+02, threshold=4.137e+02, percent-clipped=3.0 +2023-04-27 02:14:30,409 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.5876, 4.4202, 3.1925, 5.2375, 4.4882, 4.5434, 1.8094, 4.5547], + device='cuda:3'), covar=tensor([0.1636, 0.1042, 0.3272, 0.1026, 0.3631, 0.1485, 0.6094, 0.2036], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0216, 0.0248, 0.0304, 0.0297, 0.0246, 0.0268, 0.0267], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:15:05,420 INFO [finetune.py:976] (3/7) Epoch 11, batch 5600, loss[loss=0.1898, simple_loss=0.2606, pruned_loss=0.05944, over 4860.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2566, pruned_loss=0.06054, over 955500.73 frames. ], batch size: 31, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:15:15,914 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7449, 1.0698, 1.5109, 1.6509, 1.6008, 1.7276, 1.4979, 1.5197], + device='cuda:3'), covar=tensor([0.4843, 0.6813, 0.5591, 0.5430, 0.6576, 0.9184, 0.6473, 0.5855], + device='cuda:3'), in_proj_covar=tensor([0.0329, 0.0380, 0.0317, 0.0329, 0.0340, 0.0401, 0.0360, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:15:37,387 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62901.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:16:10,656 INFO [finetune.py:976] (3/7) Epoch 11, batch 5650, loss[loss=0.2086, simple_loss=0.284, pruned_loss=0.06661, over 4907.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2574, pruned_loss=0.06045, over 954953.24 frames. ], batch size: 36, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:16:11,884 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:16:23,345 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.567e+02 1.821e+02 2.303e+02 3.535e+02, threshold=3.642e+02, percent-clipped=0.0 +2023-04-27 02:16:23,980 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:17:01,895 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2802, 1.4419, 1.2475, 1.4210, 1.2088, 1.1368, 1.2800, 1.0236], + device='cuda:3'), covar=tensor([0.1114, 0.1009, 0.0873, 0.0959, 0.2640, 0.1121, 0.1140, 0.1677], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0313, 0.0224, 0.0283, 0.0310, 0.0268, 0.0251, 0.0273], + device='cuda:3'), out_proj_covar=tensor([1.1733e-04, 1.2580e-04, 8.9536e-05, 1.1334e-04, 1.2672e-04, 1.0743e-04, + 1.0234e-04, 1.0925e-04], device='cuda:3') +2023-04-27 02:17:06,580 INFO [finetune.py:976] (3/7) Epoch 11, batch 5700, loss[loss=0.1967, simple_loss=0.2467, pruned_loss=0.07333, over 4086.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2547, pruned_loss=0.0608, over 935202.89 frames. ], batch size: 17, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:17:06,613 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:17:38,299 INFO [finetune.py:976] (3/7) Epoch 12, batch 0, loss[loss=0.1782, simple_loss=0.24, pruned_loss=0.05819, over 4803.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.24, pruned_loss=0.05819, over 4803.00 frames. ], batch size: 25, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:17:38,299 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 02:17:47,678 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1914, 1.3995, 1.6709, 1.8230, 1.7184, 1.8730, 1.6996, 1.7115], + device='cuda:3'), covar=tensor([0.4657, 0.6316, 0.5578, 0.5527, 0.6309, 0.8547, 0.6846, 0.5823], + device='cuda:3'), in_proj_covar=tensor([0.0327, 0.0379, 0.0315, 0.0327, 0.0339, 0.0400, 0.0359, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:17:53,691 INFO [finetune.py:1010] (3/7) Epoch 12, validation: loss=0.1544, simple_loss=0.2267, pruned_loss=0.04099, over 2265189.00 frames. +2023-04-27 02:17:53,692 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 02:18:09,371 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9946, 1.7046, 2.0554, 2.3277, 2.4768, 1.9115, 1.7007, 2.1055], + device='cuda:3'), covar=tensor([0.0910, 0.1199, 0.0641, 0.0585, 0.0566, 0.0929, 0.0838, 0.0639], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0204, 0.0183, 0.0175, 0.0179, 0.0189, 0.0159, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:18:18,026 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63022.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:18:34,068 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9931, 1.7899, 2.1280, 2.4035, 2.4647, 1.9383, 1.6900, 2.2302], + device='cuda:3'), covar=tensor([0.0906, 0.1090, 0.0661, 0.0541, 0.0580, 0.0928, 0.0838, 0.0533], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0204, 0.0183, 0.0175, 0.0179, 0.0189, 0.0159, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:18:39,643 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.931e+01 1.655e+02 2.029e+02 2.560e+02 6.942e+02, threshold=4.058e+02, percent-clipped=5.0 +2023-04-27 02:18:44,982 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63047.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:18:54,574 INFO [finetune.py:976] (3/7) Epoch 12, batch 50, loss[loss=0.2298, simple_loss=0.291, pruned_loss=0.0843, over 4739.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2624, pruned_loss=0.06432, over 215952.43 frames. ], batch size: 26, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:19:30,953 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63095.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:19:31,937 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-27 02:19:37,861 INFO [finetune.py:976] (3/7) Epoch 12, batch 100, loss[loss=0.2087, simple_loss=0.2646, pruned_loss=0.07639, over 4894.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2549, pruned_loss=0.06215, over 377599.13 frames. ], batch size: 35, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:19:54,524 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:20:01,186 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.674e+02 1.936e+02 2.495e+02 3.786e+02, threshold=3.872e+02, percent-clipped=0.0 +2023-04-27 02:20:11,745 INFO [finetune.py:976] (3/7) Epoch 12, batch 150, loss[loss=0.1597, simple_loss=0.2292, pruned_loss=0.04512, over 4780.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2496, pruned_loss=0.05959, over 507445.26 frames. ], batch size: 26, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:20:58,537 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63201.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:01,806 INFO [finetune.py:976] (3/7) Epoch 12, batch 200, loss[loss=0.2311, simple_loss=0.299, pruned_loss=0.08155, over 4752.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2489, pruned_loss=0.05922, over 608787.58 frames. ], batch size: 59, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:21:09,218 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 02:21:13,194 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63221.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:24,494 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.629e+02 1.966e+02 2.303e+02 3.666e+02, threshold=3.932e+02, percent-clipped=0.0 +2023-04-27 02:21:25,235 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63241.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:30,505 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63249.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:34,067 INFO [finetune.py:976] (3/7) Epoch 12, batch 250, loss[loss=0.2184, simple_loss=0.3006, pruned_loss=0.0681, over 4821.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2542, pruned_loss=0.06119, over 687346.00 frames. ], batch size: 40, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:21:52,389 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63282.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:53,368 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 02:21:56,553 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:22:12,121 INFO [finetune.py:976] (3/7) Epoch 12, batch 300, loss[loss=0.2706, simple_loss=0.3126, pruned_loss=0.1143, over 4834.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2577, pruned_loss=0.06207, over 746800.33 frames. ], batch size: 49, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:22:21,947 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6873, 2.2976, 1.6979, 1.5931, 1.2694, 1.2541, 1.7187, 1.1785], + device='cuda:3'), covar=tensor([0.1694, 0.1389, 0.1462, 0.1803, 0.2413, 0.2075, 0.1067, 0.2199], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0204, 0.0203, 0.0185, 0.0159, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 02:22:36,082 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63322.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:22:46,894 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.771e+02 2.081e+02 2.585e+02 5.314e+02, threshold=4.161e+02, percent-clipped=4.0 +2023-04-27 02:23:01,681 INFO [finetune.py:976] (3/7) Epoch 12, batch 350, loss[loss=0.2008, simple_loss=0.2722, pruned_loss=0.06473, over 4831.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2588, pruned_loss=0.0621, over 794023.94 frames. ], batch size: 47, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:23:19,007 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63370.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:23:23,113 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0559, 1.6947, 2.2630, 2.4950, 2.1211, 1.9709, 2.1033, 2.0805], + device='cuda:3'), covar=tensor([0.6142, 0.8407, 0.8336, 0.7662, 0.7496, 1.0606, 1.0598, 0.9741], + device='cuda:3'), in_proj_covar=tensor([0.0407, 0.0408, 0.0494, 0.0512, 0.0438, 0.0457, 0.0465, 0.0467], + device='cuda:3'), out_proj_covar=tensor([9.9064e-05, 1.0089e-04, 1.1133e-04, 1.2155e-04, 1.0619e-04, 1.1052e-04, + 1.1153e-04, 1.1192e-04], device='cuda:3') +2023-04-27 02:23:57,900 INFO [finetune.py:976] (3/7) Epoch 12, batch 400, loss[loss=0.1629, simple_loss=0.2336, pruned_loss=0.04609, over 4923.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2597, pruned_loss=0.06195, over 830070.94 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:23:59,852 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1615, 1.4572, 1.3485, 1.7714, 1.6694, 1.6649, 1.3594, 3.1267], + device='cuda:3'), covar=tensor([0.0715, 0.0816, 0.0870, 0.1182, 0.0649, 0.0554, 0.0788, 0.0201], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:24:11,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0061, 1.6413, 2.0804, 2.3450, 2.0824, 1.9557, 2.0118, 2.0213], + device='cuda:3'), covar=tensor([0.5141, 0.6617, 0.7130, 0.7564, 0.6964, 0.9116, 0.8280, 0.7445], + device='cuda:3'), in_proj_covar=tensor([0.0408, 0.0408, 0.0495, 0.0513, 0.0439, 0.0458, 0.0467, 0.0468], + device='cuda:3'), out_proj_covar=tensor([9.9262e-05, 1.0106e-04, 1.1163e-04, 1.2186e-04, 1.0633e-04, 1.1067e-04, + 1.1191e-04, 1.1214e-04], device='cuda:3') +2023-04-27 02:24:18,826 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63420.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:24:21,002 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63422.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:24:31,388 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63429.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:24:43,213 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.266e+02 1.590e+02 1.875e+02 2.265e+02 5.610e+02, threshold=3.751e+02, percent-clipped=2.0 +2023-04-27 02:25:03,612 INFO [finetune.py:976] (3/7) Epoch 12, batch 450, loss[loss=0.1824, simple_loss=0.2393, pruned_loss=0.06277, over 4193.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.259, pruned_loss=0.06211, over 856547.47 frames. ], batch size: 18, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:25:38,226 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63477.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:25:46,556 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63481.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:25:47,779 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:26:01,235 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2190, 1.5122, 1.4190, 1.7841, 1.6874, 1.8197, 1.3732, 3.5362], + device='cuda:3'), covar=tensor([0.0651, 0.0857, 0.0865, 0.1224, 0.0661, 0.0533, 0.0832, 0.0184], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:26:12,258 INFO [finetune.py:976] (3/7) Epoch 12, batch 500, loss[loss=0.1242, simple_loss=0.1931, pruned_loss=0.02764, over 4763.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2559, pruned_loss=0.06088, over 877875.10 frames. ], batch size: 27, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:26:40,116 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6928, 3.6870, 2.7224, 4.2650, 3.6867, 3.6818, 1.7982, 3.7113], + device='cuda:3'), covar=tensor([0.1738, 0.1120, 0.2739, 0.1853, 0.3201, 0.1825, 0.5440, 0.2163], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0215, 0.0248, 0.0302, 0.0296, 0.0246, 0.0269, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:26:41,243 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.737e+01 1.732e+02 1.941e+02 2.281e+02 3.295e+02, threshold=3.881e+02, percent-clipped=0.0 +2023-04-27 02:26:50,387 INFO [finetune.py:976] (3/7) Epoch 12, batch 550, loss[loss=0.1645, simple_loss=0.2353, pruned_loss=0.0469, over 4894.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2542, pruned_loss=0.06024, over 898036.77 frames. ], batch size: 32, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:26:59,492 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63569.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:05,807 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:12,204 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63586.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:20,087 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4414, 1.3095, 1.8343, 1.7207, 1.2667, 1.1265, 1.5020, 0.9206], + device='cuda:3'), covar=tensor([0.0693, 0.0897, 0.0456, 0.0807, 0.0949, 0.1281, 0.0765, 0.0852], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0072, 0.0070, 0.0067, 0.0075, 0.0097, 0.0076, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 02:27:23,757 INFO [finetune.py:976] (3/7) Epoch 12, batch 600, loss[loss=0.1482, simple_loss=0.2177, pruned_loss=0.03932, over 4706.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.255, pruned_loss=0.06064, over 910063.53 frames. ], batch size: 23, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:27:27,503 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3173, 2.7618, 0.8520, 1.4905, 2.1365, 1.4555, 4.0191, 1.8758], + device='cuda:3'), covar=tensor([0.0684, 0.0874, 0.0944, 0.1311, 0.0551, 0.1021, 0.0191, 0.0662], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 02:27:28,766 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8418, 1.3366, 1.6862, 1.6537, 1.6105, 1.3122, 0.7120, 1.3100], + device='cuda:3'), covar=tensor([0.3815, 0.3805, 0.1894, 0.2666, 0.3048, 0.2998, 0.4831, 0.2525], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0249, 0.0222, 0.0315, 0.0214, 0.0228, 0.0231, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 02:27:41,085 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63630.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:48,534 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.750e+02 2.019e+02 2.576e+02 5.185e+02, threshold=4.039e+02, percent-clipped=2.0 +2023-04-27 02:27:52,959 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63647.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:57,757 INFO [finetune.py:976] (3/7) Epoch 12, batch 650, loss[loss=0.1674, simple_loss=0.2336, pruned_loss=0.05059, over 4796.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2569, pruned_loss=0.06132, over 917446.20 frames. ], batch size: 25, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:28:03,355 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 02:28:08,663 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5095, 1.3991, 1.6986, 1.8590, 1.4018, 1.0565, 1.4699, 0.9860], + device='cuda:3'), covar=tensor([0.0637, 0.0793, 0.0538, 0.0651, 0.0803, 0.1603, 0.0862, 0.0918], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0072, 0.0071, 0.0067, 0.0076, 0.0097, 0.0076, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 02:28:21,614 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9642, 2.4302, 0.9694, 1.2254, 1.7144, 1.1498, 3.0337, 1.5117], + device='cuda:3'), covar=tensor([0.0687, 0.0559, 0.0720, 0.1261, 0.0493, 0.0994, 0.0234, 0.0682], + device='cuda:3'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 02:28:42,264 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-04-27 02:28:42,664 INFO [finetune.py:976] (3/7) Epoch 12, batch 700, loss[loss=0.177, simple_loss=0.2439, pruned_loss=0.05508, over 4751.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.258, pruned_loss=0.06113, over 926156.32 frames. ], batch size: 28, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:29:16,503 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7818, 2.3514, 1.6773, 1.5263, 1.3144, 1.3320, 1.6718, 1.2050], + device='cuda:3'), covar=tensor([0.1753, 0.1452, 0.1641, 0.1996, 0.2438, 0.2142, 0.1147, 0.2174], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0213, 0.0169, 0.0203, 0.0201, 0.0183, 0.0158, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 02:29:23,284 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.645e+02 1.982e+02 2.413e+02 5.979e+02, threshold=3.963e+02, percent-clipped=2.0 +2023-04-27 02:29:32,956 INFO [finetune.py:976] (3/7) Epoch 12, batch 750, loss[loss=0.1868, simple_loss=0.2776, pruned_loss=0.04795, over 4782.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2597, pruned_loss=0.06162, over 933595.35 frames. ], batch size: 29, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:29:45,879 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63776.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:29:47,597 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63778.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:30:06,647 INFO [finetune.py:976] (3/7) Epoch 12, batch 800, loss[loss=0.1961, simple_loss=0.2489, pruned_loss=0.07165, over 4892.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2587, pruned_loss=0.06099, over 939405.60 frames. ], batch size: 32, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:30:09,843 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63810.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:30:37,955 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8034, 1.9612, 1.2001, 1.4150, 2.1981, 1.7405, 1.5508, 1.5681], + device='cuda:3'), covar=tensor([0.0467, 0.0344, 0.0300, 0.0529, 0.0230, 0.0500, 0.0475, 0.0553], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 02:30:40,781 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.251e+02 1.671e+02 1.936e+02 2.427e+02 3.896e+02, threshold=3.873e+02, percent-clipped=0.0 +2023-04-27 02:30:49,745 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63845.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:30:52,166 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3900, 3.3114, 0.8845, 1.8716, 1.8418, 2.2063, 1.8632, 0.8849], + device='cuda:3'), covar=tensor([0.1523, 0.0957, 0.2047, 0.1309, 0.1125, 0.1166, 0.1553, 0.2049], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0248, 0.0141, 0.0122, 0.0134, 0.0153, 0.0119, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 02:31:01,223 INFO [finetune.py:976] (3/7) Epoch 12, batch 850, loss[loss=0.1648, simple_loss=0.2223, pruned_loss=0.0537, over 4742.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2571, pruned_loss=0.06079, over 944083.02 frames. ], batch size: 23, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:31:03,800 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3055, 1.6845, 2.1137, 2.6192, 2.0995, 1.6951, 1.3128, 1.8552], + device='cuda:3'), covar=tensor([0.3600, 0.3628, 0.1871, 0.2481, 0.2960, 0.2960, 0.4557, 0.2410], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0250, 0.0223, 0.0316, 0.0215, 0.0229, 0.0231, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 02:31:21,241 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63871.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:31:24,879 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63877.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:31:33,718 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1499, 1.5498, 1.3670, 1.7149, 1.7361, 1.7159, 1.3681, 3.1787], + device='cuda:3'), covar=tensor([0.0646, 0.0776, 0.0777, 0.1149, 0.0572, 0.0554, 0.0744, 0.0161], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:32:06,884 INFO [finetune.py:976] (3/7) Epoch 12, batch 900, loss[loss=0.1254, simple_loss=0.1973, pruned_loss=0.02681, over 4903.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2537, pruned_loss=0.05947, over 945350.47 frames. ], batch size: 32, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:32:07,626 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63906.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:19,215 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:24,717 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:24,726 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:34,606 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.661e+02 1.966e+02 2.302e+02 4.372e+02, threshold=3.933e+02, percent-clipped=1.0 +2023-04-27 02:32:36,448 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63942.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:45,688 INFO [finetune.py:976] (3/7) Epoch 12, batch 950, loss[loss=0.1579, simple_loss=0.2299, pruned_loss=0.04293, over 4904.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2505, pruned_loss=0.05856, over 946634.25 frames. ], batch size: 37, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:32:53,549 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8679, 1.1015, 2.9444, 2.7211, 2.6838, 2.8837, 2.8697, 2.6143], + device='cuda:3'), covar=tensor([0.7036, 0.5179, 0.1480, 0.2260, 0.1393, 0.2461, 0.1486, 0.1795], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0305, 0.0401, 0.0408, 0.0349, 0.0405, 0.0314, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:32:58,693 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 02:32:59,650 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:33:20,843 INFO [finetune.py:976] (3/7) Epoch 12, batch 1000, loss[loss=0.1998, simple_loss=0.2691, pruned_loss=0.06525, over 4817.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2526, pruned_loss=0.05939, over 945494.32 frames. ], batch size: 38, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:33:32,489 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64023.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:33:41,628 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5355, 2.2003, 2.5407, 2.9996, 2.9094, 2.5423, 2.0599, 2.6898], + device='cuda:3'), covar=tensor([0.0838, 0.0985, 0.0653, 0.0546, 0.0588, 0.0857, 0.0898, 0.0556], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0199, 0.0181, 0.0172, 0.0176, 0.0186, 0.0158, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:33:43,219 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.750e+01 1.665e+02 1.998e+02 2.380e+02 4.049e+02, threshold=3.995e+02, percent-clipped=1.0 +2023-04-27 02:33:58,097 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8687, 1.6332, 2.1051, 2.3337, 1.9313, 1.7855, 1.9583, 1.9446], + device='cuda:3'), covar=tensor([0.5868, 0.7941, 0.8537, 0.7347, 0.7190, 1.0708, 1.0341, 1.0059], + device='cuda:3'), in_proj_covar=tensor([0.0408, 0.0408, 0.0494, 0.0512, 0.0440, 0.0459, 0.0466, 0.0468], + device='cuda:3'), out_proj_covar=tensor([9.9264e-05, 1.0100e-04, 1.1147e-04, 1.2161e-04, 1.0644e-04, 1.1096e-04, + 1.1169e-04, 1.1206e-04], device='cuda:3') +2023-04-27 02:33:59,167 INFO [finetune.py:976] (3/7) Epoch 12, batch 1050, loss[loss=0.1713, simple_loss=0.2398, pruned_loss=0.05138, over 4789.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2544, pruned_loss=0.05917, over 948341.56 frames. ], batch size: 23, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:34:21,924 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4506, 1.3315, 4.0556, 3.7953, 3.5916, 3.8668, 3.8320, 3.5866], + device='cuda:3'), covar=tensor([0.7147, 0.5799, 0.1009, 0.1571, 0.1126, 0.1560, 0.1572, 0.1631], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0303, 0.0398, 0.0405, 0.0346, 0.0403, 0.0313, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:34:29,324 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64076.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:34:30,581 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:34:40,100 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64084.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:35:02,231 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 02:35:05,820 INFO [finetune.py:976] (3/7) Epoch 12, batch 1100, loss[loss=0.1887, simple_loss=0.2584, pruned_loss=0.05949, over 4798.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2553, pruned_loss=0.0598, over 949698.82 frames. ], batch size: 45, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:35:28,513 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:35:29,729 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64126.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:35:37,874 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6672, 1.8691, 0.8628, 1.3025, 1.8934, 1.5863, 1.3731, 1.4880], + device='cuda:3'), covar=tensor([0.0513, 0.0375, 0.0359, 0.0579, 0.0271, 0.0515, 0.0517, 0.0581], + device='cuda:3'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 02:35:38,339 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.705e+02 2.099e+02 2.623e+02 4.712e+02, threshold=4.198e+02, percent-clipped=5.0 +2023-04-27 02:35:49,424 INFO [finetune.py:976] (3/7) Epoch 12, batch 1150, loss[loss=0.1759, simple_loss=0.2454, pruned_loss=0.0532, over 4786.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.256, pruned_loss=0.0598, over 950867.97 frames. ], batch size: 26, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:35:57,679 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:36:03,096 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5701, 1.9506, 1.9128, 2.3989, 2.2964, 2.2322, 1.7955, 4.7811], + device='cuda:3'), covar=tensor([0.0548, 0.0741, 0.0752, 0.1058, 0.0565, 0.0495, 0.0710, 0.0110], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:36:06,790 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9861, 2.8711, 2.6459, 2.7797, 3.2164, 2.8517, 4.0369, 2.4382], + device='cuda:3'), covar=tensor([0.3857, 0.2088, 0.3707, 0.3406, 0.1523, 0.2436, 0.1266, 0.3467], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0348, 0.0429, 0.0360, 0.0385, 0.0381, 0.0375, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:36:19,854 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64201.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:36:22,239 INFO [finetune.py:976] (3/7) Epoch 12, batch 1200, loss[loss=0.1765, simple_loss=0.2529, pruned_loss=0.05003, over 4864.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2544, pruned_loss=0.05935, over 950002.95 frames. ], batch size: 34, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:36:36,045 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:36:40,394 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:36:50,582 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.610e+02 1.933e+02 2.309e+02 4.139e+02, threshold=3.865e+02, percent-clipped=0.0 +2023-04-27 02:36:51,933 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64242.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:02,008 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7781, 1.4274, 1.9083, 2.3019, 1.9069, 1.7467, 1.8627, 1.8595], + device='cuda:3'), covar=tensor([0.5735, 0.7828, 0.8288, 0.6900, 0.6854, 0.9511, 0.9105, 0.9105], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0408, 0.0495, 0.0513, 0.0440, 0.0460, 0.0467, 0.0468], + device='cuda:3'), out_proj_covar=tensor([9.9451e-05, 1.0112e-04, 1.1161e-04, 1.2179e-04, 1.0648e-04, 1.1112e-04, + 1.1189e-04, 1.1228e-04], device='cuda:3') +2023-04-27 02:37:04,709 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-04-27 02:37:10,839 INFO [finetune.py:976] (3/7) Epoch 12, batch 1250, loss[loss=0.1801, simple_loss=0.2445, pruned_loss=0.05787, over 4910.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2529, pruned_loss=0.05895, over 952267.46 frames. ], batch size: 32, lr: 3.66e-03, grad_scale: 64.0 +2023-04-27 02:37:33,223 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:33,822 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:55,440 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64290.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:57,317 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:57,406 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-27 02:38:16,202 INFO [finetune.py:976] (3/7) Epoch 12, batch 1300, loss[loss=0.1659, simple_loss=0.2404, pruned_loss=0.0457, over 4921.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2499, pruned_loss=0.05743, over 954242.96 frames. ], batch size: 36, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:38:32,770 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-27 02:38:40,956 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.564e+02 1.871e+02 2.311e+02 4.814e+02, threshold=3.742e+02, percent-clipped=4.0 +2023-04-27 02:38:43,496 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64345.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:38:45,294 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5459, 3.3734, 0.8017, 1.8758, 1.7977, 2.4029, 1.9466, 1.0143], + device='cuda:3'), covar=tensor([0.1363, 0.1165, 0.2225, 0.1327, 0.1144, 0.1041, 0.1406, 0.2047], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0246, 0.0139, 0.0121, 0.0133, 0.0152, 0.0118, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 02:38:49,942 INFO [finetune.py:976] (3/7) Epoch 12, batch 1350, loss[loss=0.2395, simple_loss=0.3073, pruned_loss=0.08589, over 4805.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2494, pruned_loss=0.05718, over 951312.78 frames. ], batch size: 51, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:39:07,143 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64379.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:39:22,954 INFO [finetune.py:976] (3/7) Epoch 12, batch 1400, loss[loss=0.1893, simple_loss=0.2646, pruned_loss=0.05699, over 4848.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2532, pruned_loss=0.05846, over 952259.80 frames. ], batch size: 49, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:39:23,686 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9678, 2.4109, 0.9776, 1.3212, 1.7885, 1.1445, 2.9843, 1.6076], + device='cuda:3'), covar=tensor([0.0686, 0.0612, 0.0730, 0.1226, 0.0484, 0.0998, 0.0308, 0.0599], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 02:39:23,724 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64406.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:39:37,467 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1421, 2.8733, 1.9218, 2.3083, 1.5619, 1.5069, 2.0800, 1.4995], + device='cuda:3'), covar=tensor([0.1916, 0.1569, 0.1865, 0.1820, 0.2721, 0.2380, 0.1236, 0.2269], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0204, 0.0201, 0.0184, 0.0157, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 02:39:48,147 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.746e+02 2.220e+02 2.632e+02 4.800e+02, threshold=4.440e+02, percent-clipped=5.0 +2023-04-27 02:40:07,660 INFO [finetune.py:976] (3/7) Epoch 12, batch 1450, loss[loss=0.1507, simple_loss=0.2152, pruned_loss=0.0431, over 4814.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2562, pruned_loss=0.05991, over 951983.93 frames. ], batch size: 25, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:40:21,242 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64465.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:40:21,841 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:40:55,246 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64501.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:40:55,854 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0556, 1.6529, 1.5553, 1.8945, 1.8565, 2.2004, 1.5286, 3.9275], + device='cuda:3'), covar=tensor([0.0628, 0.0786, 0.0751, 0.1163, 0.0621, 0.0523, 0.0742, 0.0158], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:40:57,604 INFO [finetune.py:976] (3/7) Epoch 12, batch 1500, loss[loss=0.1747, simple_loss=0.2496, pruned_loss=0.04995, over 4858.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2563, pruned_loss=0.05952, over 951742.63 frames. ], batch size: 44, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:41:03,987 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64514.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:12,859 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64526.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:22,336 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.729e+02 2.014e+02 2.511e+02 4.004e+02, threshold=4.028e+02, percent-clipped=0.0 +2023-04-27 02:41:27,205 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64549.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:30,769 INFO [finetune.py:976] (3/7) Epoch 12, batch 1550, loss[loss=0.1591, simple_loss=0.2316, pruned_loss=0.04327, over 4822.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2575, pruned_loss=0.06023, over 953459.39 frames. ], batch size: 25, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:41:42,606 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64572.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:53,769 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64588.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:54,937 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64589.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:42:10,373 INFO [finetune.py:976] (3/7) Epoch 12, batch 1600, loss[loss=0.1787, simple_loss=0.2465, pruned_loss=0.05544, over 4895.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2551, pruned_loss=0.05934, over 954548.39 frames. ], batch size: 32, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:42:29,956 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5111, 1.1895, 0.3484, 1.2365, 1.1781, 1.3663, 1.3019, 1.2807], + device='cuda:3'), covar=tensor([0.0545, 0.0409, 0.0467, 0.0607, 0.0331, 0.0571, 0.0562, 0.0612], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 02:42:32,225 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64620.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:42:46,406 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.700e+02 1.959e+02 2.289e+02 5.207e+02, threshold=3.917e+02, percent-clipped=1.0 +2023-04-27 02:42:56,558 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64648.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:42:57,775 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:43:06,461 INFO [finetune.py:976] (3/7) Epoch 12, batch 1650, loss[loss=0.1476, simple_loss=0.2193, pruned_loss=0.03793, over 4940.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.252, pruned_loss=0.0581, over 956323.26 frames. ], batch size: 38, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:43:21,413 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0857, 1.6673, 1.5344, 1.8300, 1.7475, 1.9130, 1.4787, 3.4893], + device='cuda:3'), covar=tensor([0.0641, 0.0774, 0.0733, 0.1162, 0.0622, 0.0490, 0.0723, 0.0176], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 02:43:36,853 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:09,945 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64701.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:11,242 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64703.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:12,323 INFO [finetune.py:976] (3/7) Epoch 12, batch 1700, loss[loss=0.1162, simple_loss=0.1903, pruned_loss=0.02102, over 4784.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2503, pruned_loss=0.05744, over 958293.38 frames. ], batch size: 28, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:44:14,901 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64709.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:27,407 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64727.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:37,468 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.591e+02 1.842e+02 2.298e+02 5.493e+02, threshold=3.684e+02, percent-clipped=1.0 +2023-04-27 02:44:40,033 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64745.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:44,764 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64752.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:46,474 INFO [finetune.py:976] (3/7) Epoch 12, batch 1750, loss[loss=0.2055, simple_loss=0.2756, pruned_loss=0.06766, over 4906.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2526, pruned_loss=0.05869, over 956534.03 frames. ], batch size: 43, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:44:52,009 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64764.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:52,607 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9540, 1.5712, 1.8003, 2.2201, 1.8308, 1.5073, 1.3344, 1.6182], + device='cuda:3'), covar=tensor([0.2743, 0.2763, 0.1474, 0.1724, 0.2366, 0.2375, 0.3945, 0.2150], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0247, 0.0221, 0.0314, 0.0212, 0.0228, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 02:45:11,222 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7504, 2.0999, 1.8698, 1.9778, 1.7202, 1.6413, 1.7657, 1.4962], + device='cuda:3'), covar=tensor([0.1587, 0.1336, 0.0840, 0.1115, 0.3155, 0.1259, 0.1825, 0.2261], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0311, 0.0223, 0.0280, 0.0310, 0.0265, 0.0251, 0.0272], + device='cuda:3'), out_proj_covar=tensor([1.1710e-04, 1.2481e-04, 8.9290e-05, 1.1170e-04, 1.2656e-04, 1.0634e-04, + 1.0205e-04, 1.0866e-04], device='cuda:3') +2023-04-27 02:45:20,204 INFO [finetune.py:976] (3/7) Epoch 12, batch 1800, loss[loss=0.2054, simple_loss=0.2787, pruned_loss=0.06606, over 4863.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2561, pruned_loss=0.05972, over 957001.61 frames. ], batch size: 44, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:45:20,913 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64806.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:45:25,154 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64813.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:45:29,908 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:46:00,006 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.230e+02 1.632e+02 2.043e+02 2.586e+02 5.836e+02, threshold=4.087e+02, percent-clipped=7.0 +2023-04-27 02:46:21,067 INFO [finetune.py:976] (3/7) Epoch 12, batch 1850, loss[loss=0.1998, simple_loss=0.2685, pruned_loss=0.06553, over 4878.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2566, pruned_loss=0.06016, over 957292.56 frames. ], batch size: 32, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:46:58,522 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:47:10,163 INFO [finetune.py:976] (3/7) Epoch 12, batch 1900, loss[loss=0.2173, simple_loss=0.2862, pruned_loss=0.07421, over 4902.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2596, pruned_loss=0.06162, over 956320.07 frames. ], batch size: 37, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:47:35,552 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:47:39,046 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.650e+02 1.962e+02 2.439e+02 4.068e+02, threshold=3.925e+02, percent-clipped=0.0 +2023-04-27 02:47:48,347 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:48:00,716 INFO [finetune.py:976] (3/7) Epoch 12, batch 1950, loss[loss=0.1534, simple_loss=0.2255, pruned_loss=0.04068, over 4791.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2561, pruned_loss=0.05989, over 952797.79 frames. ], batch size: 51, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:48:31,531 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64976.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:48:42,270 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-27 02:48:49,860 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65001.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:48:51,660 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65004.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:48:52,188 INFO [finetune.py:976] (3/7) Epoch 12, batch 2000, loss[loss=0.1928, simple_loss=0.2627, pruned_loss=0.06148, over 4832.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2531, pruned_loss=0.05898, over 952016.52 frames. ], batch size: 33, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:49:12,511 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65037.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 02:49:15,252 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.562e+02 1.931e+02 2.366e+02 5.248e+02, threshold=3.862e+02, percent-clipped=2.0 +2023-04-27 02:49:21,191 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65049.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:49:26,343 INFO [finetune.py:976] (3/7) Epoch 12, batch 2050, loss[loss=0.1572, simple_loss=0.221, pruned_loss=0.04674, over 4837.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2495, pruned_loss=0.05745, over 953631.49 frames. ], batch size: 49, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:49:26,470 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:49:28,856 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65059.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:49:56,653 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65101.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:00,021 INFO [finetune.py:976] (3/7) Epoch 12, batch 2100, loss[loss=0.1798, simple_loss=0.2507, pruned_loss=0.05448, over 4886.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2502, pruned_loss=0.05815, over 955583.13 frames. ], batch size: 32, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:50:01,930 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65108.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:05,599 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5063, 3.5268, 0.9811, 1.9406, 1.8998, 2.7209, 1.9733, 0.9933], + device='cuda:3'), covar=tensor([0.1381, 0.0820, 0.1942, 0.1247, 0.1106, 0.0813, 0.1511, 0.2079], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0248, 0.0140, 0.0123, 0.0134, 0.0154, 0.0119, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 02:50:06,885 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:10,312 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65121.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:22,435 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.658e+02 1.987e+02 2.538e+02 5.198e+02, threshold=3.973e+02, percent-clipped=2.0 +2023-04-27 02:50:33,407 INFO [finetune.py:976] (3/7) Epoch 12, batch 2150, loss[loss=0.1802, simple_loss=0.2615, pruned_loss=0.04942, over 4820.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2543, pruned_loss=0.05983, over 955426.32 frames. ], batch size: 51, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:50:34,149 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5528, 1.1185, 1.2807, 1.1646, 1.6854, 1.3588, 1.1260, 1.1815], + device='cuda:3'), covar=tensor([0.1668, 0.1450, 0.2220, 0.1678, 0.1009, 0.1456, 0.2272, 0.2452], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0316, 0.0350, 0.0294, 0.0331, 0.0316, 0.0304, 0.0359], + device='cuda:3'), out_proj_covar=tensor([6.3657e-05, 6.6462e-05, 7.5270e-05, 6.0415e-05, 6.9017e-05, 6.7289e-05, + 6.4868e-05, 7.6896e-05], device='cuda:3') +2023-04-27 02:50:42,399 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65169.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:51:05,956 INFO [finetune.py:976] (3/7) Epoch 12, batch 2200, loss[loss=0.1585, simple_loss=0.2357, pruned_loss=0.04062, over 4861.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2569, pruned_loss=0.06102, over 954565.49 frames. ], batch size: 34, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:51:37,965 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 02:51:56,236 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.743e+02 2.057e+02 2.570e+02 5.251e+02, threshold=4.115e+02, percent-clipped=3.0 +2023-04-27 02:51:58,806 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65245.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:52:11,641 INFO [finetune.py:976] (3/7) Epoch 12, batch 2250, loss[loss=0.2089, simple_loss=0.2904, pruned_loss=0.06367, over 4911.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2593, pruned_loss=0.06176, over 955217.56 frames. ], batch size: 37, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:52:53,424 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65285.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:53:03,655 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:53:21,403 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65304.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:53:21,921 INFO [finetune.py:976] (3/7) Epoch 12, batch 2300, loss[loss=0.2177, simple_loss=0.2813, pruned_loss=0.07701, over 4887.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2574, pruned_loss=0.06011, over 954776.52 frames. ], batch size: 35, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:53:23,798 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1133, 1.5451, 1.9599, 2.1376, 1.8963, 1.5142, 1.0434, 1.6124], + device='cuda:3'), covar=tensor([0.3250, 0.3564, 0.1764, 0.2367, 0.2725, 0.2854, 0.4608, 0.2167], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0251, 0.0224, 0.0318, 0.0215, 0.0231, 0.0232, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 02:53:46,696 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0079, 1.9996, 1.8094, 1.6642, 2.0961, 1.7422, 2.6309, 1.5803], + device='cuda:3'), covar=tensor([0.3954, 0.1832, 0.4967, 0.3138, 0.1878, 0.2460, 0.1434, 0.4549], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0347, 0.0429, 0.0359, 0.0385, 0.0382, 0.0375, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:53:56,236 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65332.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 02:54:07,784 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.597e+02 1.792e+02 2.067e+02 3.757e+02, threshold=3.584e+02, percent-clipped=0.0 +2023-04-27 02:54:08,532 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1913, 1.8929, 2.3590, 2.5602, 2.2236, 2.0381, 2.2492, 2.1686], + device='cuda:3'), covar=tensor([0.5576, 0.8121, 0.8754, 0.7576, 0.7287, 1.0592, 0.9546, 0.9581], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0406, 0.0496, 0.0514, 0.0440, 0.0461, 0.0466, 0.0470], + device='cuda:3'), out_proj_covar=tensor([9.9441e-05, 1.0064e-04, 1.1192e-04, 1.2186e-04, 1.0658e-04, 1.1124e-04, + 1.1166e-04, 1.1246e-04], device='cuda:3') +2023-04-27 02:54:16,355 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65346.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:20,874 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65352.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:22,654 INFO [finetune.py:976] (3/7) Epoch 12, batch 2350, loss[loss=0.2275, simple_loss=0.2893, pruned_loss=0.0828, over 4834.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2544, pruned_loss=0.05925, over 954346.21 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:54:25,636 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65359.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:53,215 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65401.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:55,973 INFO [finetune.py:976] (3/7) Epoch 12, batch 2400, loss[loss=0.1632, simple_loss=0.2321, pruned_loss=0.04719, over 4761.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.251, pruned_loss=0.05734, over 954266.93 frames. ], batch size: 28, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:54:57,714 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65407.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:58,361 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65408.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:00,657 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65411.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:05,916 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1249, 2.6704, 1.1083, 1.4203, 2.0881, 1.2421, 3.6711, 1.8620], + device='cuda:3'), covar=tensor([0.0708, 0.0639, 0.0768, 0.1308, 0.0509, 0.1038, 0.0275, 0.0652], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 02:55:20,647 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.563e+02 1.962e+02 2.388e+02 3.739e+02, threshold=3.925e+02, percent-clipped=1.0 +2023-04-27 02:55:25,587 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65449.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:29,228 INFO [finetune.py:976] (3/7) Epoch 12, batch 2450, loss[loss=0.1781, simple_loss=0.2491, pruned_loss=0.05352, over 4866.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2492, pruned_loss=0.05708, over 954731.46 frames. ], batch size: 34, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:55:30,350 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65456.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:41,387 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4019, 3.3380, 2.5591, 3.9097, 3.3673, 3.4036, 1.5234, 3.3629], + device='cuda:3'), covar=tensor([0.1868, 0.1297, 0.2915, 0.2153, 0.2738, 0.1889, 0.5538, 0.2488], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0217, 0.0250, 0.0305, 0.0299, 0.0249, 0.0272, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:55:59,910 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7172, 3.7932, 0.8880, 2.2339, 2.1629, 2.5904, 2.3024, 0.8646], + device='cuda:3'), covar=tensor([0.1268, 0.0784, 0.2085, 0.1139, 0.1031, 0.0986, 0.1405, 0.2319], + device='cuda:3'), in_proj_covar=tensor([0.0119, 0.0249, 0.0141, 0.0123, 0.0135, 0.0154, 0.0119, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 02:56:02,905 INFO [finetune.py:976] (3/7) Epoch 12, batch 2500, loss[loss=0.1572, simple_loss=0.2242, pruned_loss=0.04507, over 4771.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2512, pruned_loss=0.05851, over 954107.80 frames. ], batch size: 26, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:56:09,487 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 02:56:22,234 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65532.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 02:56:26,808 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 02:56:28,027 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.847e+02 2.256e+02 2.656e+02 4.911e+02, threshold=4.512e+02, percent-clipped=5.0 +2023-04-27 02:56:36,595 INFO [finetune.py:976] (3/7) Epoch 12, batch 2550, loss[loss=0.1826, simple_loss=0.2553, pruned_loss=0.05497, over 4918.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2515, pruned_loss=0.05805, over 950167.05 frames. ], batch size: 36, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:56:51,969 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.2941, 4.3245, 3.1793, 4.9320, 4.2813, 4.2528, 2.1872, 4.2520], + device='cuda:3'), covar=tensor([0.1456, 0.0846, 0.3248, 0.0927, 0.2851, 0.1545, 0.4904, 0.2244], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0216, 0.0249, 0.0304, 0.0299, 0.0248, 0.0270, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:57:02,307 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65593.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 02:57:04,486 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.0670, 4.0453, 2.8627, 4.6438, 4.0106, 4.0059, 1.9968, 4.0174], + device='cuda:3'), covar=tensor([0.1428, 0.0982, 0.2875, 0.1238, 0.2769, 0.1715, 0.4869, 0.2208], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0215, 0.0248, 0.0303, 0.0298, 0.0248, 0.0270, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:57:10,044 INFO [finetune.py:976] (3/7) Epoch 12, batch 2600, loss[loss=0.2258, simple_loss=0.2971, pruned_loss=0.07728, over 4797.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2546, pruned_loss=0.06004, over 948956.47 frames. ], batch size: 51, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:57:20,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5408, 1.0481, 1.2527, 1.1978, 1.6480, 1.3214, 1.1192, 1.2000], + device='cuda:3'), covar=tensor([0.1622, 0.1438, 0.2229, 0.1508, 0.0961, 0.1714, 0.2225, 0.2414], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0320, 0.0356, 0.0299, 0.0335, 0.0320, 0.0309, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.4608e-05, 6.7364e-05, 7.6579e-05, 6.1477e-05, 6.9944e-05, 6.8162e-05, + 6.5857e-05, 7.7975e-05], device='cuda:3') +2023-04-27 02:57:29,337 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65632.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:57:34,707 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65641.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:57:35,234 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.708e+02 2.004e+02 2.304e+02 5.079e+02, threshold=4.008e+02, percent-clipped=1.0 +2023-04-27 02:57:54,580 INFO [finetune.py:976] (3/7) Epoch 12, batch 2650, loss[loss=0.1734, simple_loss=0.2472, pruned_loss=0.04983, over 4828.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2561, pruned_loss=0.06022, over 949620.78 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:58:17,990 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6447, 1.2738, 4.3158, 4.0190, 3.7681, 4.0877, 3.9322, 3.7381], + device='cuda:3'), covar=tensor([0.7124, 0.6038, 0.0918, 0.1603, 0.1161, 0.1521, 0.1854, 0.1358], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0302, 0.0398, 0.0403, 0.0345, 0.0404, 0.0313, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 02:58:26,148 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:58:54,680 INFO [finetune.py:976] (3/7) Epoch 12, batch 2700, loss[loss=0.2366, simple_loss=0.2944, pruned_loss=0.08941, over 4272.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2547, pruned_loss=0.0597, over 947482.37 frames. ], batch size: 65, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:59:03,989 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65711.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:59:06,980 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4664, 3.4470, 2.8132, 3.9984, 3.2999, 3.4570, 1.9838, 3.4781], + device='cuda:3'), covar=tensor([0.2042, 0.1379, 0.3969, 0.1448, 0.2825, 0.1738, 0.4726, 0.2477], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0216, 0.0250, 0.0304, 0.0299, 0.0249, 0.0272, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 02:59:42,196 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.570e+02 1.817e+02 2.252e+02 5.618e+02, threshold=3.635e+02, percent-clipped=2.0 +2023-04-27 02:59:48,439 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-04-27 03:00:01,384 INFO [finetune.py:976] (3/7) Epoch 12, batch 2750, loss[loss=0.1702, simple_loss=0.2417, pruned_loss=0.04941, over 4813.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2533, pruned_loss=0.05957, over 947443.44 frames. ], batch size: 41, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 03:00:09,045 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65759.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:01:07,122 INFO [finetune.py:976] (3/7) Epoch 12, batch 2800, loss[loss=0.1789, simple_loss=0.2503, pruned_loss=0.05369, over 4833.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2511, pruned_loss=0.05863, over 950588.19 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:01:07,237 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8698, 2.1429, 1.1855, 1.6011, 2.2839, 1.7751, 1.6339, 1.8562], + device='cuda:3'), covar=tensor([0.0508, 0.0344, 0.0299, 0.0555, 0.0233, 0.0528, 0.0509, 0.0543], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 03:01:51,930 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65838.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:01:59,670 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.664e+02 1.878e+02 2.585e+02 4.047e+02, threshold=3.756e+02, percent-clipped=2.0 +2023-04-27 03:02:08,139 INFO [finetune.py:976] (3/7) Epoch 12, batch 2850, loss[loss=0.1629, simple_loss=0.2368, pruned_loss=0.04452, over 4900.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2493, pruned_loss=0.058, over 953024.59 frames. ], batch size: 32, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:02:20,523 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 03:02:29,540 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65888.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:02:38,196 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65899.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:02:41,704 INFO [finetune.py:976] (3/7) Epoch 12, batch 2900, loss[loss=0.1682, simple_loss=0.2532, pruned_loss=0.0416, over 4748.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2519, pruned_loss=0.05954, over 952347.99 frames. ], batch size: 54, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:03:05,057 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:03:05,560 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.732e+02 2.023e+02 2.445e+02 4.251e+02, threshold=4.047e+02, percent-clipped=2.0 +2023-04-27 03:03:15,319 INFO [finetune.py:976] (3/7) Epoch 12, batch 2950, loss[loss=0.2087, simple_loss=0.2841, pruned_loss=0.06669, over 4818.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2555, pruned_loss=0.06015, over 953082.50 frames. ], batch size: 38, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:03:20,367 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.63 vs. limit=5.0 +2023-04-27 03:03:33,320 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 03:03:36,944 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:03:49,753 INFO [finetune.py:976] (3/7) Epoch 12, batch 3000, loss[loss=0.1609, simple_loss=0.2294, pruned_loss=0.0462, over 4723.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2564, pruned_loss=0.0604, over 953371.68 frames. ], batch size: 23, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:03:49,753 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 03:03:57,805 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8390, 1.6464, 1.8168, 2.1826, 2.1740, 1.8179, 1.4486, 1.9875], + device='cuda:3'), covar=tensor([0.0839, 0.1137, 0.0778, 0.0475, 0.0517, 0.0799, 0.0837, 0.0528], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0205, 0.0186, 0.0177, 0.0181, 0.0191, 0.0160, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:04:00,412 INFO [finetune.py:1010] (3/7) Epoch 12, validation: loss=0.1529, simple_loss=0.2247, pruned_loss=0.04052, over 2265189.00 frames. +2023-04-27 03:04:00,413 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 03:04:12,728 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5174, 0.9959, 1.3094, 1.2709, 1.7038, 1.3147, 1.0999, 1.2542], + device='cuda:3'), covar=tensor([0.2032, 0.1655, 0.2369, 0.1577, 0.0920, 0.1741, 0.2193, 0.2426], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0322, 0.0355, 0.0299, 0.0335, 0.0319, 0.0309, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.4673e-05, 6.7808e-05, 7.6337e-05, 6.1367e-05, 7.0027e-05, 6.7879e-05, + 6.5809e-05, 7.7380e-05], device='cuda:3') +2023-04-27 03:04:21,283 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9636, 2.4337, 1.9893, 2.2659, 1.5982, 1.9257, 2.0590, 1.5681], + device='cuda:3'), covar=tensor([0.1866, 0.1203, 0.0915, 0.1152, 0.3312, 0.1124, 0.1595, 0.2262], + device='cuda:3'), in_proj_covar=tensor([0.0289, 0.0308, 0.0222, 0.0280, 0.0310, 0.0261, 0.0250, 0.0271], + device='cuda:3'), out_proj_covar=tensor([1.1645e-04, 1.2330e-04, 8.8737e-05, 1.1206e-04, 1.2653e-04, 1.0477e-04, + 1.0141e-04, 1.0825e-04], device='cuda:3') +2023-04-27 03:04:29,439 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.787e+02 2.243e+02 2.841e+02 1.392e+03, threshold=4.486e+02, percent-clipped=4.0 +2023-04-27 03:04:38,282 INFO [finetune.py:976] (3/7) Epoch 12, batch 3050, loss[loss=0.2309, simple_loss=0.2924, pruned_loss=0.0847, over 4871.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.258, pruned_loss=0.06117, over 951158.83 frames. ], batch size: 34, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:04:46,193 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-04-27 03:05:10,512 INFO [finetune.py:976] (3/7) Epoch 12, batch 3100, loss[loss=0.1536, simple_loss=0.229, pruned_loss=0.03905, over 4826.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2553, pruned_loss=0.06014, over 951665.13 frames. ], batch size: 39, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:05:40,809 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.687e+02 1.900e+02 2.210e+02 4.094e+02, threshold=3.799e+02, percent-clipped=0.0 +2023-04-27 03:05:41,556 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1457, 1.9229, 2.1842, 2.4718, 2.4445, 1.9666, 1.6905, 2.2327], + device='cuda:3'), covar=tensor([0.0837, 0.0980, 0.0547, 0.0451, 0.0586, 0.0877, 0.0806, 0.0518], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0205, 0.0186, 0.0177, 0.0181, 0.0191, 0.0160, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:05:48,037 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:05:54,641 INFO [finetune.py:976] (3/7) Epoch 12, batch 3150, loss[loss=0.1935, simple_loss=0.2534, pruned_loss=0.06686, over 4745.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2531, pruned_loss=0.05956, over 952785.00 frames. ], batch size: 59, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:05:58,698 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66160.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:06:27,541 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66188.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:06:36,568 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66194.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:06:38,673 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 03:06:49,148 INFO [finetune.py:976] (3/7) Epoch 12, batch 3200, loss[loss=0.1841, simple_loss=0.2325, pruned_loss=0.06787, over 4821.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2493, pruned_loss=0.05788, over 951363.72 frames. ], batch size: 25, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:06:49,262 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66205.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:07:12,332 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66221.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:07:32,458 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8219, 2.4292, 1.8597, 1.6539, 1.3202, 1.3551, 1.9573, 1.2461], + device='cuda:3'), covar=tensor([0.1793, 0.1423, 0.1585, 0.1931, 0.2568, 0.2020, 0.1059, 0.2210], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0215, 0.0170, 0.0204, 0.0203, 0.0185, 0.0158, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 03:07:34,235 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66236.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:07:43,555 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.908e+01 1.563e+02 1.889e+02 2.258e+02 3.990e+02, threshold=3.778e+02, percent-clipped=1.0 +2023-04-27 03:07:57,818 INFO [finetune.py:976] (3/7) Epoch 12, batch 3250, loss[loss=0.1654, simple_loss=0.2457, pruned_loss=0.04256, over 4781.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2504, pruned_loss=0.05854, over 953102.99 frames. ], batch size: 26, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:07:59,782 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66258.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:08:47,969 INFO [finetune.py:976] (3/7) Epoch 12, batch 3300, loss[loss=0.2239, simple_loss=0.2992, pruned_loss=0.0743, over 4822.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2548, pruned_loss=0.06028, over 955341.82 frames. ], batch size: 38, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:08:57,547 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:09:11,650 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3250, 1.5340, 1.4450, 1.9654, 1.7717, 2.0418, 1.4301, 4.0943], + device='cuda:3'), covar=tensor([0.0627, 0.0843, 0.0790, 0.1179, 0.0649, 0.0663, 0.0780, 0.0121], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 03:09:13,358 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.300e+02 1.708e+02 2.133e+02 2.444e+02 6.112e+02, threshold=4.266e+02, percent-clipped=2.0 +2023-04-27 03:09:19,442 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6202, 2.2337, 1.5171, 1.4771, 1.1890, 1.2270, 1.5800, 1.1014], + device='cuda:3'), covar=tensor([0.2099, 0.1324, 0.1999, 0.2125, 0.2994, 0.2802, 0.1228, 0.2494], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0214, 0.0170, 0.0204, 0.0203, 0.0184, 0.0158, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 03:09:21,693 INFO [finetune.py:976] (3/7) Epoch 12, batch 3350, loss[loss=0.2191, simple_loss=0.2865, pruned_loss=0.07588, over 4829.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2564, pruned_loss=0.0606, over 956037.20 frames. ], batch size: 47, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:09:54,777 INFO [finetune.py:976] (3/7) Epoch 12, batch 3400, loss[loss=0.2037, simple_loss=0.2793, pruned_loss=0.06408, over 4834.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2582, pruned_loss=0.06168, over 955932.23 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:10:20,215 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.592e+02 1.894e+02 2.195e+02 3.276e+02, threshold=3.787e+02, percent-clipped=0.0 +2023-04-27 03:10:28,168 INFO [finetune.py:976] (3/7) Epoch 12, batch 3450, loss[loss=0.2066, simple_loss=0.2553, pruned_loss=0.07896, over 4824.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2581, pruned_loss=0.0615, over 955874.65 frames. ], batch size: 30, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:10:44,937 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0131, 1.5422, 1.6255, 1.7386, 2.2112, 1.7853, 1.4872, 1.5498], + device='cuda:3'), covar=tensor([0.1315, 0.1520, 0.1778, 0.1187, 0.0899, 0.1660, 0.2034, 0.2065], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0321, 0.0356, 0.0299, 0.0337, 0.0320, 0.0309, 0.0363], + device='cuda:3'), out_proj_covar=tensor([6.4852e-05, 6.7524e-05, 7.6533e-05, 6.1435e-05, 7.0484e-05, 6.8092e-05, + 6.5714e-05, 7.7671e-05], device='cuda:3') +2023-04-27 03:10:55,270 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66494.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:10:58,895 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66500.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:11:01,856 INFO [finetune.py:976] (3/7) Epoch 12, batch 3500, loss[loss=0.1906, simple_loss=0.2642, pruned_loss=0.05853, over 4820.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2577, pruned_loss=0.06214, over 956349.41 frames. ], batch size: 38, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:11:09,122 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66516.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:11:12,194 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66521.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:11:26,632 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.637e+01 1.554e+02 1.930e+02 2.326e+02 4.798e+02, threshold=3.860e+02, percent-clipped=2.0 +2023-04-27 03:11:27,202 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66542.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:11:40,353 INFO [finetune.py:976] (3/7) Epoch 12, batch 3550, loss[loss=0.1597, simple_loss=0.2228, pruned_loss=0.04825, over 4784.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2546, pruned_loss=0.06102, over 956446.81 frames. ], batch size: 26, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:11:59,812 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7491, 2.2484, 1.9879, 2.1178, 1.7177, 1.8922, 1.8718, 1.4476], + device='cuda:3'), covar=tensor([0.2088, 0.1443, 0.0932, 0.1261, 0.3426, 0.1245, 0.1817, 0.2565], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0311, 0.0223, 0.0282, 0.0311, 0.0262, 0.0251, 0.0271], + device='cuda:3'), out_proj_covar=tensor([1.1700e-04, 1.2440e-04, 8.9016e-05, 1.1257e-04, 1.2696e-04, 1.0482e-04, + 1.0201e-04, 1.0827e-04], device='cuda:3') +2023-04-27 03:12:12,779 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66582.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:12:45,655 INFO [finetune.py:976] (3/7) Epoch 12, batch 3600, loss[loss=0.1221, simple_loss=0.1902, pruned_loss=0.02698, over 4229.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.251, pruned_loss=0.05919, over 955307.93 frames. ], batch size: 18, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:12:57,588 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66614.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:13:31,412 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.620e+02 1.935e+02 2.301e+02 4.046e+02, threshold=3.870e+02, percent-clipped=1.0 +2023-04-27 03:13:42,012 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3535, 1.7262, 2.1425, 2.8278, 2.1699, 1.7739, 1.7549, 2.0617], + device='cuda:3'), covar=tensor([0.3396, 0.3599, 0.1800, 0.2715, 0.2998, 0.2775, 0.4223, 0.2569], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0250, 0.0224, 0.0319, 0.0215, 0.0230, 0.0232, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 03:13:45,503 INFO [finetune.py:976] (3/7) Epoch 12, batch 3650, loss[loss=0.1739, simple_loss=0.254, pruned_loss=0.04692, over 4818.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.253, pruned_loss=0.05983, over 953982.94 frames. ], batch size: 39, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:14:07,262 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9880, 1.4567, 1.8355, 2.0611, 1.8014, 1.4223, 1.0007, 1.5357], + device='cuda:3'), covar=tensor([0.3519, 0.3422, 0.1849, 0.2495, 0.2832, 0.2908, 0.4661, 0.2372], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0249, 0.0223, 0.0318, 0.0215, 0.0230, 0.0231, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 03:14:19,348 INFO [finetune.py:976] (3/7) Epoch 12, batch 3700, loss[loss=0.1844, simple_loss=0.2581, pruned_loss=0.05531, over 4741.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2561, pruned_loss=0.0603, over 952981.30 frames. ], batch size: 54, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:14:35,824 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66731.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:14:43,337 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.645e+02 1.962e+02 2.423e+02 5.947e+02, threshold=3.923e+02, percent-clipped=4.0 +2023-04-27 03:14:52,768 INFO [finetune.py:976] (3/7) Epoch 12, batch 3750, loss[loss=0.1814, simple_loss=0.2525, pruned_loss=0.05516, over 4781.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2578, pruned_loss=0.06084, over 953706.30 frames. ], batch size: 51, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:15:16,254 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66792.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:15:16,300 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 03:15:21,666 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66800.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:15:25,457 INFO [finetune.py:976] (3/7) Epoch 12, batch 3800, loss[loss=0.1521, simple_loss=0.2439, pruned_loss=0.03018, over 4814.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2585, pruned_loss=0.0612, over 953605.93 frames. ], batch size: 38, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:15:32,290 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66816.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:15:48,321 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.670e+02 1.892e+02 2.401e+02 5.287e+02, threshold=3.784e+02, percent-clipped=2.0 +2023-04-27 03:15:52,973 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:15:58,164 INFO [finetune.py:976] (3/7) Epoch 12, batch 3850, loss[loss=0.2154, simple_loss=0.2814, pruned_loss=0.07471, over 4909.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2568, pruned_loss=0.06022, over 951982.02 frames. ], batch size: 46, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:16:04,613 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66864.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:16:08,322 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5849, 1.8223, 1.7943, 2.3741, 2.5753, 2.1614, 2.0804, 1.9095], + device='cuda:3'), covar=tensor([0.2497, 0.2248, 0.2587, 0.2102, 0.1529, 0.2648, 0.3293, 0.3164], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0320, 0.0354, 0.0298, 0.0334, 0.0316, 0.0306, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.4435e-05, 6.7380e-05, 7.5932e-05, 6.1301e-05, 6.9650e-05, 6.7254e-05, + 6.5221e-05, 7.7530e-05], device='cuda:3') +2023-04-27 03:16:12,475 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66877.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:16:22,729 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8255, 2.3865, 1.8053, 1.7496, 1.3452, 1.3541, 1.8341, 1.2107], + device='cuda:3'), covar=tensor([0.1528, 0.1196, 0.1419, 0.1605, 0.2334, 0.1938, 0.0966, 0.2030], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0215, 0.0171, 0.0204, 0.0204, 0.0185, 0.0158, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 03:16:23,340 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4594, 1.3399, 1.7177, 1.6243, 1.3153, 1.2462, 1.3483, 0.7930], + device='cuda:3'), covar=tensor([0.0614, 0.0820, 0.0528, 0.0656, 0.0801, 0.1281, 0.0629, 0.0777], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0071, 0.0071, 0.0067, 0.0075, 0.0097, 0.0076, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 03:16:31,353 INFO [finetune.py:976] (3/7) Epoch 12, batch 3900, loss[loss=0.184, simple_loss=0.2565, pruned_loss=0.05579, over 4934.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2536, pruned_loss=0.05925, over 952050.81 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:16:38,406 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66914.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:16:40,890 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8427, 1.5065, 1.8381, 2.2536, 2.2780, 1.7392, 1.5306, 1.9971], + device='cuda:3'), covar=tensor([0.0859, 0.1312, 0.0745, 0.0502, 0.0527, 0.0917, 0.0789, 0.0550], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0201, 0.0182, 0.0173, 0.0177, 0.0185, 0.0155, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:16:59,108 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 03:17:12,162 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.631e+02 1.862e+02 2.227e+02 3.932e+02, threshold=3.724e+02, percent-clipped=1.0 +2023-04-27 03:17:24,414 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.69 vs. limit=5.0 +2023-04-27 03:17:32,061 INFO [finetune.py:976] (3/7) Epoch 12, batch 3950, loss[loss=0.1795, simple_loss=0.2415, pruned_loss=0.05874, over 4834.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2489, pruned_loss=0.0572, over 953937.93 frames. ], batch size: 30, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:17:33,393 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66957.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:17:42,916 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66962.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:18:04,829 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8046, 2.1083, 1.9629, 2.1544, 1.8668, 2.0952, 2.1352, 1.9821], + device='cuda:3'), covar=tensor([0.4689, 0.6812, 0.5299, 0.4757, 0.6178, 0.8258, 0.6648, 0.6166], + device='cuda:3'), in_proj_covar=tensor([0.0328, 0.0377, 0.0315, 0.0326, 0.0337, 0.0398, 0.0357, 0.0325], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 03:18:37,442 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 03:18:39,568 INFO [finetune.py:976] (3/7) Epoch 12, batch 4000, loss[loss=0.2183, simple_loss=0.2787, pruned_loss=0.07896, over 4870.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2497, pruned_loss=0.05819, over 952090.26 frames. ], batch size: 34, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:18:58,955 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67018.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:19:05,173 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 03:19:14,011 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.700e+02 2.045e+02 2.371e+02 4.368e+02, threshold=4.090e+02, percent-clipped=3.0 +2023-04-27 03:19:22,922 INFO [finetune.py:976] (3/7) Epoch 12, batch 4050, loss[loss=0.2092, simple_loss=0.2741, pruned_loss=0.07218, over 4821.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2528, pruned_loss=0.05933, over 952554.76 frames. ], batch size: 45, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:19:42,050 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67083.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:19:44,507 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67087.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:19:46,061 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-04-27 03:19:56,227 INFO [finetune.py:976] (3/7) Epoch 12, batch 4100, loss[loss=0.2389, simple_loss=0.3036, pruned_loss=0.08714, over 4827.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2553, pruned_loss=0.05987, over 954143.52 frames. ], batch size: 39, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:20:21,135 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.668e+02 1.950e+02 2.451e+02 4.466e+02, threshold=3.899e+02, percent-clipped=3.0 +2023-04-27 03:20:22,488 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:20:29,535 INFO [finetune.py:976] (3/7) Epoch 12, batch 4150, loss[loss=0.1688, simple_loss=0.2427, pruned_loss=0.04743, over 4792.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2575, pruned_loss=0.0612, over 954042.23 frames. ], batch size: 51, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:20:45,911 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67177.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:21:03,549 INFO [finetune.py:976] (3/7) Epoch 12, batch 4200, loss[loss=0.157, simple_loss=0.234, pruned_loss=0.04, over 4790.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2589, pruned_loss=0.06135, over 954646.15 frames. ], batch size: 51, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:21:08,279 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3396, 1.1846, 3.7516, 3.2241, 3.3444, 3.4690, 3.5237, 3.1081], + device='cuda:3'), covar=tensor([0.9183, 0.7825, 0.1786, 0.3194, 0.2402, 0.3784, 0.2747, 0.3126], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0308, 0.0406, 0.0413, 0.0353, 0.0412, 0.0316, 0.0372], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:21:11,772 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5928, 1.4768, 1.8543, 1.9049, 1.4375, 1.3071, 1.5463, 0.9569], + device='cuda:3'), covar=tensor([0.0686, 0.0692, 0.0526, 0.0718, 0.0837, 0.1250, 0.0717, 0.0870], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0071, 0.0070, 0.0067, 0.0075, 0.0096, 0.0076, 0.0071], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 03:21:18,109 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67225.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:21:28,846 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.623e+02 1.873e+02 2.449e+02 3.643e+02, threshold=3.747e+02, percent-clipped=0.0 +2023-04-27 03:21:37,114 INFO [finetune.py:976] (3/7) Epoch 12, batch 4250, loss[loss=0.188, simple_loss=0.2527, pruned_loss=0.06167, over 4883.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2559, pruned_loss=0.06045, over 954081.35 frames. ], batch size: 35, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:22:10,186 INFO [finetune.py:976] (3/7) Epoch 12, batch 4300, loss[loss=0.1734, simple_loss=0.2446, pruned_loss=0.05112, over 4754.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2536, pruned_loss=0.05987, over 954392.48 frames. ], batch size: 28, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:22:15,634 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67313.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:22:26,382 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0229, 1.5181, 5.3738, 5.0676, 4.6583, 5.2077, 4.6681, 4.7855], + device='cuda:3'), covar=tensor([0.6990, 0.5986, 0.0847, 0.1577, 0.1027, 0.1682, 0.1234, 0.1413], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0308, 0.0406, 0.0413, 0.0354, 0.0413, 0.0317, 0.0372], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:22:31,556 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8137, 1.3320, 1.8314, 2.2329, 1.8967, 1.7889, 1.8269, 1.8196], + device='cuda:3'), covar=tensor([0.5542, 0.8195, 0.8271, 0.7392, 0.6663, 0.9057, 0.9625, 1.0031], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0407, 0.0498, 0.0516, 0.0441, 0.0461, 0.0468, 0.0471], + device='cuda:3'), out_proj_covar=tensor([9.9884e-05, 1.0091e-04, 1.1206e-04, 1.2237e-04, 1.0691e-04, 1.1134e-04, + 1.1216e-04, 1.1272e-04], device='cuda:3') +2023-04-27 03:22:36,120 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.576e+02 1.859e+02 2.202e+02 4.297e+02, threshold=3.717e+02, percent-clipped=1.0 +2023-04-27 03:22:43,956 INFO [finetune.py:976] (3/7) Epoch 12, batch 4350, loss[loss=0.195, simple_loss=0.2631, pruned_loss=0.06341, over 4890.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2514, pruned_loss=0.059, over 955719.57 frames. ], batch size: 32, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:23:29,039 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67387.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:23:51,833 INFO [finetune.py:976] (3/7) Epoch 12, batch 4400, loss[loss=0.2108, simple_loss=0.2653, pruned_loss=0.0782, over 4850.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2518, pruned_loss=0.05995, over 954424.64 frames. ], batch size: 47, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:24:26,227 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67435.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:24:34,556 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67439.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:24:36,308 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.200e+02 1.735e+02 1.991e+02 2.507e+02 6.130e+02, threshold=3.981e+02, percent-clipped=5.0 +2023-04-27 03:24:56,594 INFO [finetune.py:976] (3/7) Epoch 12, batch 4450, loss[loss=0.2129, simple_loss=0.2806, pruned_loss=0.07263, over 4775.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2544, pruned_loss=0.06011, over 953650.59 frames. ], batch size: 59, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:25:48,055 INFO [finetune.py:976] (3/7) Epoch 12, batch 4500, loss[loss=0.1937, simple_loss=0.2717, pruned_loss=0.05785, over 4827.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2561, pruned_loss=0.06055, over 953147.49 frames. ], batch size: 30, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:25:53,056 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67513.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:25:55,742 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 03:26:12,839 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.294e+02 1.699e+02 1.983e+02 2.527e+02 4.329e+02, threshold=3.965e+02, percent-clipped=1.0 +2023-04-27 03:26:22,245 INFO [finetune.py:976] (3/7) Epoch 12, batch 4550, loss[loss=0.2328, simple_loss=0.2801, pruned_loss=0.09277, over 4206.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2579, pruned_loss=0.06108, over 953592.58 frames. ], batch size: 65, lr: 3.63e-03, grad_scale: 32.0 +2023-04-27 03:26:34,574 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67574.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:26:53,836 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-27 03:26:56,264 INFO [finetune.py:976] (3/7) Epoch 12, batch 4600, loss[loss=0.2295, simple_loss=0.2972, pruned_loss=0.08091, over 4815.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2564, pruned_loss=0.06023, over 954507.34 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 32.0 +2023-04-27 03:26:57,635 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3410, 1.2767, 1.4408, 1.6109, 1.6585, 1.3283, 1.0597, 1.4819], + device='cuda:3'), covar=tensor([0.0842, 0.1433, 0.0941, 0.0688, 0.0716, 0.0867, 0.0864, 0.0642], + device='cuda:3'), in_proj_covar=tensor([0.0191, 0.0200, 0.0180, 0.0172, 0.0176, 0.0184, 0.0154, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:27:00,732 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5877, 0.6465, 1.3599, 1.9612, 1.6860, 1.4688, 1.4489, 1.4806], + device='cuda:3'), covar=tensor([0.4690, 0.6868, 0.6466, 0.6630, 0.5968, 0.7969, 0.7827, 0.8213], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0405, 0.0495, 0.0514, 0.0441, 0.0461, 0.0468, 0.0470], + device='cuda:3'), out_proj_covar=tensor([9.9672e-05, 1.0065e-04, 1.1160e-04, 1.2186e-04, 1.0669e-04, 1.1116e-04, + 1.1204e-04, 1.1250e-04], device='cuda:3') +2023-04-27 03:27:01,299 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:27:15,109 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8212, 1.3190, 1.3964, 1.6035, 1.9858, 1.6707, 1.4693, 1.3177], + device='cuda:3'), covar=tensor([0.1507, 0.1496, 0.1706, 0.1248, 0.0747, 0.1495, 0.1934, 0.1863], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0317, 0.0352, 0.0295, 0.0333, 0.0315, 0.0304, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.3849e-05, 6.6771e-05, 7.5619e-05, 6.0570e-05, 6.9552e-05, 6.6873e-05, + 6.4821e-05, 7.7292e-05], device='cuda:3') +2023-04-27 03:27:20,971 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.683e+02 1.939e+02 2.322e+02 5.503e+02, threshold=3.878e+02, percent-clipped=1.0 +2023-04-27 03:27:29,855 INFO [finetune.py:976] (3/7) Epoch 12, batch 4650, loss[loss=0.1591, simple_loss=0.2252, pruned_loss=0.04647, over 4786.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2544, pruned_loss=0.05949, over 955439.85 frames. ], batch size: 25, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:27:34,156 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67661.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:27:37,131 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 03:27:46,412 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:27:54,873 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:28:04,214 INFO [finetune.py:976] (3/7) Epoch 12, batch 4700, loss[loss=0.1551, simple_loss=0.22, pruned_loss=0.04507, over 4711.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.252, pruned_loss=0.05909, over 955562.04 frames. ], batch size: 54, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:28:37,288 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67739.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:28:38,544 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67741.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:28:40,070 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.659e+02 1.914e+02 2.232e+02 4.498e+02, threshold=3.829e+02, percent-clipped=2.0 +2023-04-27 03:28:58,095 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67753.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:28:58,821 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 03:28:59,236 INFO [finetune.py:976] (3/7) Epoch 12, batch 4750, loss[loss=0.2735, simple_loss=0.3192, pruned_loss=0.1139, over 4761.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2505, pruned_loss=0.05922, over 954692.69 frames. ], batch size: 28, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:29:42,912 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67787.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:29:55,417 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6876, 1.9376, 1.0311, 1.5455, 2.0870, 1.5910, 1.5620, 1.6339], + device='cuda:3'), covar=tensor([0.0500, 0.0356, 0.0318, 0.0553, 0.0260, 0.0520, 0.0497, 0.0567], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 03:30:06,236 INFO [finetune.py:976] (3/7) Epoch 12, batch 4800, loss[loss=0.2373, simple_loss=0.2972, pruned_loss=0.08869, over 4872.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2562, pruned_loss=0.06214, over 953803.22 frames. ], batch size: 34, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:30:35,558 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.720e+02 2.004e+02 2.760e+02 5.659e+02, threshold=4.008e+02, percent-clipped=5.0 +2023-04-27 03:30:43,904 INFO [finetune.py:976] (3/7) Epoch 12, batch 4850, loss[loss=0.207, simple_loss=0.2721, pruned_loss=0.07098, over 4923.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2564, pruned_loss=0.06117, over 954026.17 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:30:54,469 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67869.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:31:05,502 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8169, 2.6348, 1.8782, 1.7977, 1.3029, 1.3297, 2.0315, 1.2457], + device='cuda:3'), covar=tensor([0.1826, 0.1441, 0.1524, 0.1974, 0.2642, 0.2191, 0.1099, 0.2213], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0214, 0.0170, 0.0204, 0.0203, 0.0185, 0.0157, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 03:31:17,828 INFO [finetune.py:976] (3/7) Epoch 12, batch 4900, loss[loss=0.2332, simple_loss=0.2945, pruned_loss=0.08592, over 4919.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2563, pruned_loss=0.06065, over 953582.79 frames. ], batch size: 33, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:31:43,017 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.676e+01 1.571e+02 1.848e+02 2.234e+02 5.872e+02, threshold=3.696e+02, percent-clipped=2.0 +2023-04-27 03:31:51,276 INFO [finetune.py:976] (3/7) Epoch 12, batch 4950, loss[loss=0.1739, simple_loss=0.2519, pruned_loss=0.04789, over 4833.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2557, pruned_loss=0.0597, over 952538.25 frames. ], batch size: 47, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:32:01,239 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8785, 1.5814, 2.0747, 2.2884, 1.5911, 1.3939, 1.7586, 1.0426], + device='cuda:3'), covar=tensor([0.0696, 0.0945, 0.0593, 0.0778, 0.0905, 0.1292, 0.0884, 0.0911], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0071, 0.0070, 0.0067, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 03:32:09,669 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:32:26,420 INFO [finetune.py:976] (3/7) Epoch 12, batch 5000, loss[loss=0.1659, simple_loss=0.2337, pruned_loss=0.04904, over 4902.00 frames. ], tot_loss[loss=0.185, simple_loss=0.253, pruned_loss=0.05852, over 952086.30 frames. ], batch size: 37, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:32:36,356 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0237, 1.0188, 1.1760, 1.1428, 0.9114, 0.8578, 1.0469, 0.7297], + device='cuda:3'), covar=tensor([0.0571, 0.0632, 0.0514, 0.0594, 0.0799, 0.1381, 0.0493, 0.0747], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0070, 0.0070, 0.0067, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 03:32:47,899 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:32:51,646 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68042.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:32:52,124 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.570e+01 1.647e+02 1.940e+02 2.431e+02 4.879e+02, threshold=3.879e+02, percent-clipped=2.0 +2023-04-27 03:32:55,702 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68048.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:32:59,804 INFO [finetune.py:976] (3/7) Epoch 12, batch 5050, loss[loss=0.1478, simple_loss=0.2193, pruned_loss=0.0381, over 4763.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2517, pruned_loss=0.05882, over 952985.12 frames. ], batch size: 26, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:33:22,294 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68087.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:33:23,466 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6804, 1.4261, 4.4764, 4.2294, 3.9014, 4.2629, 4.0816, 3.9367], + device='cuda:3'), covar=tensor([0.6731, 0.5805, 0.0946, 0.1668, 0.1126, 0.1655, 0.1441, 0.1343], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0308, 0.0404, 0.0412, 0.0352, 0.0410, 0.0316, 0.0373], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:33:26,565 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9803, 2.4929, 2.0119, 1.7931, 1.4100, 1.4208, 2.0578, 1.3637], + device='cuda:3'), covar=tensor([0.1714, 0.1417, 0.1446, 0.1889, 0.2399, 0.1981, 0.1023, 0.2056], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0212, 0.0170, 0.0203, 0.0202, 0.0184, 0.0157, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 03:33:27,713 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5115, 3.4140, 2.7923, 4.0983, 3.4666, 3.5009, 1.6537, 3.4392], + device='cuda:3'), covar=tensor([0.1953, 0.1359, 0.3907, 0.1555, 0.3366, 0.1979, 0.5738, 0.2520], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0215, 0.0248, 0.0303, 0.0298, 0.0247, 0.0271, 0.0267], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 03:33:33,622 INFO [finetune.py:976] (3/7) Epoch 12, batch 5100, loss[loss=0.2014, simple_loss=0.2753, pruned_loss=0.06376, over 4836.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2482, pruned_loss=0.0575, over 953122.99 frames. ], batch size: 33, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:33:43,247 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7985, 1.9052, 1.9345, 1.5006, 2.1403, 1.6746, 2.6756, 1.6180], + device='cuda:3'), covar=tensor([0.3914, 0.1827, 0.4924, 0.3102, 0.1600, 0.2706, 0.1418, 0.4402], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0345, 0.0427, 0.0355, 0.0378, 0.0379, 0.0373, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:34:11,408 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.851e+01 1.609e+02 1.945e+02 2.272e+02 5.546e+02, threshold=3.889e+02, percent-clipped=1.0 +2023-04-27 03:34:14,602 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68148.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:34:18,646 INFO [finetune.py:976] (3/7) Epoch 12, batch 5150, loss[loss=0.1633, simple_loss=0.2276, pruned_loss=0.04951, over 4816.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2488, pruned_loss=0.05789, over 953633.86 frames. ], batch size: 25, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:34:34,083 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68169.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:35:16,278 INFO [finetune.py:976] (3/7) Epoch 12, batch 5200, loss[loss=0.2598, simple_loss=0.3148, pruned_loss=0.1024, over 4730.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2528, pruned_loss=0.05933, over 953395.17 frames. ], batch size: 59, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:35:16,718 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 03:35:34,920 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68217.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:36:09,275 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.715e+02 2.014e+02 2.436e+02 3.767e+02, threshold=4.027e+02, percent-clipped=0.0 +2023-04-27 03:36:18,604 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5565, 3.4781, 2.6383, 4.1247, 3.5351, 3.5189, 1.4977, 3.4655], + device='cuda:3'), covar=tensor([0.1902, 0.1228, 0.3274, 0.1958, 0.3369, 0.1985, 0.6238, 0.2673], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0215, 0.0248, 0.0304, 0.0298, 0.0248, 0.0271, 0.0268], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 03:36:22,257 INFO [finetune.py:976] (3/7) Epoch 12, batch 5250, loss[loss=0.1943, simple_loss=0.2726, pruned_loss=0.05802, over 4814.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2561, pruned_loss=0.06019, over 955126.73 frames. ], batch size: 40, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:36:29,370 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0931, 0.5585, 0.9605, 0.7035, 1.1876, 0.9322, 0.8065, 0.9566], + device='cuda:3'), covar=tensor([0.1508, 0.1542, 0.1798, 0.1576, 0.0974, 0.1321, 0.1722, 0.1990], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0322, 0.0355, 0.0298, 0.0335, 0.0318, 0.0310, 0.0365], + device='cuda:3'), out_proj_covar=tensor([6.4770e-05, 6.7671e-05, 7.6197e-05, 6.1195e-05, 7.0058e-05, 6.7527e-05, + 6.5984e-05, 7.8214e-05], device='cuda:3') +2023-04-27 03:36:59,617 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-04-27 03:37:07,992 INFO [finetune.py:976] (3/7) Epoch 12, batch 5300, loss[loss=0.2022, simple_loss=0.2658, pruned_loss=0.06924, over 4803.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2574, pruned_loss=0.06065, over 955479.78 frames. ], batch size: 45, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:37:29,918 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:37:30,468 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:37:34,483 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.347e+01 1.659e+02 1.967e+02 2.376e+02 7.370e+02, threshold=3.933e+02, percent-clipped=4.0 +2023-04-27 03:37:37,614 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68348.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:37:41,847 INFO [finetune.py:976] (3/7) Epoch 12, batch 5350, loss[loss=0.1771, simple_loss=0.2509, pruned_loss=0.05165, over 4821.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2582, pruned_loss=0.06065, over 956044.16 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:38:00,874 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:38:10,025 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:38:11,943 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68399.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:38:15,660 INFO [finetune.py:976] (3/7) Epoch 12, batch 5400, loss[loss=0.1791, simple_loss=0.2405, pruned_loss=0.05879, over 4856.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2566, pruned_loss=0.06069, over 956729.85 frames. ], batch size: 47, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:38:22,486 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-04-27 03:38:41,372 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.609e+02 1.919e+02 2.411e+02 6.706e+02, threshold=3.839e+02, percent-clipped=2.0 +2023-04-27 03:38:41,452 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68443.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:38:48,652 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3695, 2.9890, 2.2591, 2.4225, 1.6245, 1.5532, 2.4123, 1.5847], + device='cuda:3'), covar=tensor([0.1760, 0.1501, 0.1454, 0.1746, 0.2473, 0.2041, 0.1047, 0.2177], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0214, 0.0170, 0.0205, 0.0204, 0.0186, 0.0158, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 03:38:49,105 INFO [finetune.py:976] (3/7) Epoch 12, batch 5450, loss[loss=0.1405, simple_loss=0.2055, pruned_loss=0.03781, over 4777.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2525, pruned_loss=0.05918, over 957745.16 frames. ], batch size: 26, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:38:52,248 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68460.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:39:33,463 INFO [finetune.py:976] (3/7) Epoch 12, batch 5500, loss[loss=0.167, simple_loss=0.2323, pruned_loss=0.0508, over 4763.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2495, pruned_loss=0.05789, over 957403.19 frames. ], batch size: 28, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:39:34,229 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0329, 2.0057, 1.9964, 1.7380, 2.3072, 1.8316, 2.8354, 1.6888], + device='cuda:3'), covar=tensor([0.3970, 0.1909, 0.4604, 0.3002, 0.1470, 0.2415, 0.1476, 0.4290], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0341, 0.0426, 0.0354, 0.0377, 0.0378, 0.0371, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:39:41,955 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 03:39:58,193 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.589e+02 1.966e+02 2.318e+02 5.103e+02, threshold=3.932e+02, percent-clipped=5.0 +2023-04-27 03:40:03,294 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 03:40:06,517 INFO [finetune.py:976] (3/7) Epoch 12, batch 5550, loss[loss=0.1718, simple_loss=0.2454, pruned_loss=0.04914, over 4906.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2512, pruned_loss=0.05873, over 958697.88 frames. ], batch size: 35, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:40:21,090 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68578.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:40:22,289 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-04-27 03:40:54,926 INFO [finetune.py:976] (3/7) Epoch 12, batch 5600, loss[loss=0.1946, simple_loss=0.2646, pruned_loss=0.06223, over 4818.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2538, pruned_loss=0.05902, over 958329.91 frames. ], batch size: 51, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:41:04,369 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 03:41:24,540 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9766, 1.2860, 1.5996, 1.6578, 2.1503, 1.7923, 1.4765, 1.4762], + device='cuda:3'), covar=tensor([0.1472, 0.1627, 0.2153, 0.1348, 0.0769, 0.1428, 0.2381, 0.2369], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0322, 0.0355, 0.0298, 0.0335, 0.0319, 0.0311, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.4699e-05, 6.7669e-05, 7.6296e-05, 6.1126e-05, 7.0076e-05, 6.7697e-05, + 6.6129e-05, 7.7863e-05], device='cuda:3') +2023-04-27 03:41:36,482 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68637.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:41:37,664 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68639.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:41:46,137 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.206e+02 1.587e+02 1.869e+02 2.457e+02 5.644e+02, threshold=3.738e+02, percent-clipped=6.0 +2023-04-27 03:41:58,626 INFO [finetune.py:976] (3/7) Epoch 12, batch 5650, loss[loss=0.1838, simple_loss=0.2615, pruned_loss=0.05311, over 4927.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2543, pruned_loss=0.0589, over 955915.89 frames. ], batch size: 36, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:42:10,144 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-04-27 03:42:33,431 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:42:33,473 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:42:45,292 INFO [finetune.py:976] (3/7) Epoch 12, batch 5700, loss[loss=0.1424, simple_loss=0.2011, pruned_loss=0.04184, over 4185.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2503, pruned_loss=0.05749, over 941224.28 frames. ], batch size: 18, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:43:00,273 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 03:43:16,428 INFO [finetune.py:976] (3/7) Epoch 13, batch 0, loss[loss=0.1986, simple_loss=0.2598, pruned_loss=0.06866, over 4889.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2598, pruned_loss=0.06866, over 4889.00 frames. ], batch size: 35, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:43:16,428 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 03:43:32,176 INFO [finetune.py:1010] (3/7) Epoch 13, validation: loss=0.1542, simple_loss=0.2264, pruned_loss=0.04102, over 2265189.00 frames. +2023-04-27 03:43:32,177 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 03:43:49,890 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.289e+01 1.578e+02 1.937e+02 2.291e+02 5.419e+02, threshold=3.875e+02, percent-clipped=2.0 +2023-04-27 03:43:50,013 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68743.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:43:51,894 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68746.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:43:56,107 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7886, 1.2768, 1.4010, 1.6496, 1.9898, 1.6114, 1.2840, 1.4175], + device='cuda:3'), covar=tensor([0.1556, 0.1582, 0.2255, 0.1213, 0.0740, 0.1549, 0.2225, 0.2040], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0318, 0.0352, 0.0294, 0.0332, 0.0316, 0.0307, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.4078e-05, 6.6941e-05, 7.5476e-05, 6.0355e-05, 6.9367e-05, 6.7107e-05, + 6.5417e-05, 7.7128e-05], device='cuda:3') +2023-04-27 03:43:57,282 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68755.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:44:12,772 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2437, 1.2548, 1.4080, 1.5497, 1.6559, 1.2204, 0.9705, 1.3389], + device='cuda:3'), covar=tensor([0.0983, 0.1473, 0.1018, 0.0687, 0.0711, 0.0997, 0.1031, 0.0741], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0201, 0.0182, 0.0173, 0.0178, 0.0184, 0.0155, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:44:15,624 INFO [finetune.py:976] (3/7) Epoch 13, batch 50, loss[loss=0.2196, simple_loss=0.2795, pruned_loss=0.07983, over 4908.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2619, pruned_loss=0.06227, over 217210.32 frames. ], batch size: 43, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:44:21,434 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68791.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:44:22,082 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8876, 1.8482, 1.7137, 1.4969, 2.0009, 1.6071, 2.5104, 1.4696], + device='cuda:3'), covar=tensor([0.4147, 0.1981, 0.5030, 0.3358, 0.1790, 0.2565, 0.1426, 0.4585], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0342, 0.0426, 0.0354, 0.0378, 0.0377, 0.0371, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:44:26,401 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9295, 1.5065, 1.9661, 2.4065, 2.0249, 1.8748, 1.9390, 1.9153], + device='cuda:3'), covar=tensor([0.5616, 0.7514, 0.7425, 0.6753, 0.6815, 0.8975, 0.9333, 0.8480], + device='cuda:3'), in_proj_covar=tensor([0.0412, 0.0407, 0.0499, 0.0514, 0.0442, 0.0462, 0.0470, 0.0473], + device='cuda:3'), out_proj_covar=tensor([9.9960e-05, 1.0106e-04, 1.1222e-04, 1.2180e-04, 1.0698e-04, 1.1149e-04, + 1.1256e-04, 1.1316e-04], device='cuda:3') +2023-04-27 03:44:46,966 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4824, 1.9221, 2.3698, 3.0210, 2.3131, 1.8454, 1.7551, 2.3489], + device='cuda:3'), covar=tensor([0.3576, 0.3587, 0.1700, 0.2726, 0.3271, 0.2954, 0.4200, 0.2331], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0248, 0.0223, 0.0317, 0.0215, 0.0229, 0.0231, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 03:44:48,041 INFO [finetune.py:976] (3/7) Epoch 13, batch 100, loss[loss=0.1797, simple_loss=0.255, pruned_loss=0.05219, over 4919.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2541, pruned_loss=0.05956, over 382082.69 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:44:55,553 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.308e+02 1.653e+02 1.937e+02 2.262e+02 3.719e+02, threshold=3.874e+02, percent-clipped=0.0 +2023-04-27 03:44:57,636 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 03:45:21,008 INFO [finetune.py:976] (3/7) Epoch 13, batch 150, loss[loss=0.1793, simple_loss=0.2509, pruned_loss=0.05388, over 4856.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2492, pruned_loss=0.058, over 509979.44 frames. ], batch size: 31, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:45:47,027 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6022, 1.4450, 0.5262, 1.3231, 1.4110, 1.4561, 1.3697, 1.3970], + device='cuda:3'), covar=tensor([0.0491, 0.0361, 0.0427, 0.0538, 0.0313, 0.0507, 0.0500, 0.0556], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0037, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 03:45:53,957 INFO [finetune.py:976] (3/7) Epoch 13, batch 200, loss[loss=0.1971, simple_loss=0.2564, pruned_loss=0.0689, over 4830.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2465, pruned_loss=0.05698, over 608366.15 frames. ], batch size: 33, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:45:55,116 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68934.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:46:01,424 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.726e+01 1.554e+02 1.965e+02 2.387e+02 1.026e+03, threshold=3.930e+02, percent-clipped=4.0 +2023-04-27 03:46:31,804 INFO [finetune.py:976] (3/7) Epoch 13, batch 250, loss[loss=0.1823, simple_loss=0.2658, pruned_loss=0.04939, over 4811.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2498, pruned_loss=0.05797, over 685837.58 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:47:12,553 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7958, 2.5541, 0.8785, 1.2061, 1.7665, 1.0971, 3.4778, 1.4150], + device='cuda:3'), covar=tensor([0.1051, 0.0847, 0.1036, 0.1946, 0.0772, 0.1537, 0.0430, 0.1075], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 03:47:36,761 INFO [finetune.py:976] (3/7) Epoch 13, batch 300, loss[loss=0.1724, simple_loss=0.2501, pruned_loss=0.04738, over 4714.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2551, pruned_loss=0.05937, over 744485.92 frames. ], batch size: 23, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:47:47,290 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69041.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:47:48,431 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.203e+02 1.674e+02 1.942e+02 2.375e+02 4.255e+02, threshold=3.885e+02, percent-clipped=1.0 +2023-04-27 03:47:48,551 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69043.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:47:59,630 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0490, 1.0035, 1.2296, 1.1238, 0.9834, 0.9078, 1.0108, 0.5173], + device='cuda:3'), covar=tensor([0.0621, 0.0601, 0.0543, 0.0538, 0.0808, 0.1231, 0.0500, 0.0754], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0071, 0.0071, 0.0067, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 03:48:05,314 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:48:39,070 INFO [finetune.py:976] (3/7) Epoch 13, batch 350, loss[loss=0.1835, simple_loss=0.2602, pruned_loss=0.0534, over 4795.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.258, pruned_loss=0.06055, over 791529.90 frames. ], batch size: 45, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:48:59,541 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69103.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:49:00,182 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69104.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:49:17,726 INFO [finetune.py:976] (3/7) Epoch 13, batch 400, loss[loss=0.1631, simple_loss=0.2413, pruned_loss=0.04238, over 4799.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2569, pruned_loss=0.05962, over 827333.20 frames. ], batch size: 45, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:49:24,721 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.707e+02 2.109e+02 2.400e+02 4.777e+02, threshold=4.219e+02, percent-clipped=1.0 +2023-04-27 03:49:38,373 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 03:49:38,764 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69162.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:49:42,427 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69168.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:49:51,421 INFO [finetune.py:976] (3/7) Epoch 13, batch 450, loss[loss=0.2375, simple_loss=0.2922, pruned_loss=0.09142, over 4850.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2555, pruned_loss=0.0595, over 854675.28 frames. ], batch size: 44, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:49:57,550 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9553, 2.5623, 1.9078, 1.7850, 1.3612, 1.3868, 1.9276, 1.2958], + device='cuda:3'), covar=tensor([0.1585, 0.1226, 0.1413, 0.1760, 0.2273, 0.1835, 0.1034, 0.1955], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0214, 0.0170, 0.0204, 0.0204, 0.0185, 0.0157, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 03:49:58,692 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 03:50:19,114 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69223.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:20,349 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:22,769 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:25,120 INFO [finetune.py:976] (3/7) Epoch 13, batch 500, loss[loss=0.1869, simple_loss=0.2497, pruned_loss=0.06204, over 4681.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2526, pruned_loss=0.0582, over 876970.95 frames. ], batch size: 23, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:50:25,825 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69234.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:31,193 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.700e+02 1.950e+02 2.352e+02 3.819e+02, threshold=3.900e+02, percent-clipped=0.0 +2023-04-27 03:50:32,976 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69245.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:50:52,429 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:57,806 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69282.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:58,358 INFO [finetune.py:976] (3/7) Epoch 13, batch 550, loss[loss=0.216, simple_loss=0.2782, pruned_loss=0.07689, over 4822.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2511, pruned_loss=0.05806, over 895105.48 frames. ], batch size: 51, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:51:00,311 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69286.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:51:05,180 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2512, 1.4374, 1.6228, 1.7911, 1.7135, 1.8308, 1.6952, 1.7048], + device='cuda:3'), covar=tensor([0.4847, 0.6047, 0.5112, 0.4850, 0.6082, 0.8340, 0.5751, 0.5574], + device='cuda:3'), in_proj_covar=tensor([0.0329, 0.0378, 0.0317, 0.0329, 0.0342, 0.0400, 0.0359, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 03:51:14,436 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69306.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:51:19,995 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8667, 2.8251, 2.2490, 3.3163, 2.8862, 2.9156, 1.3224, 2.8331], + device='cuda:3'), covar=tensor([0.2379, 0.1773, 0.3141, 0.2675, 0.3508, 0.2181, 0.6050, 0.2906], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0214, 0.0248, 0.0302, 0.0297, 0.0245, 0.0269, 0.0268], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 03:51:24,263 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:51:32,074 INFO [finetune.py:976] (3/7) Epoch 13, batch 600, loss[loss=0.1896, simple_loss=0.2453, pruned_loss=0.06692, over 4912.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2514, pruned_loss=0.05858, over 907964.83 frames. ], batch size: 32, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:51:32,805 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69334.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:51:37,070 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69341.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:51:38,138 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.579e+02 1.985e+02 2.306e+02 4.955e+02, threshold=3.970e+02, percent-clipped=1.0 +2023-04-27 03:51:38,507 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 03:51:56,889 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69368.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:04,794 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69381.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:05,905 INFO [finetune.py:976] (3/7) Epoch 13, batch 650, loss[loss=0.1913, simple_loss=0.2631, pruned_loss=0.0598, over 4723.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2546, pruned_loss=0.05996, over 917680.28 frames. ], batch size: 59, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:52:09,604 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69389.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:12,676 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69394.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:15,687 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69399.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:36,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9109, 1.6653, 2.0391, 2.3968, 2.4068, 1.7043, 1.5848, 2.0000], + device='cuda:3'), covar=tensor([0.1072, 0.1245, 0.0711, 0.0615, 0.0664, 0.1063, 0.0946, 0.0692], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0202, 0.0184, 0.0174, 0.0179, 0.0186, 0.0157, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:52:36,499 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69429.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:44,420 INFO [finetune.py:976] (3/7) Epoch 13, batch 700, loss[loss=0.1505, simple_loss=0.1889, pruned_loss=0.05601, over 4060.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2559, pruned_loss=0.06005, over 924080.74 frames. ], batch size: 17, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:52:56,108 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.279e+02 1.620e+02 1.977e+02 2.375e+02 4.653e+02, threshold=3.954e+02, percent-clipped=1.0 +2023-04-27 03:53:09,882 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69455.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:53:51,162 INFO [finetune.py:976] (3/7) Epoch 13, batch 750, loss[loss=0.239, simple_loss=0.3163, pruned_loss=0.08081, over 4911.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2568, pruned_loss=0.06014, over 931517.18 frames. ], batch size: 37, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:54:42,273 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69518.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:54:46,356 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69524.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:54:57,620 INFO [finetune.py:976] (3/7) Epoch 13, batch 800, loss[loss=0.1425, simple_loss=0.1998, pruned_loss=0.04257, over 4150.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2566, pruned_loss=0.05995, over 935793.33 frames. ], batch size: 17, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:55:09,401 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.803e+01 1.651e+02 1.914e+02 2.371e+02 3.866e+02, threshold=3.828e+02, percent-clipped=0.0 +2023-04-27 03:55:35,211 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69581.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:55:36,383 INFO [finetune.py:976] (3/7) Epoch 13, batch 850, loss[loss=0.2088, simple_loss=0.2613, pruned_loss=0.07812, over 4916.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2539, pruned_loss=0.05856, over 942737.64 frames. ], batch size: 43, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:55:47,512 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69601.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:56:07,726 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69629.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:56:10,108 INFO [finetune.py:976] (3/7) Epoch 13, batch 900, loss[loss=0.1672, simple_loss=0.2322, pruned_loss=0.0511, over 4868.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2515, pruned_loss=0.05793, over 946396.38 frames. ], batch size: 31, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:56:16,240 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.649e+02 1.999e+02 2.386e+02 5.563e+02, threshold=3.997e+02, percent-clipped=1.0 +2023-04-27 03:56:37,651 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69674.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:56:39,832 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69676.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:56:44,052 INFO [finetune.py:976] (3/7) Epoch 13, batch 950, loss[loss=0.1749, simple_loss=0.2416, pruned_loss=0.05414, over 4746.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2484, pruned_loss=0.05685, over 947917.03 frames. ], batch size: 54, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:56:49,088 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6719, 1.8577, 0.8867, 1.3270, 1.7818, 1.5746, 1.4582, 1.4987], + device='cuda:3'), covar=tensor([0.0515, 0.0351, 0.0347, 0.0573, 0.0280, 0.0531, 0.0514, 0.0581], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 03:56:53,917 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69699.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:57:11,098 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69724.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:57:12,877 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3583, 1.6057, 1.5100, 2.2539, 2.3488, 1.9155, 1.8977, 1.6898], + device='cuda:3'), covar=tensor([0.2204, 0.1792, 0.2214, 0.1596, 0.1134, 0.2011, 0.2084, 0.2232], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0316, 0.0348, 0.0291, 0.0330, 0.0313, 0.0304, 0.0358], + device='cuda:3'), out_proj_covar=tensor([6.3594e-05, 6.6397e-05, 7.4635e-05, 5.9490e-05, 6.8842e-05, 6.6584e-05, + 6.4590e-05, 7.6506e-05], device='cuda:3') +2023-04-27 03:57:18,035 INFO [finetune.py:976] (3/7) Epoch 13, batch 1000, loss[loss=0.1763, simple_loss=0.2453, pruned_loss=0.05362, over 4847.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2516, pruned_loss=0.0582, over 951445.80 frames. ], batch size: 30, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:57:19,396 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69735.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:57:24,156 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.258e+02 1.699e+02 2.018e+02 2.497e+02 4.277e+02, threshold=4.036e+02, percent-clipped=1.0 +2023-04-27 03:57:26,751 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69747.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:57:28,613 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69750.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:57:50,562 INFO [finetune.py:976] (3/7) Epoch 13, batch 1050, loss[loss=0.18, simple_loss=0.2613, pruned_loss=0.0494, over 4927.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2548, pruned_loss=0.05927, over 953038.10 frames. ], batch size: 38, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:58:04,750 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2603, 2.1377, 1.7796, 1.8093, 2.2311, 1.8652, 2.6684, 1.5780], + device='cuda:3'), covar=tensor([0.3742, 0.2168, 0.4918, 0.3563, 0.1954, 0.2335, 0.1495, 0.4669], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0344, 0.0428, 0.0356, 0.0382, 0.0380, 0.0373, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 03:58:13,046 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69818.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:58:16,690 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69824.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:58:17,381 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 03:58:23,481 INFO [finetune.py:976] (3/7) Epoch 13, batch 1100, loss[loss=0.2023, simple_loss=0.2696, pruned_loss=0.06756, over 4725.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2551, pruned_loss=0.05865, over 953330.44 frames. ], batch size: 59, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:58:30,701 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.653e+02 1.942e+02 2.253e+02 5.634e+02, threshold=3.884e+02, percent-clipped=2.0 +2023-04-27 03:58:46,674 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3831, 1.7279, 1.6080, 1.9920, 1.7782, 1.9549, 1.5795, 4.2492], + device='cuda:3'), covar=tensor([0.0570, 0.0790, 0.0796, 0.1146, 0.0668, 0.0572, 0.0727, 0.0134], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 03:58:50,906 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69866.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:59:00,300 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69872.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:59:09,274 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6442, 2.2333, 2.6718, 2.9674, 2.5878, 2.1033, 1.8784, 2.4575], + device='cuda:3'), covar=tensor([0.3487, 0.3180, 0.1581, 0.2185, 0.2624, 0.2618, 0.4021, 0.2028], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0247, 0.0222, 0.0315, 0.0213, 0.0227, 0.0230, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 03:59:12,157 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69881.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:59:13,753 INFO [finetune.py:976] (3/7) Epoch 13, batch 1150, loss[loss=0.2021, simple_loss=0.2722, pruned_loss=0.06597, over 4738.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2548, pruned_loss=0.05845, over 952223.38 frames. ], batch size: 27, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:59:41,654 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69901.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:59:57,162 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-04-27 04:00:16,221 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:00:16,247 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69929.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 04:00:18,639 INFO [finetune.py:976] (3/7) Epoch 13, batch 1200, loss[loss=0.2005, simple_loss=0.2692, pruned_loss=0.06587, over 4745.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2535, pruned_loss=0.05783, over 953053.19 frames. ], batch size: 27, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:00:32,298 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.675e+02 1.952e+02 2.332e+02 4.660e+02, threshold=3.905e+02, percent-clipped=1.0 +2023-04-27 04:00:36,039 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69949.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:00:52,482 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69976.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:00:53,026 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:00:57,154 INFO [finetune.py:976] (3/7) Epoch 13, batch 1250, loss[loss=0.1782, simple_loss=0.2423, pruned_loss=0.05704, over 4915.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2506, pruned_loss=0.05656, over 952865.71 frames. ], batch size: 37, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:01:05,966 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-04-27 04:01:26,386 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:01:26,418 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4711, 3.3427, 0.9425, 1.9543, 1.8175, 2.3567, 1.9217, 1.0216], + device='cuda:3'), covar=tensor([0.1371, 0.0984, 0.1900, 0.1144, 0.1122, 0.1023, 0.1316, 0.2056], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0245, 0.0139, 0.0121, 0.0132, 0.0152, 0.0116, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:01:26,432 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:01:30,022 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70030.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:01:31,779 INFO [finetune.py:976] (3/7) Epoch 13, batch 1300, loss[loss=0.1448, simple_loss=0.216, pruned_loss=0.03677, over 4105.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.248, pruned_loss=0.05567, over 953224.95 frames. ], batch size: 17, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:01:39,420 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.185e+02 1.660e+02 1.869e+02 2.265e+02 4.379e+02, threshold=3.739e+02, percent-clipped=1.0 +2023-04-27 04:01:45,204 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70050.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:01:48,345 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0696, 2.6088, 2.0902, 2.3952, 1.8877, 2.1110, 2.1819, 1.7101], + device='cuda:3'), covar=tensor([0.1859, 0.1117, 0.0879, 0.1079, 0.2967, 0.1069, 0.1836, 0.2545], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0314, 0.0227, 0.0286, 0.0316, 0.0266, 0.0257, 0.0274], + device='cuda:3'), out_proj_covar=tensor([1.1837e-04, 1.2531e-04, 9.0482e-05, 1.1424e-04, 1.2885e-04, 1.0638e-04, + 1.0424e-04, 1.0947e-04], device='cuda:3') +2023-04-27 04:01:58,564 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70072.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:02:05,230 INFO [finetune.py:976] (3/7) Epoch 13, batch 1350, loss[loss=0.1335, simple_loss=0.2023, pruned_loss=0.03238, over 4744.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2486, pruned_loss=0.05642, over 953062.72 frames. ], batch size: 23, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:02:16,792 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70098.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:02:24,497 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:02:34,216 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8546, 4.2509, 1.0858, 2.1856, 2.2540, 2.7422, 2.3729, 1.0684], + device='cuda:3'), covar=tensor([0.1365, 0.0852, 0.1880, 0.1223, 0.1049, 0.1050, 0.1327, 0.1946], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0245, 0.0138, 0.0121, 0.0133, 0.0152, 0.0116, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:02:38,384 INFO [finetune.py:976] (3/7) Epoch 13, batch 1400, loss[loss=0.1743, simple_loss=0.2575, pruned_loss=0.04553, over 4902.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2527, pruned_loss=0.05799, over 955166.96 frames. ], batch size: 37, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:02:45,033 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.629e+02 2.088e+02 2.462e+02 4.428e+02, threshold=4.176e+02, percent-clipped=4.0 +2023-04-27 04:02:56,693 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3746, 1.5923, 1.6268, 1.8807, 1.6581, 2.0548, 1.4584, 3.6049], + device='cuda:3'), covar=tensor([0.0612, 0.0786, 0.0844, 0.1197, 0.0664, 0.0472, 0.0757, 0.0171], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 04:03:04,709 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:03:11,826 INFO [finetune.py:976] (3/7) Epoch 13, batch 1450, loss[loss=0.1722, simple_loss=0.2455, pruned_loss=0.04948, over 4826.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2542, pruned_loss=0.05824, over 954924.75 frames. ], batch size: 33, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:03:18,056 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8029, 1.3381, 1.8337, 2.3133, 1.9281, 1.7270, 1.7841, 1.7577], + device='cuda:3'), covar=tensor([0.5155, 0.7279, 0.7321, 0.6799, 0.6531, 0.9314, 0.8857, 0.8837], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0404, 0.0491, 0.0511, 0.0439, 0.0459, 0.0465, 0.0469], + device='cuda:3'), out_proj_covar=tensor([9.9084e-05, 1.0011e-04, 1.1055e-04, 1.2106e-04, 1.0626e-04, 1.1069e-04, + 1.1123e-04, 1.1223e-04], device='cuda:3') +2023-04-27 04:03:45,321 INFO [finetune.py:976] (3/7) Epoch 13, batch 1500, loss[loss=0.1536, simple_loss=0.2267, pruned_loss=0.0402, over 4795.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.255, pruned_loss=0.05869, over 954659.80 frames. ], batch size: 29, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:03:51,423 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.692e+02 1.982e+02 2.371e+02 3.829e+02, threshold=3.965e+02, percent-clipped=0.0 +2023-04-27 04:04:10,482 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:04:29,228 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0148, 2.5645, 2.2105, 2.4001, 1.7787, 2.1351, 2.2056, 1.8512], + device='cuda:3'), covar=tensor([0.1978, 0.1183, 0.0838, 0.1246, 0.3488, 0.1083, 0.2171, 0.2628], + device='cuda:3'), in_proj_covar=tensor([0.0296, 0.0316, 0.0228, 0.0288, 0.0318, 0.0267, 0.0259, 0.0275], + device='cuda:3'), out_proj_covar=tensor([1.1948e-04, 1.2617e-04, 9.1016e-05, 1.1505e-04, 1.2960e-04, 1.0692e-04, + 1.0499e-04, 1.0994e-04], device='cuda:3') +2023-04-27 04:04:32,887 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1399, 1.5636, 1.4342, 1.7248, 1.5500, 1.8381, 1.3740, 3.3457], + device='cuda:3'), covar=tensor([0.0646, 0.0782, 0.0786, 0.1170, 0.0652, 0.0552, 0.0736, 0.0167], + device='cuda:3'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 04:04:40,470 INFO [finetune.py:976] (3/7) Epoch 13, batch 1550, loss[loss=0.1871, simple_loss=0.2544, pruned_loss=0.05987, over 4912.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.256, pruned_loss=0.05913, over 954044.60 frames. ], batch size: 46, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:04:57,735 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.5014, 4.4459, 3.2253, 5.2418, 4.5518, 4.5152, 2.0148, 4.4433], + device='cuda:3'), covar=tensor([0.1354, 0.1043, 0.2967, 0.0826, 0.3173, 0.1696, 0.5668, 0.2162], + device='cuda:3'), in_proj_covar=tensor([0.0241, 0.0213, 0.0247, 0.0300, 0.0296, 0.0246, 0.0268, 0.0267], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 04:05:17,658 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:05:18,654 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 04:05:28,448 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70330.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 04:05:30,176 INFO [finetune.py:976] (3/7) Epoch 13, batch 1600, loss[loss=0.1741, simple_loss=0.2507, pruned_loss=0.04876, over 4889.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2546, pruned_loss=0.05912, over 955951.78 frames. ], batch size: 32, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:05:41,008 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.445e+01 1.654e+02 2.055e+02 2.369e+02 4.520e+02, threshold=4.109e+02, percent-clipped=1.0 +2023-04-27 04:05:50,526 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3599, 2.6824, 1.1978, 1.6014, 2.2930, 1.2861, 3.6806, 1.7946], + device='cuda:3'), covar=tensor([0.0588, 0.0653, 0.0759, 0.1209, 0.0452, 0.0999, 0.0245, 0.0652], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 04:06:08,572 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4210, 1.2460, 1.6267, 1.5409, 1.3505, 1.2037, 1.2531, 0.7416], + device='cuda:3'), covar=tensor([0.0593, 0.0918, 0.0496, 0.0723, 0.0869, 0.1202, 0.0753, 0.0743], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0070, 0.0070, 0.0066, 0.0074, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:06:16,252 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70378.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 04:06:19,130 INFO [finetune.py:976] (3/7) Epoch 13, batch 1650, loss[loss=0.2122, simple_loss=0.2688, pruned_loss=0.07778, over 4727.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2512, pruned_loss=0.05871, over 954026.38 frames. ], batch size: 54, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:06:46,143 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 04:06:52,935 INFO [finetune.py:976] (3/7) Epoch 13, batch 1700, loss[loss=0.2028, simple_loss=0.2577, pruned_loss=0.07396, over 4775.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2489, pruned_loss=0.05815, over 950980.71 frames. ], batch size: 26, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:06:59,066 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.620e+02 1.879e+02 2.417e+02 6.028e+02, threshold=3.758e+02, percent-clipped=1.0 +2023-04-27 04:07:16,024 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:07:26,840 INFO [finetune.py:976] (3/7) Epoch 13, batch 1750, loss[loss=0.1345, simple_loss=0.2051, pruned_loss=0.03198, over 4724.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2505, pruned_loss=0.05826, over 950726.72 frames. ], batch size: 23, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:07:50,700 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5980, 0.9855, 1.5463, 2.0742, 1.7122, 1.5726, 1.5963, 1.5872], + device='cuda:3'), covar=tensor([0.5378, 0.7594, 0.7373, 0.6950, 0.6836, 0.8386, 0.9032, 0.9262], + device='cuda:3'), in_proj_covar=tensor([0.0412, 0.0408, 0.0497, 0.0517, 0.0443, 0.0464, 0.0469, 0.0474], + device='cuda:3'), out_proj_covar=tensor([9.9972e-05, 1.0107e-04, 1.1182e-04, 1.2242e-04, 1.0720e-04, 1.1180e-04, + 1.1226e-04, 1.1333e-04], device='cuda:3') +2023-04-27 04:08:00,058 INFO [finetune.py:976] (3/7) Epoch 13, batch 1800, loss[loss=0.1836, simple_loss=0.2449, pruned_loss=0.06119, over 4885.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2541, pruned_loss=0.05922, over 954196.92 frames. ], batch size: 32, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:08:06,010 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.577e+02 1.883e+02 2.332e+02 3.915e+02, threshold=3.766e+02, percent-clipped=2.0 +2023-04-27 04:08:33,367 INFO [finetune.py:976] (3/7) Epoch 13, batch 1850, loss[loss=0.1466, simple_loss=0.2334, pruned_loss=0.02991, over 4897.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2556, pruned_loss=0.05908, over 955772.68 frames. ], batch size: 43, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:08:37,178 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 04:08:54,137 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:08:55,372 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0380, 4.3324, 0.7849, 2.1818, 2.3667, 2.7486, 2.5013, 0.8928], + device='cuda:3'), covar=tensor([0.1275, 0.0799, 0.2220, 0.1287, 0.1108, 0.1139, 0.1405, 0.2161], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0247, 0.0139, 0.0122, 0.0133, 0.0152, 0.0118, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:09:06,409 INFO [finetune.py:976] (3/7) Epoch 13, batch 1900, loss[loss=0.1905, simple_loss=0.2627, pruned_loss=0.05918, over 4876.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2559, pruned_loss=0.05852, over 957098.85 frames. ], batch size: 35, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:09:12,476 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.737e+02 2.053e+02 2.436e+02 7.397e+02, threshold=4.106e+02, percent-clipped=5.0 +2023-04-27 04:10:02,034 INFO [finetune.py:976] (3/7) Epoch 13, batch 1950, loss[loss=0.1827, simple_loss=0.2257, pruned_loss=0.06985, over 4166.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2549, pruned_loss=0.05833, over 955891.37 frames. ], batch size: 18, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:10:02,781 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9263, 1.7305, 2.1350, 2.3505, 2.0280, 1.8674, 2.0499, 1.9620], + device='cuda:3'), covar=tensor([0.5342, 0.8089, 0.7952, 0.7033, 0.6555, 1.0244, 0.9434, 0.9819], + device='cuda:3'), in_proj_covar=tensor([0.0409, 0.0404, 0.0493, 0.0512, 0.0439, 0.0460, 0.0467, 0.0471], + device='cuda:3'), out_proj_covar=tensor([9.9273e-05, 1.0022e-04, 1.1089e-04, 1.2142e-04, 1.0610e-04, 1.1094e-04, + 1.1163e-04, 1.1252e-04], device='cuda:3') +2023-04-27 04:10:39,638 INFO [finetune.py:976] (3/7) Epoch 13, batch 2000, loss[loss=0.1608, simple_loss=0.2268, pruned_loss=0.04738, over 4928.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2515, pruned_loss=0.05726, over 956228.23 frames. ], batch size: 33, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:10:52,182 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.325e+01 1.560e+02 1.780e+02 2.155e+02 4.897e+02, threshold=3.560e+02, percent-clipped=2.0 +2023-04-27 04:11:15,889 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:11:28,145 INFO [finetune.py:976] (3/7) Epoch 13, batch 2050, loss[loss=0.1304, simple_loss=0.2145, pruned_loss=0.02322, over 4757.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2482, pruned_loss=0.05578, over 956619.59 frames. ], batch size: 27, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:11:29,374 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1846, 1.5386, 1.4058, 1.7480, 1.5929, 1.7365, 1.3825, 3.0563], + device='cuda:3'), covar=tensor([0.0663, 0.0783, 0.0788, 0.1141, 0.0636, 0.0482, 0.0757, 0.0157], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 04:11:48,338 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70814.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:12:01,814 INFO [finetune.py:976] (3/7) Epoch 13, batch 2100, loss[loss=0.2462, simple_loss=0.3098, pruned_loss=0.09131, over 4909.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2482, pruned_loss=0.05601, over 956627.72 frames. ], batch size: 43, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:12:08,819 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.590e+02 1.945e+02 2.393e+02 5.833e+02, threshold=3.889e+02, percent-clipped=3.0 +2023-04-27 04:12:24,912 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5272, 1.1139, 1.6445, 2.0753, 1.6663, 1.5437, 1.5929, 1.5780], + device='cuda:3'), covar=tensor([0.4980, 0.6782, 0.6235, 0.6111, 0.5994, 0.8059, 0.7946, 0.8559], + device='cuda:3'), in_proj_covar=tensor([0.0411, 0.0406, 0.0494, 0.0514, 0.0441, 0.0461, 0.0468, 0.0471], + device='cuda:3'), out_proj_covar=tensor([9.9664e-05, 1.0053e-04, 1.1126e-04, 1.2178e-04, 1.0654e-04, 1.1110e-04, + 1.1192e-04, 1.1268e-04], device='cuda:3') +2023-04-27 04:12:35,717 INFO [finetune.py:976] (3/7) Epoch 13, batch 2150, loss[loss=0.1955, simple_loss=0.28, pruned_loss=0.0555, over 4807.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2534, pruned_loss=0.05808, over 956874.36 frames. ], batch size: 41, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:12:46,831 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 04:12:54,639 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-27 04:12:56,816 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:13:00,052 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-27 04:13:08,517 INFO [finetune.py:976] (3/7) Epoch 13, batch 2200, loss[loss=0.2073, simple_loss=0.2658, pruned_loss=0.07441, over 4812.00 frames. ], tot_loss[loss=0.188, simple_loss=0.257, pruned_loss=0.0595, over 957342.98 frames. ], batch size: 38, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:13:16,547 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.168e+02 1.648e+02 2.003e+02 2.484e+02 4.937e+02, threshold=4.005e+02, percent-clipped=2.0 +2023-04-27 04:13:29,541 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70964.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:13:41,740 INFO [finetune.py:976] (3/7) Epoch 13, batch 2250, loss[loss=0.1963, simple_loss=0.2593, pruned_loss=0.06667, over 4772.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2566, pruned_loss=0.05896, over 956532.34 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:13:49,277 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70992.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:14:13,317 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71030.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:14:14,996 INFO [finetune.py:976] (3/7) Epoch 13, batch 2300, loss[loss=0.1688, simple_loss=0.2352, pruned_loss=0.05125, over 4002.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2561, pruned_loss=0.05874, over 954248.07 frames. ], batch size: 17, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:14:23,496 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.522e+02 1.891e+02 2.315e+02 3.821e+02, threshold=3.782e+02, percent-clipped=0.0 +2023-04-27 04:14:30,593 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71053.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:14:59,477 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4587, 1.4099, 4.1585, 3.8618, 3.6776, 3.9629, 3.8976, 3.6443], + device='cuda:3'), covar=tensor([0.7102, 0.5670, 0.1159, 0.1922, 0.1212, 0.1600, 0.1404, 0.1608], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0305, 0.0402, 0.0407, 0.0348, 0.0406, 0.0312, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:15:10,016 INFO [finetune.py:976] (3/7) Epoch 13, batch 2350, loss[loss=0.175, simple_loss=0.2476, pruned_loss=0.05118, over 4749.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2544, pruned_loss=0.05843, over 952517.19 frames. ], batch size: 27, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:15:20,868 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71091.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:15:44,399 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 04:15:55,081 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 04:15:57,285 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5254, 1.9886, 1.7109, 1.9090, 1.5694, 1.6745, 1.6674, 1.4813], + device='cuda:3'), covar=tensor([0.2102, 0.1345, 0.0913, 0.1173, 0.3365, 0.1133, 0.1939, 0.2490], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0314, 0.0227, 0.0285, 0.0315, 0.0266, 0.0257, 0.0274], + device='cuda:3'), out_proj_covar=tensor([1.1832e-04, 1.2528e-04, 9.0870e-05, 1.1405e-04, 1.2843e-04, 1.0660e-04, + 1.0430e-04, 1.0970e-04], device='cuda:3') +2023-04-27 04:16:15,379 INFO [finetune.py:976] (3/7) Epoch 13, batch 2400, loss[loss=0.1292, simple_loss=0.198, pruned_loss=0.03024, over 4934.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2509, pruned_loss=0.05716, over 954473.55 frames. ], batch size: 33, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:16:26,848 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.510e+02 1.853e+02 2.211e+02 4.308e+02, threshold=3.705e+02, percent-clipped=1.0 +2023-04-27 04:16:46,921 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7575, 1.3544, 1.3026, 1.5222, 1.9916, 1.6157, 1.3737, 1.3113], + device='cuda:3'), covar=tensor([0.1601, 0.1391, 0.1950, 0.1395, 0.0709, 0.1428, 0.1825, 0.2021], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0318, 0.0352, 0.0294, 0.0332, 0.0315, 0.0305, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.4233e-05, 6.6815e-05, 7.5667e-05, 6.0336e-05, 6.9437e-05, 6.6937e-05, + 6.4749e-05, 7.7435e-05], device='cuda:3') +2023-04-27 04:16:54,090 INFO [finetune.py:976] (3/7) Epoch 13, batch 2450, loss[loss=0.1619, simple_loss=0.2384, pruned_loss=0.04267, over 4938.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2493, pruned_loss=0.05702, over 956549.43 frames. ], batch size: 38, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:17:28,065 INFO [finetune.py:976] (3/7) Epoch 13, batch 2500, loss[loss=0.1862, simple_loss=0.2618, pruned_loss=0.05528, over 4792.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2513, pruned_loss=0.05795, over 955997.62 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:17:34,108 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.448e+01 1.657e+02 1.843e+02 2.125e+02 3.906e+02, threshold=3.687e+02, percent-clipped=2.0 +2023-04-27 04:18:01,396 INFO [finetune.py:976] (3/7) Epoch 13, batch 2550, loss[loss=0.2086, simple_loss=0.2798, pruned_loss=0.06867, over 4809.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2545, pruned_loss=0.05871, over 955464.29 frames. ], batch size: 45, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:18:34,377 INFO [finetune.py:976] (3/7) Epoch 13, batch 2600, loss[loss=0.1853, simple_loss=0.2557, pruned_loss=0.05747, over 4910.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.255, pruned_loss=0.05877, over 955567.06 frames. ], batch size: 37, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:18:40,523 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.287e+02 1.728e+02 1.986e+02 2.411e+02 4.998e+02, threshold=3.972e+02, percent-clipped=3.0 +2023-04-27 04:18:43,642 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71348.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:19:06,043 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-27 04:19:07,612 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7267, 1.9844, 1.0163, 1.4185, 2.0628, 1.6273, 1.5447, 1.6432], + device='cuda:3'), covar=tensor([0.0489, 0.0339, 0.0308, 0.0526, 0.0253, 0.0488, 0.0475, 0.0528], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 04:19:08,105 INFO [finetune.py:976] (3/7) Epoch 13, batch 2650, loss[loss=0.146, simple_loss=0.2293, pruned_loss=0.03138, over 4848.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2556, pruned_loss=0.05894, over 953978.23 frames. ], batch size: 49, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:19:10,021 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71386.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:19:37,263 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 04:19:41,957 INFO [finetune.py:976] (3/7) Epoch 13, batch 2700, loss[loss=0.1606, simple_loss=0.2334, pruned_loss=0.04391, over 4748.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2547, pruned_loss=0.05832, over 954468.00 frames. ], batch size: 27, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:19:47,044 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1829, 1.4073, 1.7607, 1.8691, 1.8007, 1.9129, 1.8007, 1.8140], + device='cuda:3'), covar=tensor([0.4464, 0.5958, 0.5038, 0.5245, 0.5971, 0.7824, 0.5510, 0.5331], + device='cuda:3'), in_proj_covar=tensor([0.0330, 0.0378, 0.0318, 0.0331, 0.0342, 0.0400, 0.0359, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 04:19:48,071 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.575e+02 1.881e+02 2.224e+02 4.491e+02, threshold=3.762e+02, percent-clipped=1.0 +2023-04-27 04:20:00,444 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0126, 1.7598, 2.0670, 2.3884, 2.4496, 1.8730, 1.6496, 2.1411], + device='cuda:3'), covar=tensor([0.0859, 0.1175, 0.0723, 0.0651, 0.0578, 0.0898, 0.0845, 0.0615], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0202, 0.0183, 0.0174, 0.0179, 0.0184, 0.0156, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:20:30,301 INFO [finetune.py:976] (3/7) Epoch 13, batch 2750, loss[loss=0.2076, simple_loss=0.2469, pruned_loss=0.08415, over 4063.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2515, pruned_loss=0.05714, over 954285.14 frames. ], batch size: 17, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:20:56,450 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 04:20:57,974 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-04-27 04:21:21,453 INFO [finetune.py:976] (3/7) Epoch 13, batch 2800, loss[loss=0.1636, simple_loss=0.2258, pruned_loss=0.05073, over 4766.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.248, pruned_loss=0.05653, over 954688.94 frames. ], batch size: 59, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:21:33,185 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.551e+02 1.952e+02 2.336e+02 5.892e+02, threshold=3.903e+02, percent-clipped=4.0 +2023-04-27 04:21:45,751 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6197, 3.6299, 1.0346, 1.9762, 2.0273, 2.6535, 2.0771, 1.1391], + device='cuda:3'), covar=tensor([0.1430, 0.1071, 0.2080, 0.1240, 0.1105, 0.0962, 0.1396, 0.1936], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0247, 0.0140, 0.0122, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:22:06,642 INFO [finetune.py:976] (3/7) Epoch 13, batch 2850, loss[loss=0.2462, simple_loss=0.2914, pruned_loss=0.1006, over 4898.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2476, pruned_loss=0.05681, over 953774.86 frames. ], batch size: 32, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:22:21,965 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.07 vs. limit=2.0 +2023-04-27 04:22:28,012 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:22:36,047 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5409, 1.7174, 0.8504, 1.2401, 1.7332, 1.4295, 1.3113, 1.3718], + device='cuda:3'), covar=tensor([0.0545, 0.0369, 0.0369, 0.0590, 0.0294, 0.0548, 0.0523, 0.0578], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 04:22:40,693 INFO [finetune.py:976] (3/7) Epoch 13, batch 2900, loss[loss=0.2032, simple_loss=0.2764, pruned_loss=0.06495, over 4822.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2507, pruned_loss=0.05777, over 954421.30 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:22:46,790 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.322e+02 1.703e+02 2.022e+02 2.297e+02 3.791e+02, threshold=4.044e+02, percent-clipped=0.0 +2023-04-27 04:22:49,909 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71648.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:09,476 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71677.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:13,482 INFO [finetune.py:976] (3/7) Epoch 13, batch 2950, loss[loss=0.2031, simple_loss=0.2743, pruned_loss=0.06592, over 4814.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2543, pruned_loss=0.05889, over 955770.87 frames. ], batch size: 40, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:23:15,367 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71686.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:21,434 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71696.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:30,976 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 04:23:34,311 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1635, 1.6494, 2.1108, 2.5500, 2.1830, 1.5802, 1.3330, 2.0092], + device='cuda:3'), covar=tensor([0.3315, 0.3294, 0.1615, 0.2225, 0.2455, 0.2543, 0.4378, 0.1926], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0246, 0.0221, 0.0314, 0.0213, 0.0228, 0.0229, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 04:23:34,326 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6686, 1.2966, 1.8864, 2.1848, 1.8419, 1.7577, 1.8299, 1.7628], + device='cuda:3'), covar=tensor([0.4786, 0.6728, 0.6548, 0.6338, 0.5634, 0.8041, 0.7447, 0.8266], + device='cuda:3'), in_proj_covar=tensor([0.0412, 0.0408, 0.0493, 0.0515, 0.0441, 0.0463, 0.0469, 0.0472], + device='cuda:3'), out_proj_covar=tensor([9.9758e-05, 1.0094e-04, 1.1103e-04, 1.2201e-04, 1.0668e-04, 1.1140e-04, + 1.1213e-04, 1.1291e-04], device='cuda:3') +2023-04-27 04:23:45,589 INFO [finetune.py:976] (3/7) Epoch 13, batch 3000, loss[loss=0.1582, simple_loss=0.219, pruned_loss=0.04868, over 4396.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2559, pruned_loss=0.05926, over 956590.57 frames. ], batch size: 19, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:23:45,589 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 04:23:50,756 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7950, 1.6539, 1.8733, 2.1664, 2.1314, 1.6546, 1.2858, 1.9633], + device='cuda:3'), covar=tensor([0.0830, 0.1208, 0.0755, 0.0615, 0.0619, 0.0925, 0.0968, 0.0565], + device='cuda:3'), in_proj_covar=tensor([0.0193, 0.0202, 0.0184, 0.0175, 0.0180, 0.0185, 0.0157, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:23:54,329 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1483, 1.6686, 1.9751, 2.3252, 1.9088, 1.5957, 1.1471, 1.6995], + device='cuda:3'), covar=tensor([0.3299, 0.3477, 0.1612, 0.2298, 0.2938, 0.2764, 0.4537, 0.2419], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0246, 0.0221, 0.0315, 0.0214, 0.0229, 0.0229, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 04:23:56,065 INFO [finetune.py:1010] (3/7) Epoch 13, validation: loss=0.1517, simple_loss=0.224, pruned_loss=0.03973, over 2265189.00 frames. +2023-04-27 04:23:56,066 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 04:23:56,741 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71734.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:58,066 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9180, 1.9343, 2.3757, 2.5513, 1.7964, 1.6524, 1.8366, 1.0052], + device='cuda:3'), covar=tensor([0.0665, 0.0776, 0.0447, 0.0732, 0.0943, 0.1255, 0.0802, 0.0907], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0067, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:24:02,634 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1854, 2.7384, 1.0735, 1.5163, 2.2685, 1.2592, 3.6342, 1.8616], + device='cuda:3'), covar=tensor([0.0670, 0.0662, 0.0869, 0.1376, 0.0477, 0.1062, 0.0268, 0.0643], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 04:24:03,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.669e+02 1.992e+02 2.305e+02 4.949e+02, threshold=3.985e+02, percent-clipped=1.0 +2023-04-27 04:24:27,545 INFO [finetune.py:976] (3/7) Epoch 13, batch 3050, loss[loss=0.1833, simple_loss=0.2441, pruned_loss=0.06126, over 4810.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2575, pruned_loss=0.05953, over 956473.70 frames. ], batch size: 41, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:24:47,256 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1052, 2.9101, 2.0291, 2.2239, 1.6022, 1.4646, 2.2946, 1.5092], + device='cuda:3'), covar=tensor([0.1734, 0.1422, 0.1470, 0.1738, 0.2332, 0.1953, 0.0999, 0.2095], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0213, 0.0169, 0.0203, 0.0201, 0.0183, 0.0157, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 04:25:00,523 INFO [finetune.py:976] (3/7) Epoch 13, batch 3100, loss[loss=0.2058, simple_loss=0.2703, pruned_loss=0.07068, over 4751.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2557, pruned_loss=0.05921, over 955911.08 frames. ], batch size: 27, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:25:08,968 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.880e+01 1.511e+02 1.834e+02 2.152e+02 3.267e+02, threshold=3.669e+02, percent-clipped=0.0 +2023-04-27 04:25:12,013 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3581, 1.6359, 1.4987, 1.9559, 1.8466, 2.0853, 1.5550, 3.9436], + device='cuda:3'), covar=tensor([0.0569, 0.0783, 0.0756, 0.1134, 0.0599, 0.0598, 0.0753, 0.0174], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0039, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 04:25:20,699 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0919, 2.7131, 2.2187, 2.5653, 1.9217, 2.1454, 2.1940, 1.8077], + device='cuda:3'), covar=tensor([0.2067, 0.1435, 0.0998, 0.1117, 0.3323, 0.1346, 0.2210, 0.2790], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0311, 0.0225, 0.0283, 0.0312, 0.0265, 0.0255, 0.0271], + device='cuda:3'), out_proj_covar=tensor([1.1757e-04, 1.2408e-04, 9.0021e-05, 1.1311e-04, 1.2734e-04, 1.0599e-04, + 1.0323e-04, 1.0845e-04], device='cuda:3') +2023-04-27 04:25:30,730 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-04-27 04:25:54,816 INFO [finetune.py:976] (3/7) Epoch 13, batch 3150, loss[loss=0.2127, simple_loss=0.2832, pruned_loss=0.07113, over 4821.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2525, pruned_loss=0.05776, over 954892.82 frames. ], batch size: 55, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:26:48,053 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71919.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:27:02,054 INFO [finetune.py:976] (3/7) Epoch 13, batch 3200, loss[loss=0.1766, simple_loss=0.2398, pruned_loss=0.05673, over 4781.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2489, pruned_loss=0.05647, over 953469.91 frames. ], batch size: 26, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:27:08,184 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.251e+02 1.591e+02 1.837e+02 2.276e+02 4.272e+02, threshold=3.675e+02, percent-clipped=2.0 +2023-04-27 04:27:28,807 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:27:33,782 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:27:35,468 INFO [finetune.py:976] (3/7) Epoch 13, batch 3250, loss[loss=0.1994, simple_loss=0.2584, pruned_loss=0.07021, over 4902.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2478, pruned_loss=0.05581, over 954305.52 frames. ], batch size: 35, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:28:10,359 INFO [finetune.py:976] (3/7) Epoch 13, batch 3300, loss[loss=0.1809, simple_loss=0.2625, pruned_loss=0.04966, over 4863.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.25, pruned_loss=0.05652, over 954609.78 frames. ], batch size: 34, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:28:16,990 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.593e+02 1.877e+02 2.325e+02 3.643e+02, threshold=3.754e+02, percent-clipped=0.0 +2023-04-27 04:28:28,471 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2614, 2.9844, 0.8865, 1.6512, 2.1058, 1.3428, 3.8384, 1.8930], + device='cuda:3'), covar=tensor([0.0649, 0.0737, 0.0914, 0.1216, 0.0486, 0.0927, 0.0241, 0.0554], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 04:28:43,989 INFO [finetune.py:976] (3/7) Epoch 13, batch 3350, loss[loss=0.231, simple_loss=0.2888, pruned_loss=0.08654, over 4138.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.253, pruned_loss=0.05816, over 951675.79 frames. ], batch size: 65, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:29:08,108 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0008, 2.2781, 1.8322, 1.7275, 2.1021, 1.6899, 2.7092, 1.4485], + device='cuda:3'), covar=tensor([0.3342, 0.1279, 0.4077, 0.2724, 0.1671, 0.2421, 0.1345, 0.4840], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0346, 0.0427, 0.0357, 0.0379, 0.0382, 0.0372, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:29:15,415 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:29:17,727 INFO [finetune.py:976] (3/7) Epoch 13, batch 3400, loss[loss=0.2047, simple_loss=0.2713, pruned_loss=0.0691, over 4827.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2543, pruned_loss=0.05897, over 951164.93 frames. ], batch size: 49, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:29:24,412 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.571e+02 1.878e+02 2.335e+02 4.983e+02, threshold=3.756e+02, percent-clipped=1.0 +2023-04-27 04:29:27,034 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1129, 1.5545, 1.9904, 2.2613, 1.9264, 1.5331, 1.1540, 1.7272], + device='cuda:3'), covar=tensor([0.3479, 0.3617, 0.1718, 0.2629, 0.2833, 0.2884, 0.4455, 0.2327], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0247, 0.0221, 0.0315, 0.0213, 0.0228, 0.0228, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 04:29:51,374 INFO [finetune.py:976] (3/7) Epoch 13, batch 3450, loss[loss=0.2287, simple_loss=0.2943, pruned_loss=0.08156, over 4894.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2543, pruned_loss=0.05857, over 952976.63 frames. ], batch size: 35, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:29:55,737 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72190.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:30:01,986 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.85 vs. limit=5.0 +2023-04-27 04:30:02,526 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5937, 1.4128, 1.9888, 1.9195, 1.4355, 1.3003, 1.5891, 1.1046], + device='cuda:3'), covar=tensor([0.0551, 0.0839, 0.0405, 0.0615, 0.0781, 0.1123, 0.0651, 0.0685], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:30:24,560 INFO [finetune.py:976] (3/7) Epoch 13, batch 3500, loss[loss=0.1559, simple_loss=0.2313, pruned_loss=0.04023, over 4781.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2524, pruned_loss=0.0582, over 953728.64 frames. ], batch size: 28, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:30:31,095 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.651e+02 2.080e+02 2.507e+02 4.410e+02, threshold=4.160e+02, percent-clipped=6.0 +2023-04-27 04:31:07,543 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-27 04:31:08,031 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:31:09,845 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72275.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:31:12,377 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1738, 1.5587, 1.4613, 1.7693, 1.6459, 2.0640, 1.3778, 3.7221], + device='cuda:3'), covar=tensor([0.0643, 0.0794, 0.0810, 0.1198, 0.0664, 0.0500, 0.0750, 0.0131], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 04:31:20,404 INFO [finetune.py:976] (3/7) Epoch 13, batch 3550, loss[loss=0.1832, simple_loss=0.2435, pruned_loss=0.0615, over 4838.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2508, pruned_loss=0.05815, over 956142.21 frames. ], batch size: 47, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:31:41,861 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7126, 2.1124, 1.5816, 1.4420, 1.2259, 1.2315, 1.5286, 1.1690], + device='cuda:3'), covar=tensor([0.1677, 0.1279, 0.1578, 0.1852, 0.2591, 0.2087, 0.1089, 0.2157], + device='cuda:3'), in_proj_covar=tensor([0.0194, 0.0212, 0.0167, 0.0202, 0.0201, 0.0183, 0.0156, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 04:31:44,901 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5792, 1.4445, 0.5087, 1.2473, 1.3781, 1.4504, 1.3099, 1.3410], + device='cuda:3'), covar=tensor([0.0533, 0.0398, 0.0426, 0.0590, 0.0308, 0.0537, 0.0521, 0.0614], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 04:32:06,131 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:32:16,780 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72326.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:32:26,338 INFO [finetune.py:976] (3/7) Epoch 13, batch 3600, loss[loss=0.1472, simple_loss=0.2139, pruned_loss=0.04023, over 4797.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2486, pruned_loss=0.0572, over 957030.29 frames. ], batch size: 29, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:32:33,139 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.673e+02 2.029e+02 2.570e+02 4.003e+02, threshold=4.058e+02, percent-clipped=0.0 +2023-04-27 04:32:54,505 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.9764, 3.8844, 2.9149, 4.5016, 3.9490, 3.9891, 1.9133, 3.8885], + device='cuda:3'), covar=tensor([0.1851, 0.1057, 0.2659, 0.1546, 0.2805, 0.1763, 0.5288, 0.2403], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0214, 0.0248, 0.0302, 0.0296, 0.0248, 0.0270, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 04:32:59,901 INFO [finetune.py:976] (3/7) Epoch 13, batch 3650, loss[loss=0.17, simple_loss=0.2487, pruned_loss=0.04564, over 4941.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2508, pruned_loss=0.05788, over 954956.62 frames. ], batch size: 33, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:33:00,713 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9082, 1.4925, 2.0568, 2.3789, 1.9943, 1.9052, 1.9575, 1.9555], + device='cuda:3'), covar=tensor([0.5063, 0.7188, 0.6782, 0.6305, 0.6580, 0.8536, 0.8611, 0.8491], + device='cuda:3'), in_proj_covar=tensor([0.0410, 0.0406, 0.0494, 0.0511, 0.0440, 0.0461, 0.0468, 0.0471], + device='cuda:3'), out_proj_covar=tensor([9.9444e-05, 1.0053e-04, 1.1113e-04, 1.2128e-04, 1.0632e-04, 1.1116e-04, + 1.1190e-04, 1.1271e-04], device='cuda:3') +2023-04-27 04:33:02,552 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72387.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:33:13,808 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3224, 1.5674, 1.6769, 1.8301, 1.6587, 1.7797, 1.8678, 1.7406], + device='cuda:3'), covar=tensor([0.3782, 0.5418, 0.4758, 0.4044, 0.5471, 0.7721, 0.5118, 0.4887], + device='cuda:3'), in_proj_covar=tensor([0.0330, 0.0377, 0.0319, 0.0330, 0.0343, 0.0399, 0.0357, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 04:33:33,724 INFO [finetune.py:976] (3/7) Epoch 13, batch 3700, loss[loss=0.1667, simple_loss=0.2421, pruned_loss=0.04566, over 4737.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2548, pruned_loss=0.05858, over 955152.59 frames. ], batch size: 54, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:33:40,456 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.574e+02 1.888e+02 2.267e+02 6.702e+02, threshold=3.776e+02, percent-clipped=3.0 +2023-04-27 04:33:47,291 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7269, 1.6761, 2.2170, 2.2774, 1.6220, 1.4245, 1.8111, 0.9646], + device='cuda:3'), covar=tensor([0.0728, 0.0959, 0.0469, 0.0792, 0.0841, 0.1288, 0.0851, 0.0901], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0070, 0.0070, 0.0068, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:33:56,412 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 04:34:07,019 INFO [finetune.py:976] (3/7) Epoch 13, batch 3750, loss[loss=0.1601, simple_loss=0.2365, pruned_loss=0.0419, over 4845.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2546, pruned_loss=0.05816, over 953945.48 frames. ], batch size: 44, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:34:08,313 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72485.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:34:39,244 INFO [finetune.py:976] (3/7) Epoch 13, batch 3800, loss[loss=0.1615, simple_loss=0.2297, pruned_loss=0.04665, over 4723.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2556, pruned_loss=0.05863, over 953464.67 frames. ], batch size: 23, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:34:46,446 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.671e+02 1.907e+02 2.254e+02 4.045e+02, threshold=3.813e+02, percent-clipped=1.0 +2023-04-27 04:35:04,215 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 04:35:05,802 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72575.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:35:11,559 INFO [finetune.py:976] (3/7) Epoch 13, batch 3850, loss[loss=0.1623, simple_loss=0.2265, pruned_loss=0.04912, over 4919.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2532, pruned_loss=0.0573, over 953465.95 frames. ], batch size: 37, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:35:33,469 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5261, 3.2709, 1.0864, 1.8972, 2.5452, 1.6640, 4.4673, 2.2983], + device='cuda:3'), covar=tensor([0.0651, 0.0583, 0.0850, 0.1314, 0.0520, 0.1011, 0.0213, 0.0619], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 04:35:37,013 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72623.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:35:42,929 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3525, 1.7589, 2.2609, 2.7075, 2.1840, 1.7168, 1.4522, 2.0087], + device='cuda:3'), covar=tensor([0.3304, 0.3525, 0.1642, 0.2595, 0.2798, 0.2819, 0.4455, 0.2243], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0248, 0.0223, 0.0317, 0.0215, 0.0230, 0.0231, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 04:35:44,026 INFO [finetune.py:976] (3/7) Epoch 13, batch 3900, loss[loss=0.124, simple_loss=0.1981, pruned_loss=0.0249, over 4758.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.251, pruned_loss=0.05658, over 954551.04 frames. ], batch size: 28, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:35:45,162 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7803, 1.9797, 0.9734, 1.4746, 2.1097, 1.6367, 1.5070, 1.6154], + device='cuda:3'), covar=tensor([0.0509, 0.0352, 0.0337, 0.0547, 0.0236, 0.0510, 0.0502, 0.0581], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 04:35:51,957 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.669e+02 1.861e+02 2.404e+02 3.488e+02, threshold=3.722e+02, percent-clipped=0.0 +2023-04-27 04:36:32,485 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72682.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:36:33,065 INFO [finetune.py:976] (3/7) Epoch 13, batch 3950, loss[loss=0.1685, simple_loss=0.2429, pruned_loss=0.04706, over 4869.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2475, pruned_loss=0.05495, over 952925.79 frames. ], batch size: 31, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:36:41,131 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8829, 1.0921, 1.5470, 1.7177, 1.6429, 1.7917, 1.6109, 1.5750], + device='cuda:3'), covar=tensor([0.4264, 0.5955, 0.4938, 0.4770, 0.5964, 0.7565, 0.5417, 0.5585], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0379, 0.0319, 0.0331, 0.0345, 0.0399, 0.0359, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 04:36:51,192 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 04:37:36,357 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8357, 4.0120, 0.8075, 2.2369, 2.1717, 2.9027, 2.3672, 1.0428], + device='cuda:3'), covar=tensor([0.1284, 0.1007, 0.2230, 0.1183, 0.1131, 0.0913, 0.1358, 0.2112], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0245, 0.0138, 0.0121, 0.0131, 0.0151, 0.0116, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:37:38,716 INFO [finetune.py:976] (3/7) Epoch 13, batch 4000, loss[loss=0.1843, simple_loss=0.2596, pruned_loss=0.0545, over 4766.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2482, pruned_loss=0.05566, over 952308.07 frames. ], batch size: 26, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:37:57,171 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 7.525e+01 1.494e+02 1.807e+02 2.123e+02 3.249e+02, threshold=3.613e+02, percent-clipped=0.0 +2023-04-27 04:38:27,849 INFO [finetune.py:976] (3/7) Epoch 13, batch 4050, loss[loss=0.2085, simple_loss=0.2869, pruned_loss=0.06502, over 4858.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2511, pruned_loss=0.05724, over 950557.32 frames. ], batch size: 44, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:38:28,581 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9915, 2.7761, 2.1926, 2.4420, 1.8728, 2.2676, 2.5229, 1.6320], + device='cuda:3'), covar=tensor([0.2694, 0.1390, 0.1078, 0.1550, 0.3563, 0.1360, 0.2071, 0.3327], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0312, 0.0226, 0.0285, 0.0313, 0.0266, 0.0256, 0.0272], + device='cuda:3'), out_proj_covar=tensor([1.1808e-04, 1.2441e-04, 9.0350e-05, 1.1367e-04, 1.2760e-04, 1.0643e-04, + 1.0356e-04, 1.0846e-04], device='cuda:3') +2023-04-27 04:38:29,164 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72785.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:39:01,350 INFO [finetune.py:976] (3/7) Epoch 13, batch 4100, loss[loss=0.1906, simple_loss=0.269, pruned_loss=0.05609, over 4930.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2544, pruned_loss=0.05823, over 951091.56 frames. ], batch size: 42, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:39:01,411 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72833.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:39:09,032 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.644e+02 1.849e+02 2.325e+02 6.848e+02, threshold=3.698e+02, percent-clipped=3.0 +2023-04-27 04:39:34,765 INFO [finetune.py:976] (3/7) Epoch 13, batch 4150, loss[loss=0.1568, simple_loss=0.2187, pruned_loss=0.0474, over 4462.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2558, pruned_loss=0.05926, over 952439.50 frames. ], batch size: 19, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:39:42,047 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72894.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:39:51,067 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4565, 3.8170, 0.8631, 2.0162, 1.9926, 2.5731, 2.2540, 0.8967], + device='cuda:3'), covar=tensor([0.1441, 0.1008, 0.2250, 0.1320, 0.1130, 0.1082, 0.1457, 0.2283], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0247, 0.0139, 0.0122, 0.0133, 0.0153, 0.0117, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 04:39:53,351 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2755, 1.5593, 1.4954, 1.7073, 1.6680, 1.8523, 1.4911, 3.2889], + device='cuda:3'), covar=tensor([0.0598, 0.0740, 0.0736, 0.1084, 0.0572, 0.0649, 0.0696, 0.0149], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 04:40:08,511 INFO [finetune.py:976] (3/7) Epoch 13, batch 4200, loss[loss=0.1607, simple_loss=0.2265, pruned_loss=0.04751, over 4782.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2551, pruned_loss=0.05856, over 950528.98 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:40:15,131 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.607e+02 1.987e+02 2.411e+02 4.371e+02, threshold=3.975e+02, percent-clipped=4.0 +2023-04-27 04:40:24,267 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72955.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:40:30,234 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72964.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:40:41,141 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72982.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:40:41,680 INFO [finetune.py:976] (3/7) Epoch 13, batch 4250, loss[loss=0.1578, simple_loss=0.2322, pruned_loss=0.04164, over 4903.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.251, pruned_loss=0.0567, over 950276.46 frames. ], batch size: 36, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:40:45,461 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3986, 1.6561, 2.0762, 2.6668, 2.8716, 2.1625, 1.6331, 2.2811], + device='cuda:3'), covar=tensor([0.1101, 0.1793, 0.0933, 0.0793, 0.0670, 0.1230, 0.1218, 0.0826], + device='cuda:3'), in_proj_covar=tensor([0.0190, 0.0202, 0.0182, 0.0173, 0.0177, 0.0183, 0.0155, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:40:46,687 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8289, 2.5081, 2.0409, 2.3533, 1.6761, 1.9187, 2.1186, 1.5951], + device='cuda:3'), covar=tensor([0.1891, 0.0862, 0.0781, 0.0987, 0.2969, 0.1041, 0.1427, 0.2396], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0309, 0.0226, 0.0283, 0.0312, 0.0265, 0.0255, 0.0270], + device='cuda:3'), out_proj_covar=tensor([1.1755e-04, 1.2324e-04, 9.0119e-05, 1.1307e-04, 1.2732e-04, 1.0592e-04, + 1.0309e-04, 1.0799e-04], device='cuda:3') +2023-04-27 04:41:10,641 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73025.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 04:41:13,600 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73030.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:41:15,357 INFO [finetune.py:976] (3/7) Epoch 13, batch 4300, loss[loss=0.1829, simple_loss=0.256, pruned_loss=0.05486, over 4776.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2489, pruned_loss=0.05627, over 950391.13 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:41:22,016 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.560e+02 1.991e+02 2.397e+02 3.754e+02, threshold=3.982e+02, percent-clipped=0.0 +2023-04-27 04:41:59,415 INFO [finetune.py:976] (3/7) Epoch 13, batch 4350, loss[loss=0.1878, simple_loss=0.2365, pruned_loss=0.06956, over 4303.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2467, pruned_loss=0.05634, over 949349.37 frames. ], batch size: 18, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:42:01,414 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-04-27 04:42:34,350 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4973, 0.6680, 1.3863, 1.8924, 1.6003, 1.4343, 1.4347, 1.4445], + device='cuda:3'), covar=tensor([0.4503, 0.6387, 0.6247, 0.6644, 0.5788, 0.7242, 0.6986, 0.8289], + device='cuda:3'), in_proj_covar=tensor([0.0412, 0.0408, 0.0495, 0.0513, 0.0442, 0.0464, 0.0469, 0.0475], + device='cuda:3'), out_proj_covar=tensor([9.9778e-05, 1.0099e-04, 1.1153e-04, 1.2171e-04, 1.0677e-04, 1.1166e-04, + 1.1206e-04, 1.1339e-04], device='cuda:3') +2023-04-27 04:42:38,502 INFO [finetune.py:976] (3/7) Epoch 13, batch 4400, loss[loss=0.1837, simple_loss=0.2623, pruned_loss=0.0526, over 4808.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2488, pruned_loss=0.05749, over 949247.03 frames. ], batch size: 29, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:42:47,738 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.646e+02 2.036e+02 2.477e+02 5.410e+02, threshold=4.072e+02, percent-clipped=5.0 +2023-04-27 04:43:39,918 INFO [finetune.py:976] (3/7) Epoch 13, batch 4450, loss[loss=0.2112, simple_loss=0.279, pruned_loss=0.0717, over 4755.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2524, pruned_loss=0.05761, over 950692.93 frames. ], batch size: 54, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:44:17,776 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8106, 1.8232, 1.8128, 1.5697, 2.0646, 1.7363, 2.6479, 1.5499], + device='cuda:3'), covar=tensor([0.3890, 0.1674, 0.4277, 0.2827, 0.1522, 0.2196, 0.1289, 0.4169], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0346, 0.0426, 0.0355, 0.0381, 0.0381, 0.0372, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:44:30,376 INFO [finetune.py:976] (3/7) Epoch 13, batch 4500, loss[loss=0.1957, simple_loss=0.2525, pruned_loss=0.06943, over 4146.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2545, pruned_loss=0.0585, over 952362.17 frames. ], batch size: 66, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:44:37,122 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.883e+01 1.575e+02 1.982e+02 2.231e+02 3.715e+02, threshold=3.964e+02, percent-clipped=0.0 +2023-04-27 04:44:37,815 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1344, 1.3431, 3.8015, 3.4994, 3.3363, 3.6267, 3.6475, 3.3348], + device='cuda:3'), covar=tensor([0.8098, 0.5927, 0.1443, 0.2353, 0.1359, 0.1750, 0.1630, 0.1697], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0305, 0.0401, 0.0405, 0.0344, 0.0403, 0.0311, 0.0363], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:44:40,828 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73250.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:45:04,243 INFO [finetune.py:976] (3/7) Epoch 13, batch 4550, loss[loss=0.1821, simple_loss=0.2518, pruned_loss=0.05622, over 4787.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2566, pruned_loss=0.0588, over 952663.62 frames. ], batch size: 29, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:45:10,366 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:45:16,176 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 04:45:28,148 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73320.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 04:45:37,743 INFO [finetune.py:976] (3/7) Epoch 13, batch 4600, loss[loss=0.1873, simple_loss=0.2482, pruned_loss=0.06319, over 4801.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2555, pruned_loss=0.05799, over 953038.33 frames. ], batch size: 51, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:45:44,461 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.205e+02 1.640e+02 1.996e+02 2.306e+02 3.998e+02, threshold=3.993e+02, percent-clipped=1.0 +2023-04-27 04:45:50,692 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73354.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:46:10,641 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7752, 1.9670, 1.0002, 1.4710, 1.9598, 1.5904, 1.4829, 1.6315], + device='cuda:3'), covar=tensor([0.0532, 0.0374, 0.0360, 0.0583, 0.0282, 0.0537, 0.0561, 0.0610], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0037, 0.0051, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 04:46:11,140 INFO [finetune.py:976] (3/7) Epoch 13, batch 4650, loss[loss=0.2087, simple_loss=0.2651, pruned_loss=0.07618, over 4731.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2533, pruned_loss=0.05779, over 952970.35 frames. ], batch size: 23, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:46:22,774 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0327, 2.4412, 1.0312, 1.3637, 1.9314, 1.1638, 3.0266, 1.6894], + device='cuda:3'), covar=tensor([0.0653, 0.0610, 0.0742, 0.1202, 0.0486, 0.1012, 0.0305, 0.0642], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 04:46:44,504 INFO [finetune.py:976] (3/7) Epoch 13, batch 4700, loss[loss=0.1748, simple_loss=0.2411, pruned_loss=0.05424, over 4897.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2505, pruned_loss=0.057, over 952885.76 frames. ], batch size: 35, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:46:51,528 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.168e+02 1.646e+02 2.000e+02 2.343e+02 4.660e+02, threshold=4.000e+02, percent-clipped=2.0 +2023-04-27 04:47:33,204 INFO [finetune.py:976] (3/7) Epoch 13, batch 4750, loss[loss=0.1251, simple_loss=0.1993, pruned_loss=0.02543, over 4757.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2478, pruned_loss=0.05571, over 952732.00 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:48:03,538 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 04:48:11,969 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6224, 1.2161, 1.7631, 2.0791, 1.7355, 1.6287, 1.7228, 1.7055], + device='cuda:3'), covar=tensor([0.4786, 0.6776, 0.6186, 0.6473, 0.5956, 0.8316, 0.7898, 0.8331], + device='cuda:3'), in_proj_covar=tensor([0.0410, 0.0407, 0.0493, 0.0511, 0.0441, 0.0463, 0.0469, 0.0474], + device='cuda:3'), out_proj_covar=tensor([9.9432e-05, 1.0077e-04, 1.1116e-04, 1.2142e-04, 1.0658e-04, 1.1146e-04, + 1.1212e-04, 1.1303e-04], device='cuda:3') +2023-04-27 04:48:28,609 INFO [finetune.py:976] (3/7) Epoch 13, batch 4800, loss[loss=0.1955, simple_loss=0.2517, pruned_loss=0.06961, over 4721.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2505, pruned_loss=0.05698, over 952710.25 frames. ], batch size: 23, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:48:36,792 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.600e+02 1.891e+02 2.207e+02 3.425e+02, threshold=3.783e+02, percent-clipped=0.0 +2023-04-27 04:48:41,060 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73550.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:49:02,206 INFO [finetune.py:976] (3/7) Epoch 13, batch 4850, loss[loss=0.1635, simple_loss=0.2198, pruned_loss=0.0536, over 4750.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2537, pruned_loss=0.05794, over 953521.49 frames. ], batch size: 27, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:49:17,650 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73590.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:49:28,460 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73598.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:49:30,741 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-27 04:49:53,456 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73620.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 04:49:55,966 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2829, 1.3319, 3.6818, 3.4144, 3.2644, 3.4631, 3.4067, 3.2640], + device='cuda:3'), covar=tensor([0.7201, 0.5624, 0.1188, 0.1705, 0.1243, 0.1889, 0.2608, 0.1507], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0302, 0.0398, 0.0402, 0.0341, 0.0400, 0.0310, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:50:13,628 INFO [finetune.py:976] (3/7) Epoch 13, batch 4900, loss[loss=0.1859, simple_loss=0.255, pruned_loss=0.05843, over 4193.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2533, pruned_loss=0.0573, over 952645.75 frames. ], batch size: 65, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:50:28,451 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.796e+02 2.177e+02 2.620e+02 4.604e+02, threshold=4.354e+02, percent-clipped=4.0 +2023-04-27 04:50:38,421 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73649.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:50:39,765 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73651.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:50:53,260 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3092, 1.6786, 1.5714, 1.9364, 1.7391, 2.0696, 1.4754, 3.7444], + device='cuda:3'), covar=tensor([0.0609, 0.0762, 0.0769, 0.1160, 0.0640, 0.0469, 0.0756, 0.0137], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 04:50:55,036 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73668.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 04:51:04,150 INFO [finetune.py:976] (3/7) Epoch 13, batch 4950, loss[loss=0.1417, simple_loss=0.2195, pruned_loss=0.03198, over 4781.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2542, pruned_loss=0.05757, over 950212.88 frames. ], batch size: 29, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:51:17,150 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1455, 1.7569, 2.1539, 2.5189, 2.5144, 2.0398, 1.6390, 2.2879], + device='cuda:3'), covar=tensor([0.0816, 0.1166, 0.0599, 0.0602, 0.0592, 0.0804, 0.0841, 0.0508], + device='cuda:3'), in_proj_covar=tensor([0.0190, 0.0202, 0.0182, 0.0173, 0.0178, 0.0183, 0.0156, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:51:36,578 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73732.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:51:37,095 INFO [finetune.py:976] (3/7) Epoch 13, batch 5000, loss[loss=0.1489, simple_loss=0.2209, pruned_loss=0.03844, over 4775.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.254, pruned_loss=0.05797, over 950718.37 frames. ], batch size: 51, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:51:45,203 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.282e+02 1.739e+02 2.081e+02 2.360e+02 4.914e+02, threshold=4.162e+02, percent-clipped=1.0 +2023-04-27 04:51:48,796 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2460, 1.6760, 2.1189, 2.5965, 2.0631, 1.6743, 1.4339, 1.9117], + device='cuda:3'), covar=tensor([0.3644, 0.3817, 0.1818, 0.2576, 0.3047, 0.2947, 0.4613, 0.2319], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0248, 0.0222, 0.0316, 0.0215, 0.0229, 0.0230, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 04:52:11,152 INFO [finetune.py:976] (3/7) Epoch 13, batch 5050, loss[loss=0.1792, simple_loss=0.2395, pruned_loss=0.05948, over 4821.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2512, pruned_loss=0.05765, over 950736.35 frames. ], batch size: 41, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:52:17,801 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73793.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:52:25,682 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9874, 2.6428, 2.1526, 2.4512, 1.8290, 2.1247, 2.1425, 1.7131], + device='cuda:3'), covar=tensor([0.1823, 0.0901, 0.0810, 0.1019, 0.3096, 0.1135, 0.1737, 0.2315], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0309, 0.0225, 0.0284, 0.0313, 0.0266, 0.0255, 0.0271], + device='cuda:3'), out_proj_covar=tensor([1.1725e-04, 1.2301e-04, 8.9805e-05, 1.1338e-04, 1.2753e-04, 1.0636e-04, + 1.0335e-04, 1.0833e-04], device='cuda:3') +2023-04-27 04:52:28,199 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 04:52:56,560 INFO [finetune.py:976] (3/7) Epoch 13, batch 5100, loss[loss=0.1922, simple_loss=0.2422, pruned_loss=0.07116, over 4774.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2489, pruned_loss=0.05726, over 951961.22 frames. ], batch size: 51, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:53:01,810 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-04-27 04:53:03,638 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.578e+02 1.847e+02 2.242e+02 3.779e+02, threshold=3.694e+02, percent-clipped=0.0 +2023-04-27 04:53:44,069 INFO [finetune.py:976] (3/7) Epoch 13, batch 5150, loss[loss=0.2406, simple_loss=0.2972, pruned_loss=0.09202, over 4893.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2494, pruned_loss=0.0578, over 954574.38 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:54:18,054 INFO [finetune.py:976] (3/7) Epoch 13, batch 5200, loss[loss=0.2033, simple_loss=0.2598, pruned_loss=0.07345, over 4788.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2521, pruned_loss=0.05852, over 954474.76 frames. ], batch size: 25, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:54:24,745 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.235e+02 1.597e+02 1.940e+02 2.333e+02 4.411e+02, threshold=3.880e+02, percent-clipped=2.0 +2023-04-27 04:54:26,023 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73946.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:54:28,378 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73949.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:55:07,560 INFO [finetune.py:976] (3/7) Epoch 13, batch 5250, loss[loss=0.2258, simple_loss=0.2868, pruned_loss=0.08241, over 4809.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2549, pruned_loss=0.05889, over 955288.28 frames. ], batch size: 40, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:55:16,181 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73997.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:55:42,307 INFO [finetune.py:976] (3/7) Epoch 13, batch 5300, loss[loss=0.2315, simple_loss=0.292, pruned_loss=0.08552, over 4885.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2558, pruned_loss=0.05862, over 953313.81 frames. ], batch size: 32, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:55:54,448 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.549e+02 1.803e+02 2.231e+02 3.853e+02, threshold=3.607e+02, percent-clipped=0.0 +2023-04-27 04:56:19,419 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74064.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:56:36,839 INFO [finetune.py:976] (3/7) Epoch 13, batch 5350, loss[loss=0.1954, simple_loss=0.2611, pruned_loss=0.06487, over 4870.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.255, pruned_loss=0.05792, over 952679.52 frames. ], batch size: 34, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:56:39,871 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:57:01,236 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 04:57:05,261 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74125.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:57:10,417 INFO [finetune.py:976] (3/7) Epoch 13, batch 5400, loss[loss=0.1767, simple_loss=0.248, pruned_loss=0.05274, over 4905.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2526, pruned_loss=0.05763, over 955407.28 frames. ], batch size: 46, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:57:17,120 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.609e+02 1.962e+02 2.381e+02 6.341e+02, threshold=3.923e+02, percent-clipped=2.0 +2023-04-27 04:57:43,731 INFO [finetune.py:976] (3/7) Epoch 13, batch 5450, loss[loss=0.1664, simple_loss=0.2254, pruned_loss=0.05366, over 4781.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2492, pruned_loss=0.05646, over 953330.53 frames. ], batch size: 26, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:57:47,189 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-27 04:58:22,887 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5719, 2.1410, 1.9937, 2.1705, 2.1258, 2.2315, 2.1311, 2.0284], + device='cuda:3'), covar=tensor([0.2951, 0.4314, 0.4221, 0.4377, 0.4485, 0.5492, 0.5123, 0.4949], + device='cuda:3'), in_proj_covar=tensor([0.0331, 0.0375, 0.0317, 0.0330, 0.0341, 0.0398, 0.0357, 0.0325], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 04:58:35,032 INFO [finetune.py:976] (3/7) Epoch 13, batch 5500, loss[loss=0.1656, simple_loss=0.2362, pruned_loss=0.04747, over 4828.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2455, pruned_loss=0.05495, over 950814.98 frames. ], batch size: 30, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:58:39,454 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74240.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:58:42,297 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.578e+02 1.994e+02 2.336e+02 4.147e+02, threshold=3.989e+02, percent-clipped=1.0 +2023-04-27 04:58:43,633 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74246.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:58:56,984 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7299, 1.5960, 1.7841, 2.0906, 2.0819, 1.6203, 1.3247, 1.8821], + device='cuda:3'), covar=tensor([0.0687, 0.1144, 0.0691, 0.0511, 0.0557, 0.0802, 0.0802, 0.0513], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0205, 0.0185, 0.0175, 0.0180, 0.0185, 0.0157, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 04:59:08,625 INFO [finetune.py:976] (3/7) Epoch 13, batch 5550, loss[loss=0.199, simple_loss=0.2818, pruned_loss=0.05806, over 4813.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2477, pruned_loss=0.05579, over 952355.41 frames. ], batch size: 40, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:59:15,400 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:59:20,241 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74301.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:59:39,645 INFO [finetune.py:976] (3/7) Epoch 13, batch 5600, loss[loss=0.193, simple_loss=0.2732, pruned_loss=0.05644, over 4917.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2516, pruned_loss=0.05679, over 952620.77 frames. ], batch size: 42, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:59:46,068 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.613e+02 1.797e+02 2.128e+02 3.644e+02, threshold=3.594e+02, percent-clipped=1.0 +2023-04-27 05:00:16,160 INFO [finetune.py:976] (3/7) Epoch 13, batch 5650, loss[loss=0.231, simple_loss=0.2893, pruned_loss=0.08634, over 4836.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2541, pruned_loss=0.05694, over 952379.24 frames. ], batch size: 30, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 05:00:24,520 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 05:00:25,012 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74388.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:00:44,666 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74420.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:00:52,323 INFO [finetune.py:976] (3/7) Epoch 13, batch 5700, loss[loss=0.1799, simple_loss=0.235, pruned_loss=0.06242, over 4062.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2497, pruned_loss=0.0566, over 934746.14 frames. ], batch size: 17, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 05:00:54,134 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74436.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:00:57,052 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6720, 3.3819, 1.1566, 1.9689, 2.0884, 2.3937, 2.0963, 1.2131], + device='cuda:3'), covar=tensor([0.1268, 0.0999, 0.1915, 0.1177, 0.0952, 0.1037, 0.1473, 0.1777], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0245, 0.0138, 0.0121, 0.0131, 0.0152, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 05:00:58,755 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.564e+02 1.898e+02 2.359e+02 3.680e+02, threshold=3.797e+02, percent-clipped=1.0 +2023-04-27 05:01:06,043 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 05:01:23,647 INFO [finetune.py:976] (3/7) Epoch 14, batch 0, loss[loss=0.1832, simple_loss=0.2374, pruned_loss=0.06454, over 4549.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2374, pruned_loss=0.06454, over 4549.00 frames. ], batch size: 20, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 05:01:23,647 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 05:01:45,229 INFO [finetune.py:1010] (3/7) Epoch 14, validation: loss=0.1535, simple_loss=0.226, pruned_loss=0.04054, over 2265189.00 frames. +2023-04-27 05:01:45,229 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 05:01:57,199 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7156, 2.0988, 1.1472, 1.4721, 2.1271, 1.5854, 1.5429, 1.5899], + device='cuda:3'), covar=tensor([0.0498, 0.0320, 0.0314, 0.0530, 0.0248, 0.0492, 0.0464, 0.0543], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 05:02:18,139 INFO [finetune.py:976] (3/7) Epoch 14, batch 50, loss[loss=0.2339, simple_loss=0.2822, pruned_loss=0.0928, over 4716.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2566, pruned_loss=0.06018, over 216647.91 frames. ], batch size: 59, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:02:41,160 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.646e+01 1.673e+02 1.901e+02 2.437e+02 4.267e+02, threshold=3.802e+02, percent-clipped=2.0 +2023-04-27 05:02:51,798 INFO [finetune.py:976] (3/7) Epoch 14, batch 100, loss[loss=0.1446, simple_loss=0.2174, pruned_loss=0.03586, over 4818.00 frames. ], tot_loss[loss=0.181, simple_loss=0.249, pruned_loss=0.05654, over 379715.00 frames. ], batch size: 51, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:02:55,984 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-27 05:02:56,516 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3617, 2.9431, 1.0511, 1.7077, 1.7975, 2.2370, 1.8070, 1.0069], + device='cuda:3'), covar=tensor([0.1367, 0.1009, 0.1697, 0.1196, 0.1060, 0.0900, 0.1494, 0.1821], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0244, 0.0137, 0.0120, 0.0132, 0.0151, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 05:03:29,436 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74595.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:03:30,040 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74596.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:03:51,942 INFO [finetune.py:976] (3/7) Epoch 14, batch 150, loss[loss=0.1506, simple_loss=0.2243, pruned_loss=0.03841, over 4826.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2441, pruned_loss=0.05458, over 509048.89 frames. ], batch size: 41, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:04:20,050 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.548e+02 1.852e+02 2.253e+02 4.630e+02, threshold=3.704e+02, percent-clipped=3.0 +2023-04-27 05:04:27,993 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74656.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:04:31,408 INFO [finetune.py:976] (3/7) Epoch 14, batch 200, loss[loss=0.1703, simple_loss=0.2404, pruned_loss=0.05013, over 4903.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2456, pruned_loss=0.05566, over 608714.04 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:05:05,314 INFO [finetune.py:976] (3/7) Epoch 14, batch 250, loss[loss=0.1964, simple_loss=0.2704, pruned_loss=0.06121, over 4796.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.251, pruned_loss=0.05763, over 687073.64 frames. ], batch size: 29, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:05:11,833 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:05:39,274 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.695e+01 1.601e+02 1.840e+02 2.318e+02 4.456e+02, threshold=3.680e+02, percent-clipped=1.0 +2023-04-27 05:05:48,368 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8942, 1.1072, 1.5747, 1.6850, 1.6451, 1.7333, 1.5540, 1.5868], + device='cuda:3'), covar=tensor([0.4193, 0.5578, 0.4972, 0.4850, 0.5514, 0.7592, 0.5318, 0.4816], + device='cuda:3'), in_proj_covar=tensor([0.0329, 0.0373, 0.0316, 0.0328, 0.0340, 0.0397, 0.0355, 0.0323], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:05:49,929 INFO [finetune.py:976] (3/7) Epoch 14, batch 300, loss[loss=0.2389, simple_loss=0.3119, pruned_loss=0.0829, over 4902.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2546, pruned_loss=0.05788, over 747999.11 frames. ], batch size: 38, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:05:54,766 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74768.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:06:23,249 INFO [finetune.py:976] (3/7) Epoch 14, batch 350, loss[loss=0.1538, simple_loss=0.2276, pruned_loss=0.04001, over 4868.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2567, pruned_loss=0.05894, over 793836.55 frames. ], batch size: 34, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:07:09,536 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 1.682e+02 1.913e+02 2.290e+02 4.465e+02, threshold=3.826e+02, percent-clipped=1.0 +2023-04-27 05:07:09,761 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 05:07:25,451 INFO [finetune.py:976] (3/7) Epoch 14, batch 400, loss[loss=0.2057, simple_loss=0.2675, pruned_loss=0.07193, over 4866.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2573, pruned_loss=0.05915, over 831033.70 frames. ], batch size: 31, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:07:49,827 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74896.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:07:57,749 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:07:59,307 INFO [finetune.py:976] (3/7) Epoch 14, batch 450, loss[loss=0.167, simple_loss=0.2337, pruned_loss=0.0502, over 4941.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2564, pruned_loss=0.05889, over 857842.21 frames. ], batch size: 39, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:08:37,890 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74944.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:08:38,409 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.611e+01 1.647e+02 1.819e+02 2.107e+02 4.773e+02, threshold=3.638e+02, percent-clipped=2.0 +2023-04-27 05:08:42,176 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74951.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:08:53,659 INFO [finetune.py:976] (3/7) Epoch 14, batch 500, loss[loss=0.1623, simple_loss=0.2342, pruned_loss=0.0452, over 4829.00 frames. ], tot_loss[loss=0.183, simple_loss=0.252, pruned_loss=0.057, over 880392.95 frames. ], batch size: 33, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:08:56,074 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74964.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:08:59,791 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74970.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:09:06,882 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1141, 2.5058, 0.9737, 1.3966, 1.9834, 1.2111, 3.2350, 1.6204], + device='cuda:3'), covar=tensor([0.0624, 0.0780, 0.0841, 0.1128, 0.0455, 0.0931, 0.0225, 0.0610], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 05:09:12,655 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0262, 1.7225, 2.0300, 2.3106, 2.3503, 1.8561, 1.5279, 1.9940], + device='cuda:3'), covar=tensor([0.0713, 0.1119, 0.0564, 0.0578, 0.0515, 0.0723, 0.0762, 0.0540], + device='cuda:3'), in_proj_covar=tensor([0.0191, 0.0204, 0.0184, 0.0174, 0.0179, 0.0184, 0.0156, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:09:27,833 INFO [finetune.py:976] (3/7) Epoch 14, batch 550, loss[loss=0.1639, simple_loss=0.2214, pruned_loss=0.05325, over 4826.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2492, pruned_loss=0.05683, over 895309.88 frames. ], batch size: 51, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:09:37,463 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75025.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:09:51,449 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.675e+02 1.910e+02 2.434e+02 4.302e+02, threshold=3.821e+02, percent-clipped=4.0 +2023-04-27 05:10:01,742 INFO [finetune.py:976] (3/7) Epoch 14, batch 600, loss[loss=0.2189, simple_loss=0.2808, pruned_loss=0.07852, over 4134.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.25, pruned_loss=0.05706, over 909588.26 frames. ], batch size: 65, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:10:02,208 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 05:10:35,744 INFO [finetune.py:976] (3/7) Epoch 14, batch 650, loss[loss=0.1966, simple_loss=0.2703, pruned_loss=0.0615, over 4813.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2539, pruned_loss=0.0581, over 920899.90 frames. ], batch size: 39, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:10:38,934 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5705, 1.4006, 1.7221, 2.0006, 1.4272, 1.0967, 1.4842, 0.9764], + device='cuda:3'), covar=tensor([0.0612, 0.0638, 0.0464, 0.0451, 0.0751, 0.1548, 0.0648, 0.0874], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0070, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0070], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 05:10:59,292 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.654e+02 2.108e+02 2.537e+02 4.925e+02, threshold=4.216e+02, percent-clipped=6.0 +2023-04-27 05:11:05,543 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9981, 1.4735, 1.8037, 2.1752, 1.8029, 1.4207, 0.9790, 1.5378], + device='cuda:3'), covar=tensor([0.3238, 0.3510, 0.1852, 0.2164, 0.2646, 0.2784, 0.4476, 0.2254], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0245, 0.0221, 0.0313, 0.0214, 0.0227, 0.0228, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 05:11:09,460 INFO [finetune.py:976] (3/7) Epoch 14, batch 700, loss[loss=0.1584, simple_loss=0.2218, pruned_loss=0.04751, over 4743.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2551, pruned_loss=0.05805, over 929401.80 frames. ], batch size: 27, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:11:46,819 INFO [finetune.py:976] (3/7) Epoch 14, batch 750, loss[loss=0.1764, simple_loss=0.2462, pruned_loss=0.05333, over 4796.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2568, pruned_loss=0.05874, over 936113.78 frames. ], batch size: 29, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:11:54,584 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5260, 1.0901, 1.2528, 1.1049, 1.6676, 1.3126, 1.0142, 1.2140], + device='cuda:3'), covar=tensor([0.1610, 0.1348, 0.2026, 0.1599, 0.0869, 0.1412, 0.2020, 0.2133], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0318, 0.0354, 0.0294, 0.0332, 0.0316, 0.0308, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.4050e-05, 6.6822e-05, 7.5892e-05, 6.0067e-05, 6.9240e-05, 6.6912e-05, + 6.5163e-05, 7.7689e-05], device='cuda:3') +2023-04-27 05:12:09,250 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75228.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:31,818 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.600e+02 1.912e+02 2.289e+02 3.679e+02, threshold=3.824e+02, percent-clipped=0.0 +2023-04-27 05:12:35,585 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75251.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:41,569 INFO [finetune.py:976] (3/7) Epoch 14, batch 800, loss[loss=0.1641, simple_loss=0.2355, pruned_loss=0.04635, over 4772.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2552, pruned_loss=0.05774, over 940055.24 frames. ], batch size: 28, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:12:44,560 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75265.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:47,629 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75270.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:52,580 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.01 vs. limit=5.0 +2023-04-27 05:13:00,637 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:13:07,390 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 05:13:07,657 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:13:14,864 INFO [finetune.py:976] (3/7) Epoch 14, batch 850, loss[loss=0.1706, simple_loss=0.2458, pruned_loss=0.0477, over 4889.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2524, pruned_loss=0.05678, over 944481.25 frames. ], batch size: 35, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:13:26,895 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:13:40,170 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75331.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 05:13:56,949 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.246e+02 1.618e+02 2.006e+02 2.388e+02 8.621e+02, threshold=4.012e+02, percent-clipped=6.0 +2023-04-27 05:14:01,258 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1546, 2.4236, 0.9495, 1.3588, 1.8754, 1.2699, 3.3222, 1.7840], + device='cuda:3'), covar=tensor([0.0639, 0.0683, 0.0830, 0.1370, 0.0559, 0.1033, 0.0257, 0.0663], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 05:14:12,835 INFO [finetune.py:976] (3/7) Epoch 14, batch 900, loss[loss=0.1934, simple_loss=0.2546, pruned_loss=0.06607, over 4922.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2504, pruned_loss=0.05631, over 947092.75 frames. ], batch size: 38, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:14:37,899 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6578, 1.7344, 1.8417, 1.4526, 1.9114, 1.5790, 2.4244, 1.5309], + device='cuda:3'), covar=tensor([0.3129, 0.1433, 0.3633, 0.2202, 0.1343, 0.1955, 0.1147, 0.4345], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0342, 0.0420, 0.0349, 0.0376, 0.0376, 0.0365, 0.0413], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:14:57,514 INFO [finetune.py:976] (3/7) Epoch 14, batch 950, loss[loss=0.1895, simple_loss=0.2566, pruned_loss=0.06122, over 4830.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2488, pruned_loss=0.05619, over 950108.45 frames. ], batch size: 40, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:15:18,966 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7850, 1.4062, 1.9578, 2.2737, 1.9114, 1.7903, 1.8890, 1.8281], + device='cuda:3'), covar=tensor([0.5219, 0.6925, 0.6990, 0.6491, 0.6462, 0.8533, 0.8411, 0.8054], + device='cuda:3'), in_proj_covar=tensor([0.0413, 0.0407, 0.0494, 0.0512, 0.0443, 0.0463, 0.0470, 0.0474], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:15:20,049 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.489e+02 1.870e+02 2.238e+02 3.708e+02, threshold=3.739e+02, percent-clipped=0.0 +2023-04-27 05:15:30,323 INFO [finetune.py:976] (3/7) Epoch 14, batch 1000, loss[loss=0.1723, simple_loss=0.234, pruned_loss=0.05536, over 4804.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2509, pruned_loss=0.05736, over 950214.40 frames. ], batch size: 25, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:15:50,336 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 05:16:03,286 INFO [finetune.py:976] (3/7) Epoch 14, batch 1050, loss[loss=0.1287, simple_loss=0.2111, pruned_loss=0.02314, over 4753.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2519, pruned_loss=0.05668, over 950131.98 frames. ], batch size: 27, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:16:25,346 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.730e+02 1.979e+02 2.287e+02 8.214e+02, threshold=3.958e+02, percent-clipped=1.0 +2023-04-27 05:16:26,108 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8616, 1.8013, 1.7193, 1.5328, 2.0026, 1.6153, 2.5457, 1.4806], + device='cuda:3'), covar=tensor([0.3953, 0.1795, 0.4411, 0.3016, 0.1840, 0.2542, 0.1544, 0.4761], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0345, 0.0421, 0.0352, 0.0377, 0.0377, 0.0367, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:16:36,964 INFO [finetune.py:976] (3/7) Epoch 14, batch 1100, loss[loss=0.1974, simple_loss=0.2353, pruned_loss=0.07975, over 4417.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2538, pruned_loss=0.05762, over 951333.57 frames. ], batch size: 19, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:16:39,560 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75565.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:16:57,395 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75584.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:17:31,383 INFO [finetune.py:976] (3/7) Epoch 14, batch 1150, loss[loss=0.1569, simple_loss=0.2214, pruned_loss=0.04616, over 4706.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2555, pruned_loss=0.05786, over 953431.10 frames. ], batch size: 23, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:17:32,671 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:17:36,953 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75620.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:17:40,631 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75626.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 05:17:53,272 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.637e+02 1.916e+02 2.260e+02 4.790e+02, threshold=3.833e+02, percent-clipped=2.0 +2023-04-27 05:18:05,378 INFO [finetune.py:976] (3/7) Epoch 14, batch 1200, loss[loss=0.1611, simple_loss=0.2302, pruned_loss=0.04596, over 4676.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2541, pruned_loss=0.05777, over 951170.28 frames. ], batch size: 23, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:18:09,670 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75668.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:18:43,840 INFO [finetune.py:976] (3/7) Epoch 14, batch 1250, loss[loss=0.1979, simple_loss=0.254, pruned_loss=0.07092, over 4802.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2512, pruned_loss=0.05648, over 953994.29 frames. ], batch size: 51, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:18:52,927 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1273, 1.3385, 1.2045, 1.3054, 1.1753, 1.0920, 1.2119, 0.9904], + device='cuda:3'), covar=tensor([0.1681, 0.1518, 0.1146, 0.1278, 0.3076, 0.1375, 0.1648, 0.1963], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0313, 0.0228, 0.0285, 0.0314, 0.0269, 0.0256, 0.0273], + device='cuda:3'), out_proj_covar=tensor([1.1774e-04, 1.2475e-04, 9.0931e-05, 1.1369e-04, 1.2807e-04, 1.0761e-04, + 1.0381e-04, 1.0915e-04], device='cuda:3') +2023-04-27 05:19:22,184 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 05:19:24,353 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.269e+02 1.670e+02 2.009e+02 2.395e+02 6.270e+02, threshold=4.018e+02, percent-clipped=3.0 +2023-04-27 05:19:40,590 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0863, 2.1800, 1.7964, 1.7999, 2.1288, 1.6401, 2.6535, 1.3311], + device='cuda:3'), covar=tensor([0.3763, 0.1835, 0.4621, 0.3014, 0.1929, 0.2787, 0.1300, 0.5225], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0344, 0.0422, 0.0351, 0.0378, 0.0378, 0.0367, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:19:45,790 INFO [finetune.py:976] (3/7) Epoch 14, batch 1300, loss[loss=0.1672, simple_loss=0.2437, pruned_loss=0.04541, over 4828.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.249, pruned_loss=0.056, over 955894.10 frames. ], batch size: 39, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:20:02,657 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5256, 1.8245, 1.9376, 2.0791, 1.8898, 2.0063, 2.0591, 1.9540], + device='cuda:3'), covar=tensor([0.4623, 0.6739, 0.4943, 0.5175, 0.6321, 0.8148, 0.5978, 0.5886], + device='cuda:3'), in_proj_covar=tensor([0.0330, 0.0374, 0.0317, 0.0329, 0.0342, 0.0398, 0.0355, 0.0325], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:20:50,514 INFO [finetune.py:976] (3/7) Epoch 14, batch 1350, loss[loss=0.1707, simple_loss=0.2493, pruned_loss=0.04604, over 4794.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2488, pruned_loss=0.05586, over 953395.54 frames. ], batch size: 29, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:21:18,748 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6184, 1.5350, 0.7510, 1.2890, 1.5875, 1.4696, 1.3515, 1.4911], + device='cuda:3'), covar=tensor([0.0516, 0.0376, 0.0375, 0.0562, 0.0275, 0.0504, 0.0494, 0.0573], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 05:21:19,977 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1918, 1.4518, 1.3611, 1.8357, 1.6173, 1.7242, 1.3186, 4.0147], + device='cuda:3'), covar=tensor([0.0788, 0.1138, 0.1101, 0.1449, 0.0896, 0.0908, 0.1048, 0.0198], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 05:21:39,792 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.563e+02 1.910e+02 2.423e+02 5.618e+02, threshold=3.821e+02, percent-clipped=3.0 +2023-04-27 05:21:40,984 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 05:21:55,185 INFO [finetune.py:976] (3/7) Epoch 14, batch 1400, loss[loss=0.2166, simple_loss=0.2892, pruned_loss=0.07195, over 4818.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2514, pruned_loss=0.05687, over 951543.83 frames. ], batch size: 40, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:22:34,324 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75884.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:22:39,257 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1854, 2.0025, 1.7074, 1.7052, 2.0354, 1.8029, 2.3380, 1.4350], + device='cuda:3'), covar=tensor([0.3330, 0.1429, 0.3829, 0.2550, 0.1578, 0.2013, 0.1482, 0.4504], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0343, 0.0422, 0.0352, 0.0378, 0.0378, 0.0367, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:22:51,836 INFO [finetune.py:976] (3/7) Epoch 14, batch 1450, loss[loss=0.1465, simple_loss=0.2245, pruned_loss=0.03427, over 4835.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2513, pruned_loss=0.05631, over 949414.93 frames. ], batch size: 47, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:23:03,112 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75926.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:23:04,957 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9541, 1.3514, 1.5161, 1.6437, 2.1484, 1.6964, 1.3705, 1.4667], + device='cuda:3'), covar=tensor([0.1404, 0.1643, 0.2333, 0.1456, 0.0854, 0.1742, 0.2310, 0.2120], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0315, 0.0353, 0.0291, 0.0330, 0.0313, 0.0304, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.3756e-05, 6.6103e-05, 7.5658e-05, 5.9381e-05, 6.8669e-05, 6.6318e-05, + 6.4353e-05, 7.7208e-05], device='cuda:3') +2023-04-27 05:23:06,747 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75932.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:23:14,579 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.664e+02 2.009e+02 2.321e+02 4.322e+02, threshold=4.018e+02, percent-clipped=2.0 +2023-04-27 05:23:25,299 INFO [finetune.py:976] (3/7) Epoch 14, batch 1500, loss[loss=0.1651, simple_loss=0.2354, pruned_loss=0.04742, over 4696.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2537, pruned_loss=0.05748, over 951506.77 frames. ], batch size: 23, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:23:34,211 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:24:11,481 INFO [finetune.py:976] (3/7) Epoch 14, batch 1550, loss[loss=0.1513, simple_loss=0.2241, pruned_loss=0.03926, over 4911.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2523, pruned_loss=0.05644, over 952494.42 frames. ], batch size: 37, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:24:40,108 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.630e+02 1.920e+02 2.278e+02 6.745e+02, threshold=3.839e+02, percent-clipped=3.0 +2023-04-27 05:25:01,381 INFO [finetune.py:976] (3/7) Epoch 14, batch 1600, loss[loss=0.1865, simple_loss=0.2538, pruned_loss=0.05963, over 4843.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2511, pruned_loss=0.05637, over 953813.38 frames. ], batch size: 47, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:25:34,939 INFO [finetune.py:976] (3/7) Epoch 14, batch 1650, loss[loss=0.1951, simple_loss=0.2464, pruned_loss=0.0719, over 4868.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2477, pruned_loss=0.05521, over 954502.66 frames. ], batch size: 31, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:25:58,462 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.559e+02 1.837e+02 2.256e+02 6.477e+02, threshold=3.674e+02, percent-clipped=1.0 +2023-04-27 05:26:08,245 INFO [finetune.py:976] (3/7) Epoch 14, batch 1700, loss[loss=0.2069, simple_loss=0.2629, pruned_loss=0.07548, over 4808.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2462, pruned_loss=0.05498, over 956840.46 frames. ], batch size: 40, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:26:42,126 INFO [finetune.py:976] (3/7) Epoch 14, batch 1750, loss[loss=0.2129, simple_loss=0.2807, pruned_loss=0.07257, over 4897.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.248, pruned_loss=0.05564, over 957562.33 frames. ], batch size: 32, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:26:44,572 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1341, 1.4967, 1.4011, 1.6519, 1.5335, 1.9016, 1.3146, 3.4821], + device='cuda:3'), covar=tensor([0.0665, 0.0834, 0.0861, 0.1317, 0.0707, 0.0550, 0.0821, 0.0144], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 05:27:06,454 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.610e+02 1.943e+02 2.415e+02 4.122e+02, threshold=3.886e+02, percent-clipped=3.0 +2023-04-27 05:27:16,254 INFO [finetune.py:976] (3/7) Epoch 14, batch 1800, loss[loss=0.1989, simple_loss=0.279, pruned_loss=0.05937, over 4802.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2513, pruned_loss=0.05656, over 955727.95 frames. ], batch size: 41, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:27:27,558 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76269.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 05:28:02,594 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:28:12,932 INFO [finetune.py:976] (3/7) Epoch 14, batch 1850, loss[loss=0.185, simple_loss=0.2549, pruned_loss=0.05754, over 4891.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2527, pruned_loss=0.05701, over 955052.73 frames. ], batch size: 43, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:28:20,918 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76323.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 05:28:25,686 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76330.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:28:36,249 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.653e+02 2.056e+02 2.488e+02 6.955e+02, threshold=4.112e+02, percent-clipped=5.0 +2023-04-27 05:28:43,357 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76355.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:28:46,907 INFO [finetune.py:976] (3/7) Epoch 14, batch 1900, loss[loss=0.1749, simple_loss=0.2438, pruned_loss=0.05303, over 4790.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2531, pruned_loss=0.05691, over 953973.22 frames. ], batch size: 25, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:29:01,998 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76384.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:29:17,129 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5794, 2.7363, 2.2408, 2.4365, 2.7272, 2.4610, 3.6461, 1.9777], + device='cuda:3'), covar=tensor([0.3725, 0.2159, 0.4075, 0.3085, 0.1876, 0.2614, 0.1367, 0.4109], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0348, 0.0427, 0.0358, 0.0381, 0.0383, 0.0373, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:29:20,619 INFO [finetune.py:976] (3/7) Epoch 14, batch 1950, loss[loss=0.1277, simple_loss=0.2051, pruned_loss=0.0251, over 4798.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2517, pruned_loss=0.05619, over 951374.74 frames. ], batch size: 25, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:29:30,919 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2503, 1.2186, 1.3632, 1.6196, 1.6352, 1.3063, 0.9246, 1.4321], + device='cuda:3'), covar=tensor([0.0873, 0.1329, 0.0838, 0.0567, 0.0665, 0.0739, 0.0852, 0.0597], + device='cuda:3'), in_proj_covar=tensor([0.0190, 0.0203, 0.0182, 0.0173, 0.0178, 0.0182, 0.0154, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:29:42,670 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.125e+01 1.567e+02 1.836e+02 2.067e+02 4.304e+02, threshold=3.671e+02, percent-clipped=1.0 +2023-04-27 05:29:43,330 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-04-27 05:29:55,871 INFO [finetune.py:976] (3/7) Epoch 14, batch 2000, loss[loss=0.1781, simple_loss=0.24, pruned_loss=0.05809, over 4900.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2501, pruned_loss=0.05654, over 952914.30 frames. ], batch size: 32, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:30:58,161 INFO [finetune.py:976] (3/7) Epoch 14, batch 2050, loss[loss=0.207, simple_loss=0.2628, pruned_loss=0.07562, over 4906.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2472, pruned_loss=0.05572, over 953346.24 frames. ], batch size: 37, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:31:09,479 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 05:31:18,745 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0440, 1.5553, 1.4684, 1.7673, 1.7081, 1.7446, 1.4103, 3.0876], + device='cuda:3'), covar=tensor([0.0711, 0.0798, 0.0786, 0.1162, 0.0608, 0.0489, 0.0728, 0.0186], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 05:31:19,797 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.114e+02 1.549e+02 1.874e+02 2.355e+02 5.627e+02, threshold=3.748e+02, percent-clipped=2.0 +2023-04-27 05:31:32,054 INFO [finetune.py:976] (3/7) Epoch 14, batch 2100, loss[loss=0.1466, simple_loss=0.2218, pruned_loss=0.03567, over 4827.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2467, pruned_loss=0.05568, over 953983.72 frames. ], batch size: 25, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:32:06,102 INFO [finetune.py:976] (3/7) Epoch 14, batch 2150, loss[loss=0.2325, simple_loss=0.2991, pruned_loss=0.08298, over 4806.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2514, pruned_loss=0.05707, over 955755.99 frames. ], batch size: 45, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:32:14,727 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76625.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 05:32:27,240 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2092, 1.5694, 1.4509, 1.6924, 1.6638, 1.8663, 1.2996, 3.5662], + device='cuda:3'), covar=tensor([0.0630, 0.0810, 0.0778, 0.1192, 0.0649, 0.0584, 0.0796, 0.0153], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 05:32:27,725 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.703e+02 1.997e+02 2.488e+02 3.640e+02, threshold=3.993e+02, percent-clipped=1.0 +2023-04-27 05:32:30,858 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:32:44,022 INFO [finetune.py:976] (3/7) Epoch 14, batch 2200, loss[loss=0.1851, simple_loss=0.2382, pruned_loss=0.06597, over 4701.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2538, pruned_loss=0.0578, over 954866.61 frames. ], batch size: 23, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:33:06,762 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76679.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:33:38,111 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1156, 2.5215, 0.9114, 1.3290, 1.9090, 1.2818, 3.3767, 1.8110], + device='cuda:3'), covar=tensor([0.0639, 0.0714, 0.0870, 0.1332, 0.0514, 0.0995, 0.0265, 0.0599], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 05:33:46,582 INFO [finetune.py:976] (3/7) Epoch 14, batch 2250, loss[loss=0.1789, simple_loss=0.252, pruned_loss=0.0529, over 4860.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2561, pruned_loss=0.05887, over 955513.44 frames. ], batch size: 31, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:34:30,286 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.594e+02 1.898e+02 2.176e+02 4.652e+02, threshold=3.795e+02, percent-clipped=1.0 +2023-04-27 05:34:47,192 INFO [finetune.py:976] (3/7) Epoch 14, batch 2300, loss[loss=0.1513, simple_loss=0.2159, pruned_loss=0.04331, over 4190.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2551, pruned_loss=0.05782, over 955988.67 frames. ], batch size: 18, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:34:48,384 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0201, 2.5163, 2.2034, 2.5617, 1.7260, 2.1546, 2.0354, 1.7328], + device='cuda:3'), covar=tensor([0.2211, 0.1554, 0.0825, 0.1173, 0.3566, 0.1322, 0.2088, 0.2623], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0314, 0.0228, 0.0287, 0.0316, 0.0269, 0.0258, 0.0274], + device='cuda:3'), out_proj_covar=tensor([1.1859e-04, 1.2492e-04, 9.0849e-05, 1.1450e-04, 1.2887e-04, 1.0744e-04, + 1.0451e-04, 1.0947e-04], device='cuda:3') +2023-04-27 05:35:14,280 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4121, 1.6479, 1.7849, 1.9354, 1.7812, 1.9112, 1.9254, 1.8728], + device='cuda:3'), covar=tensor([0.4259, 0.5975, 0.4775, 0.4441, 0.5819, 0.7573, 0.5599, 0.4968], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0377, 0.0318, 0.0330, 0.0344, 0.0401, 0.0357, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:35:26,138 INFO [finetune.py:976] (3/7) Epoch 14, batch 2350, loss[loss=0.1502, simple_loss=0.2151, pruned_loss=0.04265, over 4927.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2527, pruned_loss=0.05696, over 957917.12 frames. ], batch size: 38, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:35:43,325 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6918, 1.3843, 4.4796, 4.1581, 3.9104, 4.3004, 4.1752, 3.8793], + device='cuda:3'), covar=tensor([0.6600, 0.5848, 0.0957, 0.1768, 0.1060, 0.1587, 0.1176, 0.1567], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0304, 0.0399, 0.0402, 0.0344, 0.0401, 0.0313, 0.0362], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:35:56,534 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3029, 1.5591, 1.3688, 1.5396, 1.3435, 1.2172, 1.3093, 1.0901], + device='cuda:3'), covar=tensor([0.1972, 0.1635, 0.1100, 0.1336, 0.3951, 0.1598, 0.1983, 0.2424], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0313, 0.0227, 0.0286, 0.0315, 0.0268, 0.0257, 0.0274], + device='cuda:3'), out_proj_covar=tensor([1.1807e-04, 1.2485e-04, 9.0665e-05, 1.1425e-04, 1.2860e-04, 1.0725e-04, + 1.0418e-04, 1.0920e-04], device='cuda:3') +2023-04-27 05:36:05,987 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76839.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:36:09,506 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.656e+02 1.888e+02 2.270e+02 4.545e+02, threshold=3.776e+02, percent-clipped=2.0 +2023-04-27 05:36:20,259 INFO [finetune.py:976] (3/7) Epoch 14, batch 2400, loss[loss=0.1486, simple_loss=0.2143, pruned_loss=0.04147, over 4214.00 frames. ], tot_loss[loss=0.181, simple_loss=0.25, pruned_loss=0.05603, over 958093.45 frames. ], batch size: 65, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:36:46,486 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76900.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:36:52,966 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7217, 1.8971, 1.0293, 1.4091, 1.9229, 1.5774, 1.4502, 1.5243], + device='cuda:3'), covar=tensor([0.0525, 0.0369, 0.0344, 0.0547, 0.0264, 0.0520, 0.0511, 0.0594], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 05:36:54,078 INFO [finetune.py:976] (3/7) Epoch 14, batch 2450, loss[loss=0.1724, simple_loss=0.2391, pruned_loss=0.0528, over 4922.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2475, pruned_loss=0.05554, over 956736.15 frames. ], batch size: 37, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:37:04,096 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76925.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:37:06,482 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6314, 1.3920, 1.8220, 2.0512, 1.7851, 1.6124, 1.7195, 1.6793], + device='cuda:3'), covar=tensor([0.4034, 0.5787, 0.5082, 0.5292, 0.4918, 0.6676, 0.6412, 0.7060], + device='cuda:3'), in_proj_covar=tensor([0.0415, 0.0408, 0.0495, 0.0511, 0.0444, 0.0465, 0.0471, 0.0475], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:37:15,275 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8510, 3.7428, 2.7485, 4.4144, 3.7707, 3.8305, 1.8773, 3.8898], + device='cuda:3'), covar=tensor([0.1444, 0.1197, 0.2936, 0.1515, 0.4452, 0.1708, 0.5109, 0.2173], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0212, 0.0247, 0.0300, 0.0295, 0.0244, 0.0269, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:37:17,023 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.514e+02 1.972e+02 2.483e+02 3.800e+02, threshold=3.944e+02, percent-clipped=1.0 +2023-04-27 05:37:20,125 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:37:27,668 INFO [finetune.py:976] (3/7) Epoch 14, batch 2500, loss[loss=0.206, simple_loss=0.2707, pruned_loss=0.07068, over 4931.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2501, pruned_loss=0.05765, over 953788.64 frames. ], batch size: 33, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:37:36,502 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76973.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:37:40,679 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76979.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:37:52,652 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:38:07,211 INFO [finetune.py:976] (3/7) Epoch 14, batch 2550, loss[loss=0.2274, simple_loss=0.2929, pruned_loss=0.08094, over 4904.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2539, pruned_loss=0.0586, over 954574.20 frames. ], batch size: 36, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:38:18,430 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77027.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:38:30,330 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.907e+01 1.587e+02 1.892e+02 2.346e+02 6.910e+02, threshold=3.784e+02, percent-clipped=5.0 +2023-04-27 05:38:34,682 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8459, 1.8792, 1.8427, 1.5032, 1.9412, 1.6205, 2.5598, 1.5988], + device='cuda:3'), covar=tensor([0.3913, 0.1915, 0.4949, 0.2952, 0.1791, 0.2559, 0.1573, 0.4773], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0349, 0.0428, 0.0360, 0.0384, 0.0385, 0.0374, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:38:40,006 INFO [finetune.py:976] (3/7) Epoch 14, batch 2600, loss[loss=0.2071, simple_loss=0.286, pruned_loss=0.06412, over 4812.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.255, pruned_loss=0.05864, over 955751.92 frames. ], batch size: 39, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:39:20,547 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77095.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:39:41,101 INFO [finetune.py:976] (3/7) Epoch 14, batch 2650, loss[loss=0.1477, simple_loss=0.2382, pruned_loss=0.02864, over 4726.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2553, pruned_loss=0.05829, over 953700.25 frames. ], batch size: 54, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:40:19,434 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 1.638e+02 1.948e+02 2.279e+02 4.007e+02, threshold=3.896e+02, percent-clipped=1.0 +2023-04-27 05:40:21,564 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-04-27 05:40:26,269 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77156.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:40:29,225 INFO [finetune.py:976] (3/7) Epoch 14, batch 2700, loss[loss=0.2001, simple_loss=0.2653, pruned_loss=0.06744, over 4800.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2548, pruned_loss=0.058, over 951251.94 frames. ], batch size: 41, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:40:44,680 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-27 05:40:53,254 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77195.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:03,080 INFO [finetune.py:976] (3/7) Epoch 14, batch 2750, loss[loss=0.1665, simple_loss=0.2309, pruned_loss=0.05106, over 4718.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2517, pruned_loss=0.0568, over 953474.68 frames. ], batch size: 23, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:41:03,180 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77211.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:09,344 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1941, 1.4157, 1.6587, 1.8063, 1.6972, 1.8012, 1.7140, 1.7267], + device='cuda:3'), covar=tensor([0.4596, 0.6008, 0.5335, 0.4927, 0.5925, 0.7915, 0.5791, 0.5205], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0377, 0.0317, 0.0330, 0.0343, 0.0400, 0.0357, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:41:13,372 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5965, 1.7733, 1.8809, 2.0406, 1.8822, 1.9351, 1.9973, 1.9737], + device='cuda:3'), covar=tensor([0.4070, 0.6111, 0.4908, 0.4639, 0.5617, 0.7990, 0.5242, 0.5296], + device='cuda:3'), in_proj_covar=tensor([0.0331, 0.0377, 0.0317, 0.0330, 0.0342, 0.0400, 0.0357, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:41:37,789 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.631e+02 1.907e+02 2.391e+02 4.996e+02, threshold=3.813e+02, percent-clipped=2.0 +2023-04-27 05:41:48,207 INFO [finetune.py:976] (3/7) Epoch 14, batch 2800, loss[loss=0.1811, simple_loss=0.2469, pruned_loss=0.05766, over 4830.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2478, pruned_loss=0.05545, over 954367.06 frames. ], batch size: 40, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:41:55,158 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:55,740 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6302, 1.9739, 2.5178, 2.9938, 3.0454, 2.5291, 2.1061, 2.6184], + device='cuda:3'), covar=tensor([0.0778, 0.1180, 0.0641, 0.0542, 0.0490, 0.0751, 0.0725, 0.0560], + device='cuda:3'), in_proj_covar=tensor([0.0191, 0.0202, 0.0182, 0.0172, 0.0178, 0.0181, 0.0154, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:41:55,751 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:58,631 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77277.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:42:11,063 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5013, 1.7055, 1.7635, 1.9098, 1.8158, 1.8987, 1.8817, 1.8177], + device='cuda:3'), covar=tensor([0.4338, 0.6018, 0.5131, 0.4859, 0.5780, 0.7657, 0.6133, 0.5627], + device='cuda:3'), in_proj_covar=tensor([0.0329, 0.0374, 0.0314, 0.0327, 0.0340, 0.0396, 0.0355, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:42:22,267 INFO [finetune.py:976] (3/7) Epoch 14, batch 2850, loss[loss=0.1775, simple_loss=0.2432, pruned_loss=0.05586, over 4789.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2466, pruned_loss=0.05505, over 955760.20 frames. ], batch size: 26, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:42:29,094 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2716, 3.2656, 2.5112, 3.8432, 3.3115, 3.2655, 1.3828, 3.2897], + device='cuda:3'), covar=tensor([0.2022, 0.1695, 0.3413, 0.2326, 0.4754, 0.2260, 0.6392, 0.2792], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0213, 0.0249, 0.0303, 0.0297, 0.0246, 0.0271, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:42:37,042 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 05:42:37,435 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77334.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:42:39,837 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77338.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:42:44,407 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.666e+02 1.986e+02 2.404e+02 4.400e+02, threshold=3.972e+02, percent-clipped=2.0 +2023-04-27 05:42:52,749 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0737, 1.4941, 1.8712, 2.1473, 1.8699, 1.4816, 1.0693, 1.5977], + device='cuda:3'), covar=tensor([0.3176, 0.3393, 0.1736, 0.2215, 0.2645, 0.2691, 0.4191, 0.2198], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0246, 0.0222, 0.0315, 0.0214, 0.0229, 0.0228, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 05:42:55,656 INFO [finetune.py:976] (3/7) Epoch 14, batch 2900, loss[loss=0.2106, simple_loss=0.2695, pruned_loss=0.07585, over 4940.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2491, pruned_loss=0.05592, over 954976.65 frames. ], batch size: 33, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:43:21,199 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9658, 1.4002, 1.3153, 1.6270, 1.4891, 1.7187, 1.2887, 3.0558], + device='cuda:3'), covar=tensor([0.0695, 0.0830, 0.0816, 0.1225, 0.0680, 0.0560, 0.0789, 0.0166], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 05:43:29,375 INFO [finetune.py:976] (3/7) Epoch 14, batch 2950, loss[loss=0.1649, simple_loss=0.2342, pruned_loss=0.04781, over 4807.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2524, pruned_loss=0.05671, over 954617.34 frames. ], batch size: 51, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:43:50,926 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.615e+02 1.939e+02 2.289e+02 5.440e+02, threshold=3.878e+02, percent-clipped=2.0 +2023-04-27 05:43:56,099 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77451.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:44:03,066 INFO [finetune.py:976] (3/7) Epoch 14, batch 3000, loss[loss=0.2302, simple_loss=0.2896, pruned_loss=0.08538, over 4738.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2543, pruned_loss=0.05718, over 955746.13 frames. ], batch size: 54, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:44:03,067 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 05:44:09,608 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6826, 1.4548, 0.6593, 1.3411, 1.4320, 1.6026, 1.4327, 1.4381], + device='cuda:3'), covar=tensor([0.0526, 0.0395, 0.0386, 0.0545, 0.0287, 0.0516, 0.0490, 0.0566], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 05:44:11,518 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5793, 1.1374, 1.3243, 1.2038, 1.7086, 1.4166, 1.1184, 1.2986], + device='cuda:3'), covar=tensor([0.1385, 0.1412, 0.1831, 0.1427, 0.0777, 0.1232, 0.1770, 0.1896], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0317, 0.0353, 0.0291, 0.0329, 0.0315, 0.0305, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.3763e-05, 6.6617e-05, 7.5619e-05, 5.9541e-05, 6.8467e-05, 6.6835e-05, + 6.4513e-05, 7.7367e-05], device='cuda:3') +2023-04-27 05:44:19,534 INFO [finetune.py:1010] (3/7) Epoch 14, validation: loss=0.1527, simple_loss=0.224, pruned_loss=0.04073, over 2265189.00 frames. +2023-04-27 05:44:19,535 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 05:45:02,948 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:45:24,239 INFO [finetune.py:976] (3/7) Epoch 14, batch 3050, loss[loss=0.186, simple_loss=0.2592, pruned_loss=0.05641, over 4818.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2543, pruned_loss=0.05684, over 957096.60 frames. ], batch size: 47, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:45:35,392 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6681, 3.3664, 1.0654, 1.9310, 2.0464, 2.3512, 2.1035, 1.2434], + device='cuda:3'), covar=tensor([0.1293, 0.1090, 0.1838, 0.1148, 0.0974, 0.1054, 0.1479, 0.1705], + device='cuda:3'), in_proj_covar=tensor([0.0119, 0.0246, 0.0139, 0.0121, 0.0132, 0.0153, 0.0119, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 05:45:45,782 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77543.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:45:46,941 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.289e+02 1.813e+02 2.136e+02 2.467e+02 4.229e+02, threshold=4.273e+02, percent-clipped=1.0 +2023-04-27 05:45:57,727 INFO [finetune.py:976] (3/7) Epoch 14, batch 3100, loss[loss=0.1903, simple_loss=0.2623, pruned_loss=0.05918, over 4847.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2524, pruned_loss=0.05644, over 956743.71 frames. ], batch size: 49, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:46:02,400 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77567.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:46:08,253 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9699, 2.5449, 1.9003, 1.9041, 1.3633, 1.3900, 2.0032, 1.3255], + device='cuda:3'), covar=tensor([0.1840, 0.1542, 0.1664, 0.1899, 0.2612, 0.2237, 0.1181, 0.2290], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0213, 0.0168, 0.0204, 0.0200, 0.0184, 0.0156, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 05:46:13,736 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-27 05:46:33,362 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1096, 1.6417, 2.0030, 2.4034, 1.9846, 1.5480, 1.2447, 1.7648], + device='cuda:3'), covar=tensor([0.3406, 0.3470, 0.1721, 0.2211, 0.2666, 0.2700, 0.4247, 0.2198], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0247, 0.0223, 0.0317, 0.0215, 0.0229, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 05:46:36,287 INFO [finetune.py:976] (3/7) Epoch 14, batch 3150, loss[loss=0.1461, simple_loss=0.2164, pruned_loss=0.03792, over 4939.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2493, pruned_loss=0.05579, over 954963.34 frames. ], batch size: 33, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:46:45,249 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:47:05,388 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77629.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:47:07,856 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77633.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:47:18,746 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3228, 1.8193, 2.2175, 2.6710, 2.2056, 1.7251, 1.4559, 2.1410], + device='cuda:3'), covar=tensor([0.3228, 0.3334, 0.1529, 0.2578, 0.2793, 0.2687, 0.4268, 0.2183], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0246, 0.0222, 0.0315, 0.0214, 0.0228, 0.0228, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 05:47:20,987 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.865e+01 1.595e+02 1.944e+02 2.388e+02 4.637e+02, threshold=3.889e+02, percent-clipped=1.0 +2023-04-27 05:47:41,186 INFO [finetune.py:976] (3/7) Epoch 14, batch 3200, loss[loss=0.206, simple_loss=0.2622, pruned_loss=0.07485, over 4891.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2481, pruned_loss=0.05587, over 955060.63 frames. ], batch size: 36, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:48:02,288 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8036, 2.0435, 1.9085, 1.6211, 2.0420, 1.6332, 2.5946, 1.4588], + device='cuda:3'), covar=tensor([0.3424, 0.1432, 0.3993, 0.2604, 0.1398, 0.2328, 0.1184, 0.4487], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0346, 0.0428, 0.0355, 0.0382, 0.0384, 0.0372, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:48:05,069 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:48:05,079 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:48:33,215 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3723, 1.0943, 1.1323, 1.2011, 1.5576, 1.2956, 1.1036, 1.1061], + device='cuda:3'), covar=tensor([0.1381, 0.1009, 0.1402, 0.1131, 0.0641, 0.1200, 0.1379, 0.1556], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0317, 0.0352, 0.0291, 0.0329, 0.0316, 0.0305, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.4054e-05, 6.6525e-05, 7.5311e-05, 5.9394e-05, 6.8461e-05, 6.7009e-05, + 6.4597e-05, 7.7171e-05], device='cuda:3') +2023-04-27 05:48:48,235 INFO [finetune.py:976] (3/7) Epoch 14, batch 3250, loss[loss=0.2399, simple_loss=0.3069, pruned_loss=0.08649, over 4812.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2491, pruned_loss=0.05671, over 955258.72 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:49:28,776 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77738.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:49:33,562 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.660e+02 2.088e+02 2.486e+02 4.990e+02, threshold=4.175e+02, percent-clipped=6.0 +2023-04-27 05:49:36,701 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77751.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:49:42,677 INFO [finetune.py:976] (3/7) Epoch 14, batch 3300, loss[loss=0.1759, simple_loss=0.2409, pruned_loss=0.05546, over 4843.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.253, pruned_loss=0.05835, over 954118.35 frames. ], batch size: 25, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:49:45,825 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 05:50:08,405 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77799.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:50:15,838 INFO [finetune.py:976] (3/7) Epoch 14, batch 3350, loss[loss=0.1601, simple_loss=0.2309, pruned_loss=0.04465, over 4761.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2531, pruned_loss=0.05776, over 953080.53 frames. ], batch size: 28, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:50:29,691 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-27 05:50:38,223 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4781, 1.1379, 1.2637, 1.1732, 1.6759, 1.3382, 1.0893, 1.1638], + device='cuda:3'), covar=tensor([0.1657, 0.1280, 0.2087, 0.1374, 0.0824, 0.1441, 0.1815, 0.2127], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0317, 0.0353, 0.0292, 0.0330, 0.0316, 0.0304, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.4074e-05, 6.6538e-05, 7.5592e-05, 5.9578e-05, 6.8755e-05, 6.6873e-05, + 6.4366e-05, 7.7336e-05], device='cuda:3') +2023-04-27 05:50:39,907 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.645e+02 1.956e+02 2.306e+02 5.075e+02, threshold=3.912e+02, percent-clipped=1.0 +2023-04-27 05:50:43,794 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 05:50:44,364 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8555, 2.3099, 1.8107, 1.5982, 1.3264, 1.3145, 1.9333, 1.2331], + device='cuda:3'), covar=tensor([0.1624, 0.1432, 0.1487, 0.1855, 0.2476, 0.2133, 0.0997, 0.2112], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0213, 0.0168, 0.0204, 0.0201, 0.0184, 0.0157, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 05:50:49,108 INFO [finetune.py:976] (3/7) Epoch 14, batch 3400, loss[loss=0.1957, simple_loss=0.2676, pruned_loss=0.06192, over 4789.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2536, pruned_loss=0.05776, over 954153.98 frames. ], batch size: 29, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:50:52,846 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77867.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:51:14,649 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7079, 2.1785, 1.8426, 2.0841, 1.4992, 1.8863, 1.9944, 1.3891], + device='cuda:3'), covar=tensor([0.2492, 0.1960, 0.1322, 0.1689, 0.3634, 0.1603, 0.2005, 0.2847], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0311, 0.0224, 0.0283, 0.0312, 0.0265, 0.0254, 0.0271], + device='cuda:3'), out_proj_covar=tensor([1.1759e-04, 1.2368e-04, 8.9617e-05, 1.1309e-04, 1.2729e-04, 1.0604e-04, + 1.0301e-04, 1.0835e-04], device='cuda:3') +2023-04-27 05:51:22,437 INFO [finetune.py:976] (3/7) Epoch 14, batch 3450, loss[loss=0.1698, simple_loss=0.2416, pruned_loss=0.04903, over 4797.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2532, pruned_loss=0.0568, over 954544.88 frames. ], batch size: 51, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:51:24,871 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77915.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:51:33,834 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:51:36,786 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:51:43,795 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6144, 2.1794, 2.5224, 2.8510, 2.8457, 2.5651, 2.0362, 2.6510], + device='cuda:3'), covar=tensor([0.0737, 0.1022, 0.0522, 0.0526, 0.0528, 0.0684, 0.0685, 0.0456], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0203, 0.0183, 0.0173, 0.0179, 0.0183, 0.0155, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:51:45,495 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.602e+02 1.891e+02 2.396e+02 5.235e+02, threshold=3.783e+02, percent-clipped=1.0 +2023-04-27 05:51:54,677 INFO [finetune.py:976] (3/7) Epoch 14, batch 3500, loss[loss=0.1958, simple_loss=0.2626, pruned_loss=0.06449, over 4799.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2509, pruned_loss=0.05599, over 954177.53 frames. ], batch size: 51, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:52:01,405 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:52:04,451 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:52:07,431 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:52:28,914 INFO [finetune.py:976] (3/7) Epoch 14, batch 3550, loss[loss=0.1572, simple_loss=0.2363, pruned_loss=0.03907, over 4903.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2488, pruned_loss=0.05557, over 954228.99 frames. ], batch size: 36, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:52:42,974 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78033.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:52:56,888 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.685e+02 1.967e+02 2.304e+02 5.361e+02, threshold=3.934e+02, percent-clipped=3.0 +2023-04-27 05:53:17,517 INFO [finetune.py:976] (3/7) Epoch 14, batch 3600, loss[loss=0.1797, simple_loss=0.2445, pruned_loss=0.05748, over 4793.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2463, pruned_loss=0.05537, over 952035.43 frames. ], batch size: 25, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:53:31,507 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 05:54:18,584 INFO [finetune.py:976] (3/7) Epoch 14, batch 3650, loss[loss=0.205, simple_loss=0.2724, pruned_loss=0.06878, over 4859.00 frames. ], tot_loss[loss=0.181, simple_loss=0.249, pruned_loss=0.05652, over 954280.75 frames. ], batch size: 44, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:54:22,381 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7078, 2.2905, 2.8240, 3.0437, 3.1434, 2.6701, 2.0713, 2.5719], + device='cuda:3'), covar=tensor([0.0854, 0.1154, 0.0581, 0.0584, 0.0579, 0.0746, 0.0754, 0.0643], + device='cuda:3'), in_proj_covar=tensor([0.0191, 0.0202, 0.0182, 0.0173, 0.0178, 0.0182, 0.0154, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 05:54:48,422 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.0346, 2.2313, 2.1517, 2.3135, 2.0868, 2.2416, 2.2990, 2.1688], + device='cuda:3'), covar=tensor([0.4202, 0.7184, 0.5652, 0.5441, 0.6682, 0.8633, 0.6755, 0.6625], + device='cuda:3'), in_proj_covar=tensor([0.0330, 0.0374, 0.0317, 0.0328, 0.0340, 0.0397, 0.0353, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 05:54:56,789 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.566e+02 1.842e+02 2.272e+02 4.887e+02, threshold=3.683e+02, percent-clipped=1.0 +2023-04-27 05:55:18,207 INFO [finetune.py:976] (3/7) Epoch 14, batch 3700, loss[loss=0.2394, simple_loss=0.3001, pruned_loss=0.08937, over 4801.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2526, pruned_loss=0.05758, over 951377.01 frames. ], batch size: 45, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:55:18,961 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1353, 1.6107, 2.0404, 2.6065, 1.9962, 1.5238, 1.3981, 1.9503], + device='cuda:3'), covar=tensor([0.3270, 0.3446, 0.1746, 0.2284, 0.2950, 0.2896, 0.4254, 0.2173], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0247, 0.0223, 0.0316, 0.0216, 0.0230, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 05:55:44,099 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 05:55:50,537 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6664, 1.1368, 1.7601, 2.1327, 1.7466, 1.6327, 1.7251, 1.6829], + device='cuda:3'), covar=tensor([0.4744, 0.6733, 0.6670, 0.6207, 0.5953, 0.8458, 0.8125, 0.8762], + device='cuda:3'), in_proj_covar=tensor([0.0412, 0.0406, 0.0494, 0.0509, 0.0442, 0.0464, 0.0471, 0.0473], + device='cuda:3'), out_proj_covar=tensor([9.9698e-05, 1.0050e-04, 1.1113e-04, 1.2094e-04, 1.0667e-04, 1.1174e-04, + 1.1232e-04, 1.1275e-04], device='cuda:3') +2023-04-27 05:55:56,843 INFO [finetune.py:976] (3/7) Epoch 14, batch 3750, loss[loss=0.2208, simple_loss=0.297, pruned_loss=0.07228, over 4809.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2544, pruned_loss=0.05818, over 952972.98 frames. ], batch size: 45, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:56:09,789 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:56:18,616 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.305e+02 1.679e+02 1.977e+02 2.535e+02 5.677e+02, threshold=3.954e+02, percent-clipped=2.0 +2023-04-27 05:56:30,168 INFO [finetune.py:976] (3/7) Epoch 14, batch 3800, loss[loss=0.2177, simple_loss=0.284, pruned_loss=0.07567, over 4814.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2557, pruned_loss=0.05814, over 953981.80 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:56:37,454 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:56:50,292 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78293.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:56:50,336 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.70 vs. limit=5.0 +2023-04-27 05:56:51,663 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 05:56:53,269 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8389, 1.4062, 1.5114, 1.5982, 2.0258, 1.6314, 1.3341, 1.4021], + device='cuda:3'), covar=tensor([0.1479, 0.1229, 0.1860, 0.1265, 0.0668, 0.1396, 0.1776, 0.1829], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0315, 0.0351, 0.0290, 0.0327, 0.0314, 0.0303, 0.0359], + device='cuda:3'), out_proj_covar=tensor([6.3475e-05, 6.6138e-05, 7.5261e-05, 5.9312e-05, 6.8075e-05, 6.6447e-05, + 6.4073e-05, 7.6792e-05], device='cuda:3') +2023-04-27 05:56:55,046 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7161, 3.4276, 2.5707, 2.7714, 2.0205, 1.9852, 2.8609, 2.0549], + device='cuda:3'), covar=tensor([0.1412, 0.1229, 0.1239, 0.1358, 0.2009, 0.1758, 0.0746, 0.1706], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0215, 0.0169, 0.0206, 0.0202, 0.0186, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 05:57:03,403 INFO [finetune.py:976] (3/7) Epoch 14, batch 3850, loss[loss=0.2076, simple_loss=0.2761, pruned_loss=0.0696, over 4802.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2532, pruned_loss=0.05691, over 952167.02 frames. ], batch size: 39, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:57:09,828 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:57:17,839 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78333.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:57:25,786 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.524e+02 1.884e+02 2.249e+02 3.539e+02, threshold=3.767e+02, percent-clipped=0.0 +2023-04-27 05:57:36,828 INFO [finetune.py:976] (3/7) Epoch 14, batch 3900, loss[loss=0.1738, simple_loss=0.2506, pruned_loss=0.04855, over 4820.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2503, pruned_loss=0.05601, over 953620.51 frames. ], batch size: 30, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:57:50,412 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78381.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:57:50,481 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3862, 1.3162, 1.5779, 1.5885, 1.2794, 1.1741, 1.2908, 0.7696], + device='cuda:3'), covar=tensor([0.0504, 0.0638, 0.0424, 0.0629, 0.0767, 0.1187, 0.0535, 0.0668], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0070, 0.0069, 0.0068, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 05:57:59,683 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8538, 2.2278, 1.1051, 1.5491, 2.1071, 1.7269, 1.6407, 1.7840], + device='cuda:3'), covar=tensor([0.0507, 0.0332, 0.0303, 0.0542, 0.0239, 0.0520, 0.0487, 0.0536], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 05:58:09,807 INFO [finetune.py:976] (3/7) Epoch 14, batch 3950, loss[loss=0.21, simple_loss=0.2638, pruned_loss=0.07805, over 4711.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.247, pruned_loss=0.05486, over 954625.78 frames. ], batch size: 54, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:58:50,265 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.573e+02 1.912e+02 2.257e+02 5.641e+02, threshold=3.824e+02, percent-clipped=2.0 +2023-04-27 05:59:04,625 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78460.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:59:05,121 INFO [finetune.py:976] (3/7) Epoch 14, batch 4000, loss[loss=0.1757, simple_loss=0.2488, pruned_loss=0.05126, over 4800.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.246, pruned_loss=0.05459, over 954738.48 frames. ], batch size: 29, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:00:00,329 INFO [finetune.py:976] (3/7) Epoch 14, batch 4050, loss[loss=0.1968, simple_loss=0.2783, pruned_loss=0.05762, over 4813.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2493, pruned_loss=0.05615, over 954497.24 frames. ], batch size: 51, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:00:19,666 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78521.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:00:29,322 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1819, 2.6284, 0.9939, 1.3288, 2.0597, 1.2695, 3.5776, 1.6239], + device='cuda:3'), covar=tensor([0.0656, 0.0710, 0.0839, 0.1329, 0.0508, 0.1085, 0.0284, 0.0759], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 06:00:51,058 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.700e+02 2.035e+02 2.375e+02 5.106e+02, threshold=4.071e+02, percent-clipped=4.0 +2023-04-27 06:01:03,063 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78556.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:01:06,005 INFO [finetune.py:976] (3/7) Epoch 14, batch 4100, loss[loss=0.1639, simple_loss=0.2331, pruned_loss=0.04734, over 4710.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2522, pruned_loss=0.0569, over 954091.95 frames. ], batch size: 23, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:01:33,687 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1996, 1.7061, 2.0821, 2.5440, 2.6364, 2.0670, 1.7638, 2.2422], + device='cuda:3'), covar=tensor([0.0873, 0.1178, 0.0658, 0.0500, 0.0535, 0.0792, 0.0729, 0.0559], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0204, 0.0183, 0.0173, 0.0178, 0.0184, 0.0154, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:01:34,974 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6100, 1.1269, 1.7100, 2.0893, 1.7189, 1.5648, 1.6449, 1.6382], + device='cuda:3'), covar=tensor([0.5098, 0.7464, 0.6881, 0.6200, 0.6487, 0.8765, 0.8171, 0.9490], + device='cuda:3'), in_proj_covar=tensor([0.0413, 0.0406, 0.0495, 0.0510, 0.0443, 0.0463, 0.0472, 0.0475], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:01:45,288 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78588.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:02:05,269 INFO [finetune.py:976] (3/7) Epoch 14, batch 4150, loss[loss=0.1996, simple_loss=0.276, pruned_loss=0.06161, over 4821.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2526, pruned_loss=0.05675, over 954133.55 frames. ], batch size: 55, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:02:09,552 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:02:17,923 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.62 vs. limit=5.0 +2023-04-27 06:02:29,498 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.351e+02 1.709e+02 1.960e+02 2.352e+02 3.930e+02, threshold=3.920e+02, percent-clipped=0.0 +2023-04-27 06:02:38,719 INFO [finetune.py:976] (3/7) Epoch 14, batch 4200, loss[loss=0.1829, simple_loss=0.2567, pruned_loss=0.05459, over 4879.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2523, pruned_loss=0.0563, over 954753.83 frames. ], batch size: 35, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:02:50,113 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8541, 2.3789, 1.7797, 1.6943, 1.2662, 1.3298, 1.8826, 1.2322], + device='cuda:3'), covar=tensor([0.1589, 0.1223, 0.1446, 0.1726, 0.2367, 0.1976, 0.0995, 0.2087], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0213, 0.0168, 0.0205, 0.0201, 0.0185, 0.0157, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 06:03:03,634 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6502, 2.3662, 2.8538, 3.0662, 3.1484, 2.4623, 2.0033, 2.6257], + device='cuda:3'), covar=tensor([0.0748, 0.0890, 0.0466, 0.0496, 0.0485, 0.0776, 0.0695, 0.0559], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0203, 0.0183, 0.0173, 0.0178, 0.0184, 0.0154, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:03:12,004 INFO [finetune.py:976] (3/7) Epoch 14, batch 4250, loss[loss=0.1399, simple_loss=0.2096, pruned_loss=0.03515, over 4900.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2507, pruned_loss=0.0556, over 954950.31 frames. ], batch size: 37, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:03:36,180 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.058e+02 1.588e+02 1.905e+02 2.240e+02 4.270e+02, threshold=3.810e+02, percent-clipped=2.0 +2023-04-27 06:03:42,497 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7246, 1.6254, 2.0801, 2.1140, 1.5068, 1.3666, 1.6304, 1.2680], + device='cuda:3'), covar=tensor([0.0582, 0.0798, 0.0454, 0.0549, 0.0804, 0.1284, 0.0750, 0.0652], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0070, 0.0069, 0.0067, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:03:45,442 INFO [finetune.py:976] (3/7) Epoch 14, batch 4300, loss[loss=0.1406, simple_loss=0.2211, pruned_loss=0.03003, over 4862.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2476, pruned_loss=0.05467, over 956059.26 frames. ], batch size: 31, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:04:19,293 INFO [finetune.py:976] (3/7) Epoch 14, batch 4350, loss[loss=0.1995, simple_loss=0.2667, pruned_loss=0.06611, over 4939.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2459, pruned_loss=0.05385, over 958822.16 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:04:22,429 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78816.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:04:43,556 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.581e+02 1.858e+02 2.218e+02 4.471e+02, threshold=3.716e+02, percent-clipped=1.0 +2023-04-27 06:04:55,416 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3338, 1.3981, 3.8916, 3.5803, 3.4225, 3.7534, 3.7193, 3.4120], + device='cuda:3'), covar=tensor([0.6896, 0.5529, 0.1135, 0.1992, 0.1179, 0.2151, 0.1345, 0.1672], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0305, 0.0403, 0.0405, 0.0347, 0.0405, 0.0312, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:05:04,243 INFO [finetune.py:976] (3/7) Epoch 14, batch 4400, loss[loss=0.2201, simple_loss=0.2861, pruned_loss=0.07704, over 4937.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.247, pruned_loss=0.05465, over 957645.02 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:05:38,860 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:06:10,128 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 06:06:11,507 INFO [finetune.py:976] (3/7) Epoch 14, batch 4450, loss[loss=0.1786, simple_loss=0.2464, pruned_loss=0.05539, over 4744.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.251, pruned_loss=0.05598, over 959335.25 frames. ], batch size: 27, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:06:12,194 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78912.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:06:21,379 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3359, 1.6448, 1.5839, 1.7505, 1.6967, 1.9450, 1.5415, 3.1957], + device='cuda:3'), covar=tensor([0.0584, 0.0665, 0.0696, 0.1044, 0.0558, 0.0648, 0.0667, 0.0179], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 06:06:23,861 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7504, 1.6919, 0.7464, 1.4106, 1.7781, 1.6282, 1.4786, 1.5133], + device='cuda:3'), covar=tensor([0.0494, 0.0388, 0.0359, 0.0547, 0.0280, 0.0512, 0.0534, 0.0583], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 06:06:44,186 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:07:03,713 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.730e+02 2.046e+02 2.407e+02 5.260e+02, threshold=4.092e+02, percent-clipped=6.0 +2023-04-27 06:07:18,901 INFO [finetune.py:976] (3/7) Epoch 14, batch 4500, loss[loss=0.1933, simple_loss=0.2588, pruned_loss=0.06393, over 4140.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2508, pruned_loss=0.05553, over 955997.90 frames. ], batch size: 66, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:07:21,397 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78965.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:07:43,788 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0080, 1.9180, 2.4058, 2.4123, 1.8246, 1.6326, 1.9593, 1.1303], + device='cuda:3'), covar=tensor([0.0627, 0.0794, 0.0448, 0.0607, 0.0802, 0.1166, 0.0716, 0.0837], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0069, 0.0069, 0.0067, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:07:52,199 INFO [finetune.py:976] (3/7) Epoch 14, batch 4550, loss[loss=0.1878, simple_loss=0.2667, pruned_loss=0.05444, over 4901.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2523, pruned_loss=0.05642, over 957223.41 frames. ], batch size: 36, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:07:56,085 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-27 06:08:01,460 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79026.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:08:14,993 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.614e+01 1.580e+02 1.894e+02 2.383e+02 3.819e+02, threshold=3.787e+02, percent-clipped=0.0 +2023-04-27 06:08:26,107 INFO [finetune.py:976] (3/7) Epoch 14, batch 4600, loss[loss=0.1773, simple_loss=0.2502, pruned_loss=0.05221, over 4883.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2532, pruned_loss=0.05642, over 958668.34 frames. ], batch size: 32, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:08:39,519 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8696, 1.5603, 1.4102, 1.7052, 2.0979, 1.6047, 1.4198, 1.3167], + device='cuda:3'), covar=tensor([0.1519, 0.1402, 0.1905, 0.1188, 0.0691, 0.1723, 0.2201, 0.2307], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0317, 0.0354, 0.0292, 0.0329, 0.0315, 0.0303, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.3950e-05, 6.6487e-05, 7.6048e-05, 5.9606e-05, 6.8388e-05, 6.6646e-05, + 6.4084e-05, 7.7189e-05], device='cuda:3') +2023-04-27 06:08:59,225 INFO [finetune.py:976] (3/7) Epoch 14, batch 4650, loss[loss=0.1506, simple_loss=0.2263, pruned_loss=0.03742, over 4750.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2508, pruned_loss=0.05558, over 957560.24 frames. ], batch size: 26, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:09:01,788 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8704, 2.0611, 0.9317, 1.6429, 2.2577, 1.7241, 1.7651, 1.7237], + device='cuda:3'), covar=tensor([0.0457, 0.0344, 0.0319, 0.0528, 0.0231, 0.0489, 0.0447, 0.0522], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 06:09:02,358 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:09:21,104 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8164, 2.3008, 0.9552, 1.5510, 2.3276, 1.6794, 1.6871, 1.7247], + device='cuda:3'), covar=tensor([0.0486, 0.0310, 0.0335, 0.0553, 0.0249, 0.0515, 0.0506, 0.0552], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 06:09:21,596 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.186e+02 1.677e+02 1.942e+02 2.274e+02 5.469e+02, threshold=3.883e+02, percent-clipped=3.0 +2023-04-27 06:09:22,950 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3785, 1.8963, 2.2562, 2.6716, 2.2560, 1.7088, 1.4301, 2.0929], + device='cuda:3'), covar=tensor([0.3087, 0.3210, 0.1622, 0.2447, 0.2587, 0.2724, 0.4219, 0.2031], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0247, 0.0223, 0.0316, 0.0215, 0.0229, 0.0229, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 06:09:32,658 INFO [finetune.py:976] (3/7) Epoch 14, batch 4700, loss[loss=0.1672, simple_loss=0.2329, pruned_loss=0.05077, over 4853.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2475, pruned_loss=0.05419, over 957923.49 frames. ], batch size: 44, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:09:34,544 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79164.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:10:05,832 INFO [finetune.py:976] (3/7) Epoch 14, batch 4750, loss[loss=0.2083, simple_loss=0.2831, pruned_loss=0.06675, over 4860.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2462, pruned_loss=0.05415, over 958112.39 frames. ], batch size: 44, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:10:06,533 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79212.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:10:09,415 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8718, 2.3152, 0.8517, 1.2334, 1.4839, 1.1100, 2.5172, 1.4137], + device='cuda:3'), covar=tensor([0.0688, 0.0564, 0.0695, 0.1233, 0.0483, 0.1061, 0.0334, 0.0689], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 06:10:37,980 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.597e+02 1.966e+02 2.340e+02 3.997e+02, threshold=3.932e+02, percent-clipped=2.0 +2023-04-27 06:10:58,673 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:10:59,212 INFO [finetune.py:976] (3/7) Epoch 14, batch 4800, loss[loss=0.2112, simple_loss=0.279, pruned_loss=0.0717, over 4769.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.249, pruned_loss=0.05543, over 955365.61 frames. ], batch size: 28, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:11:10,711 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1374, 1.4231, 1.3415, 1.6923, 1.6001, 1.6894, 1.3247, 3.0652], + device='cuda:3'), covar=tensor([0.0668, 0.0803, 0.0848, 0.1193, 0.0644, 0.0524, 0.0751, 0.0170], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 06:11:25,278 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79285.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:11:58,653 INFO [finetune.py:976] (3/7) Epoch 14, batch 4850, loss[loss=0.1869, simple_loss=0.2629, pruned_loss=0.05546, over 4811.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2514, pruned_loss=0.05641, over 954833.48 frames. ], batch size: 51, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:12:08,294 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6559, 2.8626, 2.4599, 2.5325, 2.9458, 2.6536, 3.8343, 2.1162], + device='cuda:3'), covar=tensor([0.4104, 0.2217, 0.4292, 0.3368, 0.2229, 0.2623, 0.1477, 0.4412], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0350, 0.0431, 0.0358, 0.0386, 0.0385, 0.0376, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:12:10,431 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:12:23,701 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79333.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:12:36,866 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.635e+02 2.018e+02 2.486e+02 3.725e+02, threshold=4.037e+02, percent-clipped=0.0 +2023-04-27 06:12:37,017 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79346.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:12:53,971 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7235, 1.7795, 1.0140, 1.3614, 1.8554, 1.5524, 1.4350, 1.5514], + device='cuda:3'), covar=tensor([0.0510, 0.0374, 0.0334, 0.0566, 0.0274, 0.0533, 0.0495, 0.0590], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 06:12:56,922 INFO [finetune.py:976] (3/7) Epoch 14, batch 4900, loss[loss=0.1652, simple_loss=0.2452, pruned_loss=0.04254, over 4824.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2526, pruned_loss=0.05681, over 954573.65 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:13:00,245 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79364.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:13:07,475 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79375.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:13:08,191 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 06:13:19,226 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79394.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:13:29,999 INFO [finetune.py:976] (3/7) Epoch 14, batch 4950, loss[loss=0.1834, simple_loss=0.2664, pruned_loss=0.05014, over 4758.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2533, pruned_loss=0.05636, over 955172.01 frames. ], batch size: 59, lr: 3.52e-03, grad_scale: 32.0 +2023-04-27 06:13:40,498 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:13:47,630 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79436.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:13:53,546 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.668e+02 1.946e+02 2.385e+02 4.906e+02, threshold=3.893e+02, percent-clipped=2.0 +2023-04-27 06:14:03,213 INFO [finetune.py:976] (3/7) Epoch 14, batch 5000, loss[loss=0.1695, simple_loss=0.2274, pruned_loss=0.05578, over 4748.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2518, pruned_loss=0.0561, over 953790.03 frames. ], batch size: 26, lr: 3.52e-03, grad_scale: 32.0 +2023-04-27 06:14:14,515 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-27 06:14:30,832 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0906, 2.0311, 1.7811, 1.6542, 2.2057, 1.6545, 2.6536, 1.5856], + device='cuda:3'), covar=tensor([0.3547, 0.1982, 0.4763, 0.3030, 0.1668, 0.2716, 0.1440, 0.4511], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0350, 0.0432, 0.0357, 0.0385, 0.0384, 0.0376, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:14:33,270 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1924, 1.7390, 2.0558, 2.6039, 2.1224, 1.5796, 1.3265, 1.9725], + device='cuda:3'), covar=tensor([0.2978, 0.2932, 0.1557, 0.1988, 0.2404, 0.2555, 0.4138, 0.1916], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0248, 0.0224, 0.0317, 0.0216, 0.0231, 0.0230, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 06:14:35,549 INFO [finetune.py:976] (3/7) Epoch 14, batch 5050, loss[loss=0.1558, simple_loss=0.2305, pruned_loss=0.0405, over 4764.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2495, pruned_loss=0.05578, over 952478.40 frames. ], batch size: 26, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:14:59,516 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.519e+02 1.891e+02 2.332e+02 5.445e+02, threshold=3.783e+02, percent-clipped=1.0 +2023-04-27 06:15:08,088 INFO [finetune.py:976] (3/7) Epoch 14, batch 5100, loss[loss=0.1642, simple_loss=0.2327, pruned_loss=0.0478, over 4819.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2457, pruned_loss=0.05408, over 952793.34 frames. ], batch size: 38, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:15:25,154 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79585.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:15:41,434 INFO [finetune.py:976] (3/7) Epoch 14, batch 5150, loss[loss=0.1762, simple_loss=0.2632, pruned_loss=0.04461, over 4827.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.246, pruned_loss=0.05449, over 953109.05 frames. ], batch size: 39, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:15:48,511 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79621.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:02,434 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79641.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:16:06,012 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79646.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:06,492 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.596e+02 1.950e+02 2.248e+02 3.363e+02, threshold=3.899e+02, percent-clipped=0.0 +2023-04-27 06:16:12,026 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6118, 2.0523, 1.7582, 1.8990, 1.4504, 1.6712, 1.7112, 1.4030], + device='cuda:3'), covar=tensor([0.1647, 0.1030, 0.0728, 0.0968, 0.2819, 0.0989, 0.1585, 0.1958], + device='cuda:3'), in_proj_covar=tensor([0.0289, 0.0309, 0.0223, 0.0281, 0.0311, 0.0263, 0.0253, 0.0268], + device='cuda:3'), out_proj_covar=tensor([1.1662e-04, 1.2320e-04, 8.8784e-05, 1.1193e-04, 1.2656e-04, 1.0498e-04, + 1.0239e-04, 1.0686e-04], device='cuda:3') +2023-04-27 06:16:12,606 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79657.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:14,969 INFO [finetune.py:976] (3/7) Epoch 14, batch 5200, loss[loss=0.208, simple_loss=0.2594, pruned_loss=0.07825, over 4805.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.25, pruned_loss=0.05659, over 953117.24 frames. ], batch size: 25, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:16:19,855 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79669.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:31,045 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4131, 1.8662, 1.7957, 1.9513, 1.7619, 1.9640, 1.8399, 1.8248], + device='cuda:3'), covar=tensor([0.4438, 0.5697, 0.5153, 0.4656, 0.5838, 0.7170, 0.6149, 0.5672], + device='cuda:3'), in_proj_covar=tensor([0.0333, 0.0377, 0.0319, 0.0333, 0.0342, 0.0400, 0.0357, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 06:16:34,496 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79689.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:54,349 INFO [finetune.py:976] (3/7) Epoch 14, batch 5250, loss[loss=0.1442, simple_loss=0.2192, pruned_loss=0.03461, over 4763.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2527, pruned_loss=0.05723, over 953175.33 frames. ], batch size: 27, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:17:03,907 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:17:05,053 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:17:19,832 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79731.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 06:17:36,847 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.641e+02 2.072e+02 2.610e+02 5.253e+02, threshold=4.143e+02, percent-clipped=2.0 +2023-04-27 06:17:45,876 INFO [finetune.py:976] (3/7) Epoch 14, batch 5300, loss[loss=0.2626, simple_loss=0.3059, pruned_loss=0.1097, over 4894.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2547, pruned_loss=0.05863, over 954163.05 frames. ], batch size: 32, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:17:46,588 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79762.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:18:38,535 INFO [finetune.py:976] (3/7) Epoch 14, batch 5350, loss[loss=0.2537, simple_loss=0.3009, pruned_loss=0.1033, over 4884.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2555, pruned_loss=0.05881, over 955560.42 frames. ], batch size: 32, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:18:45,989 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79823.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:18:46,797 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.93 vs. limit=5.0 +2023-04-27 06:19:00,015 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4859, 1.3285, 4.3833, 4.0679, 3.7719, 4.0967, 4.0347, 3.8265], + device='cuda:3'), covar=tensor([0.7119, 0.5775, 0.0950, 0.1596, 0.1179, 0.1545, 0.1392, 0.1422], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0303, 0.0400, 0.0401, 0.0346, 0.0402, 0.0309, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:19:02,342 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.530e+02 1.860e+02 2.217e+02 4.388e+02, threshold=3.721e+02, percent-clipped=2.0 +2023-04-27 06:19:11,410 INFO [finetune.py:976] (3/7) Epoch 14, batch 5400, loss[loss=0.1545, simple_loss=0.2179, pruned_loss=0.0455, over 4793.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2529, pruned_loss=0.05782, over 954956.63 frames. ], batch size: 25, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:19:23,157 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-04-27 06:19:45,248 INFO [finetune.py:976] (3/7) Epoch 14, batch 5450, loss[loss=0.1631, simple_loss=0.2249, pruned_loss=0.0507, over 4902.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2505, pruned_loss=0.05747, over 954959.22 frames. ], batch size: 32, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:20:04,589 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:20:04,628 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79941.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:20:09,137 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.493e+02 1.714e+02 2.129e+02 3.310e+02, threshold=3.428e+02, percent-clipped=0.0 +2023-04-27 06:20:18,680 INFO [finetune.py:976] (3/7) Epoch 14, batch 5500, loss[loss=0.1747, simple_loss=0.2445, pruned_loss=0.05251, over 4850.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2461, pruned_loss=0.05559, over 955227.69 frames. ], batch size: 49, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:20:32,079 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4760, 1.3875, 1.7392, 1.7918, 1.4225, 1.1075, 1.2873, 0.8422], + device='cuda:3'), covar=tensor([0.0632, 0.0597, 0.0477, 0.0520, 0.0670, 0.1585, 0.0724, 0.0796], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0067, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:20:36,320 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79989.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:20:36,338 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:20:53,195 INFO [finetune.py:976] (3/7) Epoch 14, batch 5550, loss[loss=0.2536, simple_loss=0.3085, pruned_loss=0.09932, over 4762.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2485, pruned_loss=0.05699, over 955722.55 frames. ], batch size: 59, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:20:54,495 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80013.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:20:59,228 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80020.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:21:06,029 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80031.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:21:09,683 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80037.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:21:16,566 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.402e+01 1.777e+02 2.079e+02 2.504e+02 5.110e+02, threshold=4.158e+02, percent-clipped=3.0 +2023-04-27 06:21:24,823 INFO [finetune.py:976] (3/7) Epoch 14, batch 5600, loss[loss=0.1942, simple_loss=0.2667, pruned_loss=0.0609, over 4715.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2516, pruned_loss=0.05789, over 954654.61 frames. ], batch size: 59, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:21:28,973 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80068.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:21:35,802 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80079.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:21:46,828 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5978, 1.7817, 0.6803, 1.2776, 1.8531, 1.4054, 1.3459, 1.4031], + device='cuda:3'), covar=tensor([0.0538, 0.0378, 0.0390, 0.0585, 0.0281, 0.0550, 0.0542, 0.0612], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:3') +2023-04-27 06:22:01,085 INFO [finetune.py:976] (3/7) Epoch 14, batch 5650, loss[loss=0.2242, simple_loss=0.2784, pruned_loss=0.08504, over 4837.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2542, pruned_loss=0.05808, over 955772.99 frames. ], batch size: 49, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:22:11,110 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80118.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:22:28,365 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.596e+02 1.966e+02 2.394e+02 5.142e+02, threshold=3.932e+02, percent-clipped=2.0 +2023-04-27 06:22:47,957 INFO [finetune.py:976] (3/7) Epoch 14, batch 5700, loss[loss=0.1658, simple_loss=0.2185, pruned_loss=0.0566, over 4388.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2499, pruned_loss=0.0569, over 937109.55 frames. ], batch size: 19, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:23:36,604 INFO [finetune.py:976] (3/7) Epoch 15, batch 0, loss[loss=0.1676, simple_loss=0.2481, pruned_loss=0.04359, over 4775.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2481, pruned_loss=0.04359, over 4775.00 frames. ], batch size: 28, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:23:36,604 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 06:23:39,101 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7350, 1.8584, 1.7410, 1.2834, 1.9389, 1.5323, 2.3498, 1.5753], + device='cuda:3'), covar=tensor([0.3717, 0.1655, 0.5072, 0.2844, 0.1451, 0.2406, 0.1497, 0.4791], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0346, 0.0427, 0.0355, 0.0381, 0.0381, 0.0372, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:23:53,231 INFO [finetune.py:1010] (3/7) Epoch 15, validation: loss=0.1536, simple_loss=0.2258, pruned_loss=0.04063, over 2265189.00 frames. +2023-04-27 06:23:53,231 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 06:23:53,378 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3809, 1.8418, 2.2372, 2.8314, 2.2010, 1.7741, 1.7331, 2.1703], + device='cuda:3'), covar=tensor([0.3284, 0.3500, 0.1768, 0.2995, 0.3088, 0.2732, 0.4307, 0.2406], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0249, 0.0225, 0.0318, 0.0216, 0.0231, 0.0231, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 06:24:27,726 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-27 06:24:56,492 INFO [finetune.py:976] (3/7) Epoch 15, batch 50, loss[loss=0.1655, simple_loss=0.2442, pruned_loss=0.04339, over 4815.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2516, pruned_loss=0.0563, over 216343.22 frames. ], batch size: 39, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:25:05,171 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80241.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:25:08,721 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.528e+02 1.806e+02 2.187e+02 6.322e+02, threshold=3.611e+02, percent-clipped=3.0 +2023-04-27 06:25:36,398 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80283.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:25:40,292 INFO [finetune.py:976] (3/7) Epoch 15, batch 100, loss[loss=0.192, simple_loss=0.2549, pruned_loss=0.0646, over 4873.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2476, pruned_loss=0.05527, over 382158.23 frames. ], batch size: 34, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:25:41,452 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:25:45,565 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 06:25:56,997 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80313.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:26:12,986 INFO [finetune.py:976] (3/7) Epoch 15, batch 150, loss[loss=0.1891, simple_loss=0.2569, pruned_loss=0.06068, over 4922.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2435, pruned_loss=0.05403, over 510292.33 frames. ], batch size: 43, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:26:18,203 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80344.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:26:20,364 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.629e+02 1.838e+02 2.299e+02 3.632e+02, threshold=3.676e+02, percent-clipped=1.0 +2023-04-27 06:26:22,978 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1537, 2.2293, 1.9347, 1.8026, 2.4893, 1.7693, 2.9063, 1.7394], + device='cuda:3'), covar=tensor([0.3778, 0.1755, 0.4374, 0.2995, 0.1520, 0.2617, 0.1163, 0.4438], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0348, 0.0429, 0.0356, 0.0383, 0.0382, 0.0373, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:26:28,922 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80361.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:26:46,348 INFO [finetune.py:976] (3/7) Epoch 15, batch 200, loss[loss=0.1655, simple_loss=0.2368, pruned_loss=0.04703, over 4867.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2439, pruned_loss=0.05446, over 609822.95 frames. ], batch size: 31, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:26:50,391 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80393.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:27:06,644 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80418.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:27:11,533 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4783, 1.5317, 1.8091, 1.7932, 1.4631, 1.2035, 1.5798, 1.1631], + device='cuda:3'), covar=tensor([0.0692, 0.0722, 0.0448, 0.0732, 0.0766, 0.1347, 0.0587, 0.0664], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0071, 0.0070, 0.0068, 0.0076, 0.0098, 0.0076, 0.0070], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:27:16,977 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 06:27:18,653 INFO [finetune.py:976] (3/7) Epoch 15, batch 250, loss[loss=0.1808, simple_loss=0.2769, pruned_loss=0.04242, over 4821.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.247, pruned_loss=0.05538, over 688275.74 frames. ], batch size: 40, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:27:25,585 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.693e+02 2.067e+02 2.411e+02 4.714e+02, threshold=4.133e+02, percent-clipped=3.0 +2023-04-27 06:27:31,399 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80454.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:27:36,888 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80463.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:27:38,599 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:27:51,746 INFO [finetune.py:976] (3/7) Epoch 15, batch 300, loss[loss=0.152, simple_loss=0.2125, pruned_loss=0.04579, over 4720.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2509, pruned_loss=0.05647, over 747180.30 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:27:56,454 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80494.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:28:13,002 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1090, 1.5322, 1.4144, 1.6083, 1.5324, 1.6956, 1.3641, 3.2686], + device='cuda:3'), covar=tensor([0.0659, 0.0775, 0.0790, 0.1224, 0.0656, 0.0521, 0.0730, 0.0162], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 06:28:13,029 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4150, 1.3879, 1.6522, 1.6218, 1.2800, 1.1657, 1.4583, 0.9671], + device='cuda:3'), covar=tensor([0.0669, 0.0686, 0.0475, 0.0643, 0.0834, 0.1145, 0.0534, 0.0616], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0067, 0.0075, 0.0097, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:28:27,733 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80524.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:28:33,930 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 06:28:36,207 INFO [finetune.py:976] (3/7) Epoch 15, batch 350, loss[loss=0.2243, simple_loss=0.2882, pruned_loss=0.08023, over 4915.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2538, pruned_loss=0.0573, over 793066.86 frames. ], batch size: 37, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:28:42,200 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.655e+02 1.989e+02 2.509e+02 3.787e+02, threshold=3.978e+02, percent-clipped=0.0 +2023-04-27 06:28:48,613 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:28:56,125 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.3105, 4.2553, 2.9479, 4.9559, 4.2264, 4.2622, 1.8675, 4.3307], + device='cuda:3'), covar=tensor([0.1350, 0.0963, 0.3370, 0.0888, 0.2654, 0.1432, 0.5228, 0.1720], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0214, 0.0249, 0.0302, 0.0297, 0.0246, 0.0270, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 06:29:15,645 INFO [finetune.py:976] (3/7) Epoch 15, batch 400, loss[loss=0.1935, simple_loss=0.2496, pruned_loss=0.06868, over 4175.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2534, pruned_loss=0.05717, over 827703.84 frames. ], batch size: 65, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:29:34,709 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2672, 2.7346, 1.0528, 1.4299, 2.2952, 1.2519, 3.6687, 1.7817], + device='cuda:3'), covar=tensor([0.0614, 0.0566, 0.0723, 0.1202, 0.0448, 0.0996, 0.0273, 0.0620], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 06:30:07,686 INFO [finetune.py:976] (3/7) Epoch 15, batch 450, loss[loss=0.1932, simple_loss=0.254, pruned_loss=0.06614, over 4853.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2519, pruned_loss=0.05628, over 855560.78 frames. ], batch size: 31, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:30:13,731 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80639.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:30:18,557 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.255e+02 1.612e+02 1.974e+02 2.306e+02 4.694e+02, threshold=3.949e+02, percent-clipped=1.0 +2023-04-27 06:31:01,361 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8016, 3.6478, 1.1043, 1.9877, 2.0485, 2.6562, 2.0031, 1.0996], + device='cuda:3'), covar=tensor([0.1248, 0.0876, 0.1874, 0.1230, 0.0936, 0.0965, 0.1581, 0.1894], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0246, 0.0138, 0.0122, 0.0131, 0.0153, 0.0119, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:31:13,676 INFO [finetune.py:976] (3/7) Epoch 15, batch 500, loss[loss=0.167, simple_loss=0.2384, pruned_loss=0.04777, over 4767.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2496, pruned_loss=0.05553, over 879206.15 frames. ], batch size: 27, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:31:20,217 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4917, 2.4694, 2.1579, 2.1731, 2.6675, 2.1772, 3.3248, 1.9415], + device='cuda:3'), covar=tensor([0.3973, 0.2564, 0.4581, 0.3847, 0.1824, 0.2712, 0.2038, 0.4329], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0349, 0.0431, 0.0357, 0.0385, 0.0383, 0.0375, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:31:43,335 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4506, 1.6414, 1.4918, 1.8333, 1.7330, 1.8783, 1.4862, 3.6844], + device='cuda:3'), covar=tensor([0.0587, 0.0796, 0.0816, 0.1191, 0.0635, 0.0531, 0.0733, 0.0136], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 06:32:06,676 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:32:07,800 INFO [finetune.py:976] (3/7) Epoch 15, batch 550, loss[loss=0.1956, simple_loss=0.2606, pruned_loss=0.06534, over 4862.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.246, pruned_loss=0.05451, over 896327.52 frames. ], batch size: 34, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:32:13,248 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.610e+02 1.836e+02 2.201e+02 4.321e+02, threshold=3.672e+02, percent-clipped=1.0 +2023-04-27 06:32:14,548 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80749.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:32:32,580 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80773.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:32:41,568 INFO [finetune.py:976] (3/7) Epoch 15, batch 600, loss[loss=0.1504, simple_loss=0.2115, pruned_loss=0.04464, over 4033.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2461, pruned_loss=0.0548, over 909450.35 frames. ], batch size: 17, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:32:47,139 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:33:03,035 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:33:12,674 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80834.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:33:15,002 INFO [finetune.py:976] (3/7) Epoch 15, batch 650, loss[loss=0.1984, simple_loss=0.2674, pruned_loss=0.06471, over 4917.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.25, pruned_loss=0.0561, over 918585.21 frames. ], batch size: 38, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:33:20,400 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.693e+02 1.997e+02 2.392e+02 3.553e+02, threshold=3.994e+02, percent-clipped=0.0 +2023-04-27 06:33:21,952 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 06:33:22,326 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:33:25,979 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9954, 1.6258, 1.3865, 1.9787, 2.2297, 1.8165, 1.7494, 1.3594], + device='cuda:3'), covar=tensor([0.1966, 0.1648, 0.2239, 0.1375, 0.1303, 0.1888, 0.2151, 0.2600], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0318, 0.0354, 0.0292, 0.0331, 0.0314, 0.0304, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.3972e-05, 6.6635e-05, 7.5830e-05, 5.9660e-05, 6.9048e-05, 6.6414e-05, + 6.4229e-05, 7.7472e-05], device='cuda:3') +2023-04-27 06:33:40,485 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-04-27 06:33:44,803 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-04-27 06:33:48,176 INFO [finetune.py:976] (3/7) Epoch 15, batch 700, loss[loss=0.1909, simple_loss=0.2536, pruned_loss=0.06413, over 4904.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2513, pruned_loss=0.0564, over 924641.55 frames. ], batch size: 32, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:34:21,753 INFO [finetune.py:976] (3/7) Epoch 15, batch 750, loss[loss=0.2181, simple_loss=0.2907, pruned_loss=0.07273, over 4906.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2537, pruned_loss=0.05766, over 930943.69 frames. ], batch size: 36, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:34:22,437 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80939.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:34:27,258 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.649e+02 1.974e+02 2.385e+02 5.578e+02, threshold=3.948e+02, percent-clipped=5.0 +2023-04-27 06:34:54,973 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:34:55,527 INFO [finetune.py:976] (3/7) Epoch 15, batch 800, loss[loss=0.1739, simple_loss=0.2413, pruned_loss=0.05322, over 4793.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2538, pruned_loss=0.05723, over 936526.91 frames. ], batch size: 51, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:35:35,657 INFO [finetune.py:976] (3/7) Epoch 15, batch 850, loss[loss=0.1872, simple_loss=0.2554, pruned_loss=0.05949, over 4824.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2517, pruned_loss=0.05638, over 942073.30 frames. ], batch size: 39, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:35:46,470 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.971e+01 1.581e+02 1.945e+02 2.309e+02 3.593e+02, threshold=3.889e+02, percent-clipped=0.0 +2023-04-27 06:35:47,815 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81049.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:36:41,754 INFO [finetune.py:976] (3/7) Epoch 15, batch 900, loss[loss=0.1853, simple_loss=0.2627, pruned_loss=0.05393, over 4751.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2496, pruned_loss=0.0558, over 946130.03 frames. ], batch size: 26, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:36:49,895 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81092.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:36:52,938 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:15,239 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:22,517 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81119.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:34,999 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:48,250 INFO [finetune.py:976] (3/7) Epoch 15, batch 950, loss[loss=0.1553, simple_loss=0.2307, pruned_loss=0.03991, over 4907.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2479, pruned_loss=0.05535, over 947796.80 frames. ], batch size: 37, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:38:04,672 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.625e+02 2.019e+02 2.368e+02 4.092e+02, threshold=4.038e+02, percent-clipped=2.0 +2023-04-27 06:38:05,386 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.4900, 4.2685, 3.0455, 5.0475, 4.3905, 4.3914, 1.7654, 4.3337], + device='cuda:3'), covar=tensor([0.1563, 0.1134, 0.3553, 0.1000, 0.3691, 0.1656, 0.6020, 0.2318], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0214, 0.0248, 0.0301, 0.0296, 0.0246, 0.0269, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 06:38:06,659 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81150.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:38:17,045 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81167.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:38:23,013 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1619, 2.2092, 1.7892, 1.8925, 2.4199, 1.9256, 2.8273, 1.4426], + device='cuda:3'), covar=tensor([0.3529, 0.1873, 0.4884, 0.2643, 0.1464, 0.2294, 0.1184, 0.5409], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0349, 0.0429, 0.0355, 0.0384, 0.0384, 0.0373, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:38:23,609 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81177.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:38:30,448 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4493, 1.4698, 3.8043, 3.5181, 3.3918, 3.7055, 3.6842, 3.3675], + device='cuda:3'), covar=tensor([0.6748, 0.5452, 0.1210, 0.1962, 0.1208, 0.1434, 0.1284, 0.1479], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0306, 0.0400, 0.0404, 0.0348, 0.0404, 0.0314, 0.0360], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:38:32,221 INFO [finetune.py:976] (3/7) Epoch 15, batch 1000, loss[loss=0.1839, simple_loss=0.258, pruned_loss=0.05493, over 4105.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2498, pruned_loss=0.05633, over 946250.62 frames. ], batch size: 66, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:38:38,869 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81198.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:39:05,803 INFO [finetune.py:976] (3/7) Epoch 15, batch 1050, loss[loss=0.1619, simple_loss=0.2299, pruned_loss=0.04694, over 4784.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2512, pruned_loss=0.05573, over 949417.51 frames. ], batch size: 29, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:39:11,772 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.267e+02 1.726e+02 1.946e+02 2.343e+02 3.308e+02, threshold=3.891e+02, percent-clipped=0.0 +2023-04-27 06:39:38,630 INFO [finetune.py:976] (3/7) Epoch 15, batch 1100, loss[loss=0.2272, simple_loss=0.2907, pruned_loss=0.08188, over 4811.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2517, pruned_loss=0.05568, over 951669.01 frames. ], batch size: 40, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:39:46,744 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:39:47,952 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81301.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:40:00,289 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 06:40:11,235 INFO [finetune.py:976] (3/7) Epoch 15, batch 1150, loss[loss=0.201, simple_loss=0.2668, pruned_loss=0.06764, over 4734.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2525, pruned_loss=0.05561, over 952084.57 frames. ], batch size: 54, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:40:18,548 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.629e+02 1.924e+02 2.399e+02 3.865e+02, threshold=3.848e+02, percent-clipped=0.0 +2023-04-27 06:40:27,087 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81360.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:40:27,775 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 06:40:28,330 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81362.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:40:44,971 INFO [finetune.py:976] (3/7) Epoch 15, batch 1200, loss[loss=0.1715, simple_loss=0.2522, pruned_loss=0.04533, over 4790.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.251, pruned_loss=0.05532, over 951858.08 frames. ], batch size: 51, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:40:48,475 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81392.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:41:12,546 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81429.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:41:18,366 INFO [finetune.py:976] (3/7) Epoch 15, batch 1250, loss[loss=0.1579, simple_loss=0.2258, pruned_loss=0.04499, over 4683.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2492, pruned_loss=0.05539, over 952038.60 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:41:20,060 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81440.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:41:24,138 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2288, 1.5485, 1.4813, 1.8455, 1.7260, 1.9420, 1.4158, 3.5794], + device='cuda:3'), covar=tensor([0.0626, 0.0782, 0.0795, 0.1115, 0.0604, 0.0470, 0.0730, 0.0145], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 06:41:31,411 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.659e+02 1.910e+02 2.320e+02 4.141e+02, threshold=3.819e+02, percent-clipped=2.0 +2023-04-27 06:41:35,688 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6699, 3.6513, 2.6448, 4.2104, 3.7047, 3.6242, 1.7112, 3.6333], + device='cuda:3'), covar=tensor([0.1781, 0.1162, 0.2988, 0.1838, 0.2671, 0.1789, 0.5503, 0.2339], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0213, 0.0248, 0.0300, 0.0296, 0.0246, 0.0269, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 06:41:52,825 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-27 06:42:05,045 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81472.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:42:08,091 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81477.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:42:20,542 INFO [finetune.py:976] (3/7) Epoch 15, batch 1300, loss[loss=0.1872, simple_loss=0.2552, pruned_loss=0.05961, over 4873.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2475, pruned_loss=0.05522, over 952194.61 frames. ], batch size: 34, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:43:12,244 INFO [finetune.py:976] (3/7) Epoch 15, batch 1350, loss[loss=0.1832, simple_loss=0.2346, pruned_loss=0.06584, over 4703.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2463, pruned_loss=0.05467, over 954892.71 frames. ], batch size: 23, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:43:24,620 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.644e+02 2.016e+02 2.436e+02 4.078e+02, threshold=4.033e+02, percent-clipped=1.0 +2023-04-27 06:43:42,875 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1001, 0.6624, 0.9101, 0.7112, 1.2156, 0.9730, 0.8223, 0.9450], + device='cuda:3'), covar=tensor([0.1879, 0.1673, 0.2148, 0.1716, 0.1117, 0.1684, 0.2055, 0.2426], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0317, 0.0353, 0.0292, 0.0330, 0.0314, 0.0304, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.4353e-05, 6.6330e-05, 7.5705e-05, 5.9775e-05, 6.8765e-05, 6.6491e-05, + 6.4389e-05, 7.7325e-05], device='cuda:3') +2023-04-27 06:44:07,693 INFO [finetune.py:976] (3/7) Epoch 15, batch 1400, loss[loss=0.2068, simple_loss=0.2799, pruned_loss=0.06688, over 4907.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2495, pruned_loss=0.05522, over 955113.14 frames. ], batch size: 43, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:44:29,735 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.4961, 1.4120, 1.3597, 1.0597, 1.5101, 1.2690, 1.7173, 1.2944], + device='cuda:3'), covar=tensor([0.3517, 0.1572, 0.4698, 0.2392, 0.1280, 0.1868, 0.1619, 0.4344], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0350, 0.0431, 0.0358, 0.0386, 0.0385, 0.0375, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:44:30,989 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8005, 1.8146, 1.8204, 1.4127, 2.0357, 1.5417, 2.5470, 1.5837], + device='cuda:3'), covar=tensor([0.3784, 0.2063, 0.4562, 0.3033, 0.1791, 0.2381, 0.1296, 0.4576], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0350, 0.0431, 0.0358, 0.0386, 0.0385, 0.0375, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:44:33,418 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6547, 1.8404, 1.8492, 2.0252, 2.0619, 2.2451, 1.7652, 3.6790], + device='cuda:3'), covar=tensor([0.0573, 0.0751, 0.0721, 0.1051, 0.0557, 0.0427, 0.0647, 0.0161], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 06:44:41,266 INFO [finetune.py:976] (3/7) Epoch 15, batch 1450, loss[loss=0.2487, simple_loss=0.2897, pruned_loss=0.1039, over 4897.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2514, pruned_loss=0.05604, over 953941.14 frames. ], batch size: 32, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:44:47,200 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.626e+02 2.006e+02 2.422e+02 4.010e+02, threshold=4.013e+02, percent-clipped=0.0 +2023-04-27 06:44:53,616 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81655.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:44:55,302 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81657.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:45:04,217 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81670.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:45:15,083 INFO [finetune.py:976] (3/7) Epoch 15, batch 1500, loss[loss=0.1716, simple_loss=0.2467, pruned_loss=0.0482, over 4905.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2534, pruned_loss=0.05654, over 956241.50 frames. ], batch size: 37, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:45:18,218 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7994, 1.2563, 1.3854, 1.3910, 1.9248, 1.5667, 1.2711, 1.3769], + device='cuda:3'), covar=tensor([0.1734, 0.1584, 0.2062, 0.1504, 0.0940, 0.1410, 0.2009, 0.2023], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0321, 0.0357, 0.0296, 0.0333, 0.0318, 0.0308, 0.0366], + device='cuda:3'), out_proj_covar=tensor([6.5173e-05, 6.7197e-05, 7.6557e-05, 6.0464e-05, 6.9395e-05, 6.7296e-05, + 6.5116e-05, 7.8265e-05], device='cuda:3') +2023-04-27 06:45:44,067 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81731.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:45:48,177 INFO [finetune.py:976] (3/7) Epoch 15, batch 1550, loss[loss=0.1824, simple_loss=0.2554, pruned_loss=0.05468, over 4924.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2527, pruned_loss=0.05612, over 955184.40 frames. ], batch size: 38, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:45:52,553 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2850, 1.6550, 1.4492, 1.8592, 1.7393, 1.9917, 1.4689, 3.6805], + device='cuda:3'), covar=tensor([0.0615, 0.0776, 0.0801, 0.1180, 0.0623, 0.0468, 0.0751, 0.0135], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 06:45:53,673 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.553e+02 1.872e+02 2.288e+02 6.577e+02, threshold=3.744e+02, percent-clipped=2.0 +2023-04-27 06:46:11,812 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81772.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:46:21,441 INFO [finetune.py:976] (3/7) Epoch 15, batch 1600, loss[loss=0.19, simple_loss=0.2491, pruned_loss=0.06545, over 4823.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2504, pruned_loss=0.05541, over 956818.36 frames. ], batch size: 25, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:46:42,457 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:46:48,550 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9379, 1.4130, 1.7875, 1.7398, 1.7180, 1.3926, 0.7904, 1.4140], + device='cuda:3'), covar=tensor([0.3326, 0.3389, 0.1623, 0.2357, 0.2404, 0.2657, 0.4366, 0.2163], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0245, 0.0222, 0.0315, 0.0214, 0.0229, 0.0229, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 06:46:53,851 INFO [finetune.py:976] (3/7) Epoch 15, batch 1650, loss[loss=0.1611, simple_loss=0.2302, pruned_loss=0.04599, over 4921.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2483, pruned_loss=0.05469, over 956485.19 frames. ], batch size: 36, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:47:04,628 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.625e+02 1.858e+02 2.341e+02 4.893e+02, threshold=3.717e+02, percent-clipped=2.0 +2023-04-27 06:47:18,709 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-04-27 06:47:59,312 INFO [finetune.py:976] (3/7) Epoch 15, batch 1700, loss[loss=0.1926, simple_loss=0.2579, pruned_loss=0.06371, over 4937.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2466, pruned_loss=0.05444, over 958754.83 frames. ], batch size: 33, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:48:21,918 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5893, 1.4596, 0.6692, 1.2614, 1.6105, 1.4439, 1.3518, 1.3889], + device='cuda:3'), covar=tensor([0.0500, 0.0390, 0.0391, 0.0574, 0.0294, 0.0490, 0.0519, 0.0578], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 06:48:24,125 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-04-27 06:48:42,056 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3351, 3.2326, 2.5606, 3.7965, 3.3038, 3.2541, 1.3534, 3.2251], + device='cuda:3'), covar=tensor([0.1776, 0.1456, 0.3210, 0.2183, 0.2885, 0.1959, 0.5733, 0.2611], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0215, 0.0250, 0.0302, 0.0297, 0.0248, 0.0270, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 06:49:06,083 INFO [finetune.py:976] (3/7) Epoch 15, batch 1750, loss[loss=0.1882, simple_loss=0.2524, pruned_loss=0.06199, over 4739.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2494, pruned_loss=0.05568, over 957155.48 frames. ], batch size: 23, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:49:16,875 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.080e+01 1.647e+02 2.011e+02 2.347e+02 5.729e+02, threshold=4.021e+02, percent-clipped=2.0 +2023-04-27 06:49:26,288 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81955.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:49:26,959 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7775, 1.8301, 1.7942, 1.4236, 1.9776, 1.5180, 2.6019, 1.6538], + device='cuda:3'), covar=tensor([0.3827, 0.1783, 0.4620, 0.2858, 0.1538, 0.2497, 0.1302, 0.3730], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0348, 0.0427, 0.0354, 0.0383, 0.0382, 0.0373, 0.0416], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:49:27,515 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81957.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:49:48,438 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:49:48,926 INFO [finetune.py:976] (3/7) Epoch 15, batch 1800, loss[loss=0.1787, simple_loss=0.254, pruned_loss=0.05169, over 4827.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2508, pruned_loss=0.05578, over 953164.44 frames. ], batch size: 33, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:49:59,486 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82003.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:00,738 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82005.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:16,423 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82026.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:50:24,047 INFO [finetune.py:976] (3/7) Epoch 15, batch 1850, loss[loss=0.1708, simple_loss=0.2404, pruned_loss=0.05058, over 4755.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2526, pruned_loss=0.05629, over 953096.01 frames. ], batch size: 27, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:50:29,487 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.760e+02 1.989e+02 2.247e+02 4.902e+02, threshold=3.979e+02, percent-clipped=1.0 +2023-04-27 06:50:30,223 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82048.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:34,550 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:56,616 INFO [finetune.py:976] (3/7) Epoch 15, batch 1900, loss[loss=0.2001, simple_loss=0.261, pruned_loss=0.0696, over 4820.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2536, pruned_loss=0.05649, over 953965.97 frames. ], batch size: 49, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:51:14,319 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:51:22,446 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-04-27 06:51:30,519 INFO [finetune.py:976] (3/7) Epoch 15, batch 1950, loss[loss=0.1934, simple_loss=0.2638, pruned_loss=0.06148, over 4722.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2524, pruned_loss=0.05615, over 954048.93 frames. ], batch size: 59, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:51:36,461 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.535e+02 1.831e+02 2.341e+02 3.747e+02, threshold=3.661e+02, percent-clipped=0.0 +2023-04-27 06:51:41,705 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 06:51:57,298 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-04-27 06:52:04,204 INFO [finetune.py:976] (3/7) Epoch 15, batch 2000, loss[loss=0.2068, simple_loss=0.2646, pruned_loss=0.07446, over 4816.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2509, pruned_loss=0.05614, over 956329.11 frames. ], batch size: 41, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:52:04,514 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-27 06:52:33,079 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 06:52:37,843 INFO [finetune.py:976] (3/7) Epoch 15, batch 2050, loss[loss=0.1561, simple_loss=0.2386, pruned_loss=0.03679, over 4817.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2491, pruned_loss=0.05552, over 956823.13 frames. ], batch size: 40, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:52:43,310 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.629e+02 1.931e+02 2.371e+02 3.947e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-04-27 06:53:21,150 INFO [finetune.py:976] (3/7) Epoch 15, batch 2100, loss[loss=0.2541, simple_loss=0.3102, pruned_loss=0.09897, over 4916.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2484, pruned_loss=0.05556, over 955784.20 frames. ], batch size: 37, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:54:04,940 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82326.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:54:13,974 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8403, 1.5334, 1.9983, 2.1829, 1.8756, 1.7244, 1.8782, 1.8178], + device='cuda:3'), covar=tensor([0.5528, 0.7709, 0.7932, 0.6925, 0.6813, 0.9643, 0.9880, 1.0935], + device='cuda:3'), in_proj_covar=tensor([0.0417, 0.0407, 0.0498, 0.0511, 0.0445, 0.0468, 0.0476, 0.0479], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:54:24,280 INFO [finetune.py:976] (3/7) Epoch 15, batch 2150, loss[loss=0.1784, simple_loss=0.2558, pruned_loss=0.05051, over 4835.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2509, pruned_loss=0.05628, over 951522.65 frames. ], batch size: 30, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:54:33,153 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82343.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:54:35,053 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82346.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:54:35,535 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.316e+02 1.732e+02 1.970e+02 2.465e+02 9.359e+02, threshold=3.940e+02, percent-clipped=1.0 +2023-04-27 06:54:37,279 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 06:54:46,134 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4419, 2.8633, 1.1653, 1.7992, 1.7945, 2.2987, 1.7979, 1.2342], + device='cuda:3'), covar=tensor([0.1208, 0.0808, 0.1605, 0.1066, 0.0990, 0.0809, 0.1392, 0.1709], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0246, 0.0137, 0.0121, 0.0132, 0.0154, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:54:58,342 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:55:13,104 INFO [finetune.py:976] (3/7) Epoch 15, batch 2200, loss[loss=0.1899, simple_loss=0.2596, pruned_loss=0.0601, over 4891.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.253, pruned_loss=0.05713, over 950171.92 frames. ], batch size: 35, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:55:42,647 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82407.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:55:45,045 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82411.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:55:54,019 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4091, 1.9283, 2.3615, 2.9553, 2.3326, 1.8363, 1.6628, 2.3804], + device='cuda:3'), covar=tensor([0.3352, 0.3129, 0.1585, 0.2791, 0.2684, 0.2712, 0.4178, 0.2157], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0245, 0.0223, 0.0315, 0.0214, 0.0228, 0.0229, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 06:56:12,256 INFO [finetune.py:976] (3/7) Epoch 15, batch 2250, loss[loss=0.1635, simple_loss=0.2324, pruned_loss=0.04733, over 4795.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2536, pruned_loss=0.05699, over 950964.79 frames. ], batch size: 26, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:56:19,194 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.596e+02 1.887e+02 2.124e+02 5.729e+02, threshold=3.773e+02, percent-clipped=1.0 +2023-04-27 06:56:28,079 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8590, 2.4215, 1.9934, 2.3245, 1.6202, 2.0683, 1.9902, 1.5054], + device='cuda:3'), covar=tensor([0.2057, 0.1145, 0.0900, 0.1124, 0.3449, 0.1156, 0.2150, 0.2614], + device='cuda:3'), in_proj_covar=tensor([0.0289, 0.0311, 0.0223, 0.0284, 0.0313, 0.0266, 0.0255, 0.0269], + device='cuda:3'), out_proj_covar=tensor([1.1637e-04, 1.2348e-04, 8.8891e-05, 1.1298e-04, 1.2754e-04, 1.0607e-04, + 1.0326e-04, 1.0707e-04], device='cuda:3') +2023-04-27 06:56:43,953 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.40 vs. limit=5.0 +2023-04-27 06:56:45,467 INFO [finetune.py:976] (3/7) Epoch 15, batch 2300, loss[loss=0.1823, simple_loss=0.2447, pruned_loss=0.05993, over 4899.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2529, pruned_loss=0.05621, over 951773.96 frames. ], batch size: 35, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:56:49,017 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82493.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:57:00,257 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6448, 1.3626, 1.2707, 1.3630, 1.8512, 1.4882, 1.2164, 1.2468], + device='cuda:3'), covar=tensor([0.1409, 0.1114, 0.1554, 0.1295, 0.0676, 0.1175, 0.1745, 0.1873], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0318, 0.0353, 0.0293, 0.0331, 0.0315, 0.0304, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.4100e-05, 6.6549e-05, 7.5616e-05, 5.9970e-05, 6.8880e-05, 6.6744e-05, + 6.4416e-05, 7.7380e-05], device='cuda:3') +2023-04-27 06:57:12,847 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0322, 2.0001, 1.8026, 1.7202, 2.0936, 1.8163, 2.6068, 1.6234], + device='cuda:3'), covar=tensor([0.3858, 0.1944, 0.4703, 0.3065, 0.1907, 0.2406, 0.1427, 0.4397], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0349, 0.0430, 0.0357, 0.0384, 0.0386, 0.0375, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:57:18,140 INFO [finetune.py:976] (3/7) Epoch 15, batch 2350, loss[loss=0.1423, simple_loss=0.2063, pruned_loss=0.0391, over 4761.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2516, pruned_loss=0.05629, over 954371.24 frames. ], batch size: 28, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:57:25,043 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.877e+01 1.633e+02 1.887e+02 2.224e+02 3.641e+02, threshold=3.774e+02, percent-clipped=0.0 +2023-04-27 06:57:30,341 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82554.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 06:57:38,097 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-04-27 06:57:39,322 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 06:57:40,473 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82569.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:57:51,903 INFO [finetune.py:976] (3/7) Epoch 15, batch 2400, loss[loss=0.1608, simple_loss=0.2283, pruned_loss=0.04659, over 4913.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2494, pruned_loss=0.05558, over 956536.43 frames. ], batch size: 36, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:57:52,672 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2699, 1.9118, 2.2802, 2.5870, 2.2166, 1.8467, 1.5305, 2.0440], + device='cuda:3'), covar=tensor([0.3724, 0.3365, 0.1683, 0.2382, 0.2738, 0.2717, 0.4362, 0.2154], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0249, 0.0225, 0.0318, 0.0217, 0.0231, 0.0232, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 06:58:17,081 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 06:58:21,019 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82630.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:58:21,593 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5362, 3.4892, 1.1736, 1.9580, 1.9400, 2.6363, 1.9147, 0.9899], + device='cuda:3'), covar=tensor([0.1249, 0.0911, 0.1693, 0.1198, 0.0977, 0.0885, 0.1498, 0.1978], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0246, 0.0137, 0.0121, 0.0131, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 06:58:25,773 INFO [finetune.py:976] (3/7) Epoch 15, batch 2450, loss[loss=0.199, simple_loss=0.268, pruned_loss=0.065, over 4863.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.246, pruned_loss=0.05443, over 952463.45 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 06:58:28,867 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:58:31,168 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.750e+01 1.533e+02 1.809e+02 2.233e+02 3.484e+02, threshold=3.618e+02, percent-clipped=0.0 +2023-04-27 06:58:51,161 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1002, 1.1986, 2.9351, 2.5933, 2.6958, 2.6743, 2.7632, 2.5575], + device='cuda:3'), covar=tensor([0.8721, 0.6230, 0.2534, 0.3675, 0.2492, 0.4058, 0.6289, 0.3619], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0305, 0.0400, 0.0403, 0.0345, 0.0403, 0.0311, 0.0360], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:59:01,878 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 06:59:02,103 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0741, 1.9912, 1.6307, 1.7051, 1.9823, 1.7197, 2.5072, 1.4317], + device='cuda:3'), covar=tensor([0.3998, 0.1846, 0.5472, 0.3182, 0.1895, 0.2600, 0.1548, 0.5312], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0348, 0.0430, 0.0356, 0.0384, 0.0385, 0.0373, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 06:59:04,415 INFO [finetune.py:976] (3/7) Epoch 15, batch 2500, loss[loss=0.2083, simple_loss=0.2769, pruned_loss=0.06981, over 4845.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2474, pruned_loss=0.05485, over 953083.04 frames. ], batch size: 49, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 06:59:11,878 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82691.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:59:24,993 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82702.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:59:38,022 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82711.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:00:09,105 INFO [finetune.py:976] (3/7) Epoch 15, batch 2550, loss[loss=0.2597, simple_loss=0.3107, pruned_loss=0.1043, over 4908.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2519, pruned_loss=0.05663, over 954805.88 frames. ], batch size: 36, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:00:14,537 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.661e+02 1.980e+02 2.344e+02 4.065e+02, threshold=3.961e+02, percent-clipped=3.0 +2023-04-27 07:00:24,402 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82759.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:00:47,932 INFO [finetune.py:976] (3/7) Epoch 15, batch 2600, loss[loss=0.2029, simple_loss=0.2687, pruned_loss=0.06857, over 4886.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2519, pruned_loss=0.05648, over 951377.41 frames. ], batch size: 32, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:01:27,753 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6891, 4.6506, 1.3784, 3.0125, 3.1801, 3.6082, 2.9517, 1.5028], + device='cuda:3'), covar=tensor([0.0917, 0.0754, 0.1812, 0.0877, 0.0659, 0.0773, 0.1219, 0.1718], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0245, 0.0136, 0.0120, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:01:43,236 INFO [finetune.py:976] (3/7) Epoch 15, batch 2650, loss[loss=0.2075, simple_loss=0.279, pruned_loss=0.06804, over 4787.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2535, pruned_loss=0.05728, over 949418.88 frames. ], batch size: 45, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:01:48,663 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.684e+02 1.897e+02 2.238e+02 3.554e+02, threshold=3.793e+02, percent-clipped=0.0 +2023-04-27 07:01:49,963 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82849.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:02:33,356 INFO [finetune.py:976] (3/7) Epoch 15, batch 2700, loss[loss=0.15, simple_loss=0.2142, pruned_loss=0.04286, over 4345.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2526, pruned_loss=0.05631, over 951136.44 frames. ], batch size: 18, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:03:04,631 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:03:12,571 INFO [finetune.py:976] (3/7) Epoch 15, batch 2750, loss[loss=0.1562, simple_loss=0.2378, pruned_loss=0.03729, over 4775.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2491, pruned_loss=0.05489, over 953306.98 frames. ], batch size: 28, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:03:18,528 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.634e+02 2.005e+02 2.343e+02 4.529e+02, threshold=4.010e+02, percent-clipped=3.0 +2023-04-27 07:03:28,544 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82963.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:03:44,345 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4193, 3.3966, 0.8681, 1.8214, 1.9023, 2.5267, 1.8528, 1.0059], + device='cuda:3'), covar=tensor([0.1417, 0.0855, 0.2148, 0.1264, 0.1035, 0.0945, 0.1540, 0.1944], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0246, 0.0137, 0.0121, 0.0131, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:03:46,114 INFO [finetune.py:976] (3/7) Epoch 15, batch 2800, loss[loss=0.1972, simple_loss=0.2568, pruned_loss=0.06878, over 4869.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2454, pruned_loss=0.05378, over 952578.56 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:03:55,381 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83002.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:04:09,812 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83024.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:04:19,800 INFO [finetune.py:976] (3/7) Epoch 15, batch 2850, loss[loss=0.1486, simple_loss=0.2128, pruned_loss=0.04223, over 4738.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2448, pruned_loss=0.05373, over 955346.54 frames. ], batch size: 27, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:04:25,721 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.603e+01 1.615e+02 1.794e+02 2.149e+02 4.463e+02, threshold=3.588e+02, percent-clipped=2.0 +2023-04-27 07:04:27,624 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83050.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:04:53,189 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4103, 1.6712, 1.6395, 2.2493, 2.3826, 1.9181, 1.8930, 1.6475], + device='cuda:3'), covar=tensor([0.1863, 0.1749, 0.2093, 0.1235, 0.1154, 0.1873, 0.2031, 0.2100], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0317, 0.0353, 0.0292, 0.0331, 0.0314, 0.0305, 0.0363], + device='cuda:3'), out_proj_covar=tensor([6.3995e-05, 6.6366e-05, 7.5657e-05, 5.9682e-05, 6.8887e-05, 6.6489e-05, + 6.4621e-05, 7.7424e-05], device='cuda:3') +2023-04-27 07:05:03,760 INFO [finetune.py:976] (3/7) Epoch 15, batch 2900, loss[loss=0.1398, simple_loss=0.219, pruned_loss=0.0303, over 4748.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2487, pruned_loss=0.05504, over 956569.00 frames. ], batch size: 27, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:05:17,382 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4643, 1.3877, 4.2292, 3.9856, 3.7804, 4.0490, 4.0200, 3.7391], + device='cuda:3'), covar=tensor([0.6734, 0.5653, 0.0925, 0.1523, 0.0922, 0.1608, 0.0974, 0.1225], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0308, 0.0401, 0.0405, 0.0348, 0.0404, 0.0312, 0.0360], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:05:27,767 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 07:06:06,734 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 07:06:07,041 INFO [finetune.py:976] (3/7) Epoch 15, batch 2950, loss[loss=0.1896, simple_loss=0.2691, pruned_loss=0.05508, over 4841.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2503, pruned_loss=0.05565, over 955141.66 frames. ], batch size: 47, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:06:16,724 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 07:06:18,120 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.270e+02 1.793e+02 2.066e+02 2.627e+02 6.571e+02, threshold=4.132e+02, percent-clipped=5.0 +2023-04-27 07:06:19,455 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83149.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:06:47,867 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 07:06:56,295 INFO [finetune.py:976] (3/7) Epoch 15, batch 3000, loss[loss=0.1146, simple_loss=0.176, pruned_loss=0.0266, over 3994.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2524, pruned_loss=0.05685, over 954256.33 frames. ], batch size: 17, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:06:56,296 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 07:07:05,472 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9312, 2.2423, 1.8403, 1.5691, 1.4882, 1.5180, 1.8361, 1.4239], + device='cuda:3'), covar=tensor([0.1714, 0.1453, 0.1460, 0.1789, 0.2339, 0.1907, 0.1045, 0.2075], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0214, 0.0169, 0.0205, 0.0201, 0.0184, 0.0156, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 07:07:06,878 INFO [finetune.py:1010] (3/7) Epoch 15, validation: loss=0.1516, simple_loss=0.2237, pruned_loss=0.03975, over 2265189.00 frames. +2023-04-27 07:07:06,879 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 07:07:12,345 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83196.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:07:12,903 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83197.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:07:15,782 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-27 07:07:30,743 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:07:38,609 INFO [finetune.py:976] (3/7) Epoch 15, batch 3050, loss[loss=0.1752, simple_loss=0.2505, pruned_loss=0.04999, over 4859.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2519, pruned_loss=0.0559, over 953503.84 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:07:50,803 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.173e+02 1.583e+02 1.890e+02 2.106e+02 3.614e+02, threshold=3.781e+02, percent-clipped=0.0 +2023-04-27 07:08:02,509 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83257.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:08:24,058 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:08:36,018 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83284.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:08:42,493 INFO [finetune.py:976] (3/7) Epoch 15, batch 3100, loss[loss=0.1823, simple_loss=0.2403, pruned_loss=0.06219, over 4889.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2506, pruned_loss=0.0553, over 954630.85 frames. ], batch size: 32, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:09:08,143 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-04-27 07:09:15,039 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:09:26,584 INFO [finetune.py:976] (3/7) Epoch 15, batch 3150, loss[loss=0.1798, simple_loss=0.2495, pruned_loss=0.05511, over 4924.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2488, pruned_loss=0.05484, over 954897.82 frames. ], batch size: 37, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:09:31,033 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83345.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:09:32,637 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.218e+02 1.641e+02 1.865e+02 2.202e+02 3.570e+02, threshold=3.730e+02, percent-clipped=0.0 +2023-04-27 07:09:48,698 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6060, 2.7034, 2.3677, 2.3870, 2.7225, 2.6008, 3.7522, 2.0664], + device='cuda:3'), covar=tensor([0.4037, 0.2207, 0.4551, 0.3626, 0.2045, 0.2525, 0.1411, 0.4428], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0349, 0.0429, 0.0357, 0.0386, 0.0384, 0.0374, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:09:59,878 INFO [finetune.py:976] (3/7) Epoch 15, batch 3200, loss[loss=0.1635, simple_loss=0.2354, pruned_loss=0.04575, over 4776.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2447, pruned_loss=0.05309, over 954022.32 frames. ], batch size: 26, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:10:45,248 INFO [finetune.py:976] (3/7) Epoch 15, batch 3250, loss[loss=0.1815, simple_loss=0.2532, pruned_loss=0.05492, over 4924.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2455, pruned_loss=0.05371, over 952192.14 frames. ], batch size: 38, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:10:56,131 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.623e+02 1.932e+02 2.416e+02 4.902e+02, threshold=3.863e+02, percent-clipped=4.0 +2023-04-27 07:11:08,267 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6456, 3.5467, 2.5470, 4.2196, 3.6442, 3.6738, 1.5851, 3.6492], + device='cuda:3'), covar=tensor([0.1761, 0.1450, 0.3221, 0.1901, 0.4225, 0.2007, 0.5794, 0.2455], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0213, 0.0248, 0.0300, 0.0295, 0.0245, 0.0267, 0.0268], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:11:17,952 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 07:11:18,757 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8084, 2.5273, 1.8569, 1.8435, 1.3081, 1.3582, 1.9928, 1.2759], + device='cuda:3'), covar=tensor([0.1651, 0.1290, 0.1437, 0.1638, 0.2372, 0.1975, 0.0943, 0.2093], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0215, 0.0169, 0.0206, 0.0202, 0.0185, 0.0158, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 07:11:50,146 INFO [finetune.py:976] (3/7) Epoch 15, batch 3300, loss[loss=0.1411, simple_loss=0.213, pruned_loss=0.03466, over 4829.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2482, pruned_loss=0.05466, over 952938.74 frames. ], batch size: 25, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:12:17,739 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5659, 3.5665, 2.5512, 4.1570, 3.6328, 3.5470, 1.5182, 3.5837], + device='cuda:3'), covar=tensor([0.1828, 0.1341, 0.3376, 0.1837, 0.3011, 0.1900, 0.6215, 0.2624], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0215, 0.0251, 0.0303, 0.0298, 0.0248, 0.0270, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:12:20,672 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2525, 1.4924, 1.3445, 1.7203, 1.6561, 1.8184, 1.3902, 3.3774], + device='cuda:3'), covar=tensor([0.0606, 0.0814, 0.0837, 0.1245, 0.0653, 0.0518, 0.0771, 0.0149], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 07:12:23,029 INFO [finetune.py:976] (3/7) Epoch 15, batch 3350, loss[loss=0.1765, simple_loss=0.2489, pruned_loss=0.05211, over 4806.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2514, pruned_loss=0.05621, over 953413.93 frames. ], batch size: 25, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:12:28,445 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.803e+02 2.175e+02 2.734e+02 4.804e+02, threshold=4.350e+02, percent-clipped=5.0 +2023-04-27 07:12:31,592 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83552.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:12:45,370 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0393, 2.5831, 0.9044, 1.2972, 1.9092, 1.2079, 3.4248, 1.6738], + device='cuda:3'), covar=tensor([0.0677, 0.0736, 0.0928, 0.1270, 0.0532, 0.1033, 0.0218, 0.0631], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0051, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 07:12:52,060 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0685, 2.5662, 1.0486, 1.3433, 2.0396, 1.2261, 3.3486, 1.6590], + device='cuda:3'), covar=tensor([0.0664, 0.0855, 0.0904, 0.1230, 0.0479, 0.1021, 0.0194, 0.0621], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0051, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 07:12:56,720 INFO [finetune.py:976] (3/7) Epoch 15, batch 3400, loss[loss=0.2197, simple_loss=0.2899, pruned_loss=0.07471, over 4840.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2523, pruned_loss=0.05659, over 951962.56 frames. ], batch size: 49, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:12:59,285 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7494, 1.0592, 1.2577, 1.3896, 1.7514, 1.4072, 1.2239, 1.1923], + device='cuda:3'), covar=tensor([0.1390, 0.1856, 0.2164, 0.1503, 0.1145, 0.1726, 0.2119, 0.2442], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0319, 0.0354, 0.0295, 0.0333, 0.0315, 0.0307, 0.0365], + device='cuda:3'), out_proj_covar=tensor([6.4136e-05, 6.6648e-05, 7.5716e-05, 6.0329e-05, 6.9430e-05, 6.6619e-05, + 6.4942e-05, 7.7934e-05], device='cuda:3') +2023-04-27 07:13:18,199 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83619.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:13:30,155 INFO [finetune.py:976] (3/7) Epoch 15, batch 3450, loss[loss=0.1984, simple_loss=0.2678, pruned_loss=0.06452, over 4745.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2529, pruned_loss=0.05679, over 952784.93 frames. ], batch size: 54, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:13:31,412 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83640.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:13:35,582 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.561e+02 1.911e+02 2.345e+02 4.135e+02, threshold=3.821e+02, percent-clipped=0.0 +2023-04-27 07:13:49,254 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83667.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:14:13,519 INFO [finetune.py:976] (3/7) Epoch 15, batch 3500, loss[loss=0.2224, simple_loss=0.2715, pruned_loss=0.08659, over 4841.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2515, pruned_loss=0.05672, over 953379.57 frames. ], batch size: 44, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:14:33,633 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 07:15:15,482 INFO [finetune.py:976] (3/7) Epoch 15, batch 3550, loss[loss=0.1583, simple_loss=0.2305, pruned_loss=0.04307, over 4895.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2477, pruned_loss=0.05564, over 954447.12 frames. ], batch size: 43, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:15:21,401 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.464e+02 1.732e+02 2.080e+02 4.644e+02, threshold=3.463e+02, percent-clipped=1.0 +2023-04-27 07:15:29,357 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83760.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:15:32,423 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4350, 2.9760, 0.7769, 1.5448, 2.3018, 1.5202, 4.4240, 2.2237], + device='cuda:3'), covar=tensor([0.0687, 0.0797, 0.0946, 0.1368, 0.0573, 0.1047, 0.0246, 0.0636], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0051, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 07:15:32,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9706, 2.2168, 2.1356, 2.3022, 2.0644, 2.3027, 2.1777, 2.1628], + device='cuda:3'), covar=tensor([0.4283, 0.6631, 0.5478, 0.5200, 0.6313, 0.7743, 0.7096, 0.6862], + device='cuda:3'), in_proj_covar=tensor([0.0327, 0.0371, 0.0317, 0.0331, 0.0341, 0.0396, 0.0352, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:15:44,582 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 07:15:49,272 INFO [finetune.py:976] (3/7) Epoch 15, batch 3600, loss[loss=0.1629, simple_loss=0.2406, pruned_loss=0.04258, over 4714.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2456, pruned_loss=0.05466, over 955276.92 frames. ], batch size: 59, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:16:32,027 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:16:55,918 INFO [finetune.py:976] (3/7) Epoch 15, batch 3650, loss[loss=0.2022, simple_loss=0.2754, pruned_loss=0.06446, over 4876.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2477, pruned_loss=0.0552, over 956554.63 frames. ], batch size: 34, lr: 3.48e-03, grad_scale: 64.0 +2023-04-27 07:16:57,942 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8549, 1.5584, 2.0280, 2.2933, 1.8986, 1.7847, 1.9177, 1.9285], + device='cuda:3'), covar=tensor([0.5103, 0.7716, 0.7569, 0.6248, 0.6620, 0.9066, 0.9596, 1.0109], + device='cuda:3'), in_proj_covar=tensor([0.0414, 0.0405, 0.0493, 0.0505, 0.0444, 0.0465, 0.0473, 0.0477], + device='cuda:3'), out_proj_covar=tensor([9.9963e-05, 1.0014e-04, 1.1098e-04, 1.2026e-04, 1.0675e-04, 1.1212e-04, + 1.1263e-04, 1.1335e-04], device='cuda:3') +2023-04-27 07:17:01,428 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.661e+01 1.633e+02 1.914e+02 2.329e+02 4.921e+02, threshold=3.827e+02, percent-clipped=4.0 +2023-04-27 07:17:05,011 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:17:29,863 INFO [finetune.py:976] (3/7) Epoch 15, batch 3700, loss[loss=0.1727, simple_loss=0.2546, pruned_loss=0.04536, over 4923.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2518, pruned_loss=0.05588, over 954681.95 frames. ], batch size: 38, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:17:31,758 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.9586, 3.9353, 2.8003, 4.5456, 3.8886, 3.9483, 1.7392, 3.8513], + device='cuda:3'), covar=tensor([0.1715, 0.1118, 0.3272, 0.1520, 0.3236, 0.1728, 0.5527, 0.2409], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0215, 0.0251, 0.0303, 0.0299, 0.0247, 0.0270, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:17:37,201 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83900.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:18:03,517 INFO [finetune.py:976] (3/7) Epoch 15, batch 3750, loss[loss=0.2284, simple_loss=0.2913, pruned_loss=0.08276, over 4925.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2535, pruned_loss=0.05686, over 955772.79 frames. ], batch size: 42, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:18:04,866 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83940.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:18:09,569 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.640e+02 2.069e+02 2.492e+02 6.161e+02, threshold=4.139e+02, percent-clipped=2.0 +2023-04-27 07:18:10,304 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83949.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:18:16,900 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2940, 1.6598, 1.5946, 2.0442, 1.8337, 1.9238, 1.5912, 4.1890], + device='cuda:3'), covar=tensor([0.0546, 0.0738, 0.0754, 0.1065, 0.0612, 0.0586, 0.0691, 0.0110], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 07:18:16,941 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4948, 1.8879, 2.3204, 2.9824, 2.4114, 1.8693, 1.8427, 2.2381], + device='cuda:3'), covar=tensor([0.3125, 0.3203, 0.1594, 0.2548, 0.2645, 0.2619, 0.3861, 0.2161], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0249, 0.0226, 0.0319, 0.0217, 0.0230, 0.0231, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 07:18:27,274 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0474, 1.7295, 2.2167, 2.5908, 2.0828, 1.9276, 2.0834, 2.0095], + device='cuda:3'), covar=tensor([0.5095, 0.7566, 0.7983, 0.6077, 0.6176, 0.9375, 0.9348, 1.0391], + device='cuda:3'), in_proj_covar=tensor([0.0415, 0.0407, 0.0494, 0.0508, 0.0446, 0.0467, 0.0475, 0.0479], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:18:31,803 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0847, 1.9814, 2.3552, 2.5294, 1.8347, 1.6184, 1.9661, 1.0952], + device='cuda:3'), covar=tensor([0.0562, 0.0895, 0.0480, 0.0662, 0.0847, 0.1311, 0.0775, 0.0865], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0076, 0.0097, 0.0076, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:18:34,098 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83985.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:18:36,294 INFO [finetune.py:976] (3/7) Epoch 15, batch 3800, loss[loss=0.2541, simple_loss=0.3142, pruned_loss=0.09701, over 4318.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2536, pruned_loss=0.05638, over 955793.03 frames. ], batch size: 66, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:18:36,354 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83988.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:18:36,743 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-04-27 07:18:51,937 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84010.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:19:10,674 INFO [finetune.py:976] (3/7) Epoch 15, batch 3850, loss[loss=0.1442, simple_loss=0.2129, pruned_loss=0.03778, over 4924.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2508, pruned_loss=0.05516, over 954509.58 frames. ], batch size: 46, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:19:21,639 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84046.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:19:22,760 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.590e+02 1.914e+02 2.294e+02 5.295e+02, threshold=3.828e+02, percent-clipped=1.0 +2023-04-27 07:19:31,187 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 07:20:15,363 INFO [finetune.py:976] (3/7) Epoch 15, batch 3900, loss[loss=0.1884, simple_loss=0.2545, pruned_loss=0.06113, over 4849.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2497, pruned_loss=0.05576, over 956023.22 frames. ], batch size: 44, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:20:27,928 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:20:46,488 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.75 vs. limit=5.0 +2023-04-27 07:20:50,572 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:21:20,858 INFO [finetune.py:976] (3/7) Epoch 15, batch 3950, loss[loss=0.1397, simple_loss=0.2157, pruned_loss=0.03188, over 4909.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2465, pruned_loss=0.05462, over 955218.10 frames. ], batch size: 36, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:21:34,731 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.122e+01 1.668e+02 1.985e+02 2.633e+02 4.581e+02, threshold=3.971e+02, percent-clipped=4.0 +2023-04-27 07:21:41,474 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:22:12,111 INFO [finetune.py:976] (3/7) Epoch 15, batch 4000, loss[loss=0.1896, simple_loss=0.2627, pruned_loss=0.05824, over 4926.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2461, pruned_loss=0.05416, over 954851.60 frames. ], batch size: 38, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:22:31,712 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84199.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:23:00,623 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84236.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:23:02,206 INFO [finetune.py:976] (3/7) Epoch 15, batch 4050, loss[loss=0.1566, simple_loss=0.2321, pruned_loss=0.04052, over 4795.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2491, pruned_loss=0.05526, over 952108.96 frames. ], batch size: 29, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:23:09,757 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.868e+01 1.686e+02 2.049e+02 2.449e+02 3.463e+02, threshold=4.099e+02, percent-clipped=0.0 +2023-04-27 07:23:18,191 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:23:19,013 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-04-27 07:23:28,430 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4414, 1.6376, 1.5729, 2.1478, 2.3449, 1.9340, 1.8369, 1.6020], + device='cuda:3'), covar=tensor([0.1611, 0.1992, 0.2163, 0.1956, 0.1364, 0.2072, 0.2365, 0.2404], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0318, 0.0354, 0.0294, 0.0332, 0.0315, 0.0305, 0.0365], + device='cuda:3'), out_proj_covar=tensor([6.4319e-05, 6.6428e-05, 7.5793e-05, 6.0078e-05, 6.9225e-05, 6.6595e-05, + 6.4621e-05, 7.7928e-05], device='cuda:3') +2023-04-27 07:23:32,118 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8100, 1.5318, 1.8235, 2.1138, 2.1412, 1.8256, 1.5565, 2.0231], + device='cuda:3'), covar=tensor([0.0970, 0.1318, 0.0740, 0.0656, 0.0689, 0.0830, 0.0796, 0.0586], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0200, 0.0179, 0.0171, 0.0176, 0.0181, 0.0152, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:23:35,685 INFO [finetune.py:976] (3/7) Epoch 15, batch 4100, loss[loss=0.2319, simple_loss=0.2997, pruned_loss=0.0821, over 4808.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2519, pruned_loss=0.05584, over 954166.65 frames. ], batch size: 40, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:23:42,205 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84297.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:23:43,510 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-04-27 07:23:48,051 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84305.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:24:08,381 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:24:09,501 INFO [finetune.py:976] (3/7) Epoch 15, batch 4150, loss[loss=0.2214, simple_loss=0.2821, pruned_loss=0.08036, over 4897.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2529, pruned_loss=0.05602, over 953015.17 frames. ], batch size: 37, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:24:11,379 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84341.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:24:16,002 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.201e+02 1.658e+02 1.986e+02 2.406e+02 7.031e+02, threshold=3.971e+02, percent-clipped=3.0 +2023-04-27 07:24:36,807 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9567, 2.3434, 2.0363, 2.2031, 1.7139, 2.0064, 2.0800, 1.6882], + device='cuda:3'), covar=tensor([0.1522, 0.0883, 0.0617, 0.0957, 0.2656, 0.0900, 0.1578, 0.1993], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0308, 0.0218, 0.0281, 0.0310, 0.0264, 0.0251, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1546e-04, 1.2242e-04, 8.6898e-05, 1.1163e-04, 1.2609e-04, 1.0529e-04, + 1.0140e-04, 1.0561e-04], device='cuda:3') +2023-04-27 07:24:43,282 INFO [finetune.py:976] (3/7) Epoch 15, batch 4200, loss[loss=0.1796, simple_loss=0.2541, pruned_loss=0.05258, over 4861.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2537, pruned_loss=0.05573, over 953294.91 frames. ], batch size: 31, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:24:49,263 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84397.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:25:00,646 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 07:25:03,292 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84416.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:25:12,455 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1029, 1.2863, 1.1956, 1.6003, 1.4369, 1.4796, 1.2333, 2.4793], + device='cuda:3'), covar=tensor([0.0673, 0.0931, 0.0962, 0.1334, 0.0761, 0.0524, 0.0861, 0.0275], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 07:25:22,955 INFO [finetune.py:976] (3/7) Epoch 15, batch 4250, loss[loss=0.1571, simple_loss=0.2219, pruned_loss=0.04612, over 4910.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2517, pruned_loss=0.05533, over 956033.18 frames. ], batch size: 36, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:25:33,107 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.587e+02 1.989e+02 2.520e+02 5.183e+02, threshold=3.979e+02, percent-clipped=3.0 +2023-04-27 07:25:42,002 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84453.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:25:56,124 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84464.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:26:28,018 INFO [finetune.py:976] (3/7) Epoch 15, batch 4300, loss[loss=0.1636, simple_loss=0.2322, pruned_loss=0.04753, over 4913.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2489, pruned_loss=0.05495, over 956313.81 frames. ], batch size: 37, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:27:19,044 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84525.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:27:32,533 INFO [finetune.py:976] (3/7) Epoch 15, batch 4350, loss[loss=0.1837, simple_loss=0.2498, pruned_loss=0.05884, over 4913.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2464, pruned_loss=0.05452, over 954679.80 frames. ], batch size: 43, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:27:42,400 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4599, 1.4609, 1.7298, 1.7652, 1.3523, 1.1957, 1.4943, 1.0440], + device='cuda:3'), covar=tensor([0.0548, 0.0580, 0.0478, 0.0534, 0.0717, 0.1004, 0.0590, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0076, 0.0096, 0.0075, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:27:44,745 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.607e+02 1.900e+02 2.324e+02 5.232e+02, threshold=3.801e+02, percent-clipped=2.0 +2023-04-27 07:27:54,644 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:26,839 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84586.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:27,951 INFO [finetune.py:976] (3/7) Epoch 15, batch 4400, loss[loss=0.2592, simple_loss=0.3253, pruned_loss=0.0966, over 4737.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2488, pruned_loss=0.05586, over 953132.38 frames. ], batch size: 59, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:28:30,485 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84592.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:37,873 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0644, 0.6252, 0.9095, 0.7798, 1.2013, 0.9716, 0.8145, 0.9318], + device='cuda:3'), covar=tensor([0.1773, 0.1456, 0.1871, 0.1592, 0.0887, 0.1261, 0.1628, 0.2025], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0318, 0.0354, 0.0293, 0.0332, 0.0314, 0.0305, 0.0366], + device='cuda:3'), out_proj_covar=tensor([6.4216e-05, 6.6521e-05, 7.5837e-05, 5.9780e-05, 6.9259e-05, 6.6304e-05, + 6.4446e-05, 7.8219e-05], device='cuda:3') +2023-04-27 07:28:39,072 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84605.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:59,200 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84633.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:02,147 INFO [finetune.py:976] (3/7) Epoch 15, batch 4450, loss[loss=0.2011, simple_loss=0.2827, pruned_loss=0.05975, over 4762.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2517, pruned_loss=0.05649, over 952011.99 frames. ], batch size: 54, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:29:04,111 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84641.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:08,289 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.631e+02 1.866e+02 2.330e+02 4.316e+02, threshold=3.732e+02, percent-clipped=2.0 +2023-04-27 07:29:11,996 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84653.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:36,302 INFO [finetune.py:976] (3/7) Epoch 15, batch 4500, loss[loss=0.1826, simple_loss=0.265, pruned_loss=0.05009, over 4798.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2525, pruned_loss=0.05604, over 953708.10 frames. ], batch size: 45, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:29:37,002 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84689.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:38,815 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:40,088 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84694.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:45,385 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5831, 0.6986, 1.5128, 1.9401, 1.6564, 1.4823, 1.5383, 1.5503], + device='cuda:3'), covar=tensor([0.4414, 0.6630, 0.5848, 0.6147, 0.5873, 0.7262, 0.7490, 0.7441], + device='cuda:3'), in_proj_covar=tensor([0.0416, 0.0407, 0.0493, 0.0507, 0.0446, 0.0467, 0.0476, 0.0478], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:30:09,977 INFO [finetune.py:976] (3/7) Epoch 15, batch 4550, loss[loss=0.1315, simple_loss=0.2013, pruned_loss=0.03085, over 4736.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2529, pruned_loss=0.05647, over 954719.95 frames. ], batch size: 23, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:30:16,091 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.700e+02 1.997e+02 2.528e+02 3.461e+02, threshold=3.994e+02, percent-clipped=0.0 +2023-04-27 07:30:19,276 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84753.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:30:36,253 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9677, 2.3492, 1.2969, 1.5858, 2.3793, 1.8993, 1.7357, 1.8111], + device='cuda:3'), covar=tensor([0.0481, 0.0317, 0.0294, 0.0544, 0.0223, 0.0497, 0.0529, 0.0558], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 07:30:43,751 INFO [finetune.py:976] (3/7) Epoch 15, batch 4600, loss[loss=0.1593, simple_loss=0.2317, pruned_loss=0.04344, over 4843.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2516, pruned_loss=0.05555, over 954961.06 frames. ], batch size: 49, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:30:51,294 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84800.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:30:51,854 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84801.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:31:32,412 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2675, 1.7636, 2.1366, 2.6047, 2.1588, 1.7135, 1.5014, 1.9686], + device='cuda:3'), covar=tensor([0.3068, 0.3080, 0.1584, 0.2171, 0.2632, 0.2473, 0.4447, 0.2052], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0250, 0.0226, 0.0318, 0.0218, 0.0232, 0.0232, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 07:31:33,987 INFO [finetune.py:976] (3/7) Epoch 15, batch 4650, loss[loss=0.1768, simple_loss=0.2467, pruned_loss=0.05349, over 4868.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2494, pruned_loss=0.05545, over 953473.06 frames. ], batch size: 31, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:31:45,944 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.628e+02 1.889e+02 2.261e+02 4.327e+02, threshold=3.778e+02, percent-clipped=1.0 +2023-04-27 07:31:55,726 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84855.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:32:05,698 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84861.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:32:06,279 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2258, 2.0174, 2.3853, 2.6761, 2.6515, 2.2860, 1.9076, 2.3874], + device='cuda:3'), covar=tensor([0.0852, 0.1022, 0.0583, 0.0548, 0.0618, 0.0732, 0.0794, 0.0516], + device='cuda:3'), in_proj_covar=tensor([0.0191, 0.0203, 0.0183, 0.0175, 0.0179, 0.0184, 0.0156, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:32:30,053 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84881.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:32:37,960 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4244, 1.6605, 1.4864, 1.8290, 1.7648, 1.9990, 1.4340, 3.7579], + device='cuda:3'), covar=tensor([0.0573, 0.0761, 0.0817, 0.1122, 0.0622, 0.0484, 0.0699, 0.0119], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 07:32:39,700 INFO [finetune.py:976] (3/7) Epoch 15, batch 4700, loss[loss=0.1713, simple_loss=0.2331, pruned_loss=0.05479, over 4786.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2467, pruned_loss=0.05457, over 955901.79 frames. ], batch size: 26, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:32:48,376 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84892.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:32:59,468 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6027, 1.1710, 1.3490, 1.3884, 1.8030, 1.5169, 1.2600, 1.2889], + device='cuda:3'), covar=tensor([0.1733, 0.1402, 0.1722, 0.1400, 0.0756, 0.1329, 0.1558, 0.1790], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0319, 0.0356, 0.0295, 0.0334, 0.0314, 0.0305, 0.0367], + device='cuda:3'), out_proj_covar=tensor([6.4577e-05, 6.6622e-05, 7.6221e-05, 6.0246e-05, 6.9595e-05, 6.6416e-05, + 6.4482e-05, 7.8314e-05], device='cuda:3') +2023-04-27 07:33:00,658 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84903.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:33:37,581 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1863, 1.6847, 2.0607, 2.3299, 2.0180, 1.5813, 1.2249, 1.7635], + device='cuda:3'), covar=tensor([0.3382, 0.3342, 0.1669, 0.2515, 0.2682, 0.2697, 0.4592, 0.2102], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0248, 0.0225, 0.0316, 0.0217, 0.0230, 0.0231, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 07:33:46,042 INFO [finetune.py:976] (3/7) Epoch 15, batch 4750, loss[loss=0.1793, simple_loss=0.253, pruned_loss=0.05285, over 4808.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2465, pruned_loss=0.0552, over 957854.51 frames. ], batch size: 41, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:33:47,831 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84940.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:33:53,645 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.954e+01 1.625e+02 2.007e+02 2.365e+02 3.914e+02, threshold=4.014e+02, percent-clipped=2.0 +2023-04-27 07:33:55,662 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84951.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:34:05,925 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2787, 1.4520, 5.1479, 4.8612, 4.5109, 4.9491, 4.6066, 4.6317], + device='cuda:3'), covar=tensor([0.5782, 0.5919, 0.1103, 0.1775, 0.1099, 0.1800, 0.1299, 0.1450], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0306, 0.0402, 0.0407, 0.0349, 0.0405, 0.0312, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:34:11,745 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 07:34:19,536 INFO [finetune.py:976] (3/7) Epoch 15, batch 4800, loss[loss=0.1964, simple_loss=0.2666, pruned_loss=0.06313, over 4151.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2498, pruned_loss=0.05635, over 957757.69 frames. ], batch size: 65, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:34:20,212 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:34:23,044 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84992.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:34:33,957 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1241, 2.1674, 1.7051, 1.8505, 2.0367, 1.6926, 2.5619, 1.3351], + device='cuda:3'), covar=tensor([0.3721, 0.1922, 0.5391, 0.3326, 0.2164, 0.2771, 0.1548, 0.5257], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0351, 0.0430, 0.0357, 0.0386, 0.0383, 0.0376, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:34:36,386 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85012.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:34:53,702 INFO [finetune.py:976] (3/7) Epoch 15, batch 4850, loss[loss=0.1743, simple_loss=0.2508, pruned_loss=0.04886, over 4891.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2525, pruned_loss=0.05679, over 956995.54 frames. ], batch size: 43, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:34:54,454 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-27 07:34:54,927 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:35:01,196 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.698e+02 1.957e+02 2.327e+02 4.552e+02, threshold=3.914e+02, percent-clipped=3.0 +2023-04-27 07:35:16,356 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 07:35:27,220 INFO [finetune.py:976] (3/7) Epoch 15, batch 4900, loss[loss=0.2271, simple_loss=0.3036, pruned_loss=0.07534, over 4912.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2533, pruned_loss=0.0569, over 953810.85 frames. ], batch size: 38, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:35:40,343 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85107.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:35:42,720 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5507, 3.2787, 0.9102, 1.7683, 1.8956, 2.4284, 1.8219, 1.1054], + device='cuda:3'), covar=tensor([0.1357, 0.0807, 0.2002, 0.1281, 0.1062, 0.0982, 0.1502, 0.2038], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0242, 0.0136, 0.0120, 0.0130, 0.0151, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:35:44,001 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9291, 1.7173, 2.1801, 2.3919, 1.9999, 1.8786, 2.0117, 1.9736], + device='cuda:3'), covar=tensor([0.4979, 0.7320, 0.7161, 0.6292, 0.6031, 0.8925, 0.9383, 0.9365], + device='cuda:3'), in_proj_covar=tensor([0.0416, 0.0406, 0.0493, 0.0506, 0.0444, 0.0468, 0.0475, 0.0478], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:35:44,605 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5669, 1.5139, 1.8460, 1.8923, 1.3944, 1.3104, 1.6005, 1.0882], + device='cuda:3'), covar=tensor([0.0568, 0.0730, 0.0441, 0.0588, 0.0768, 0.1092, 0.0753, 0.0734], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0070, 0.0069, 0.0067, 0.0075, 0.0096, 0.0075, 0.0068], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:35:45,208 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:36:00,015 INFO [finetune.py:976] (3/7) Epoch 15, batch 4950, loss[loss=0.1745, simple_loss=0.2458, pruned_loss=0.05161, over 4712.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2548, pruned_loss=0.05699, over 952661.70 frames. ], batch size: 54, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:36:07,093 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.205e+02 1.668e+02 1.905e+02 2.304e+02 4.878e+02, threshold=3.810e+02, percent-clipped=1.0 +2023-04-27 07:36:12,985 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85156.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:36:20,317 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85168.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:36:25,197 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:36:28,194 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85181.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:36:33,340 INFO [finetune.py:976] (3/7) Epoch 15, batch 5000, loss[loss=0.2059, simple_loss=0.2714, pruned_loss=0.07017, over 4842.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2537, pruned_loss=0.05629, over 955011.17 frames. ], batch size: 49, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:37:00,637 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:37:12,536 INFO [finetune.py:976] (3/7) Epoch 15, batch 5050, loss[loss=0.1561, simple_loss=0.2261, pruned_loss=0.04301, over 4890.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2504, pruned_loss=0.05536, over 955640.60 frames. ], batch size: 35, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:37:25,139 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.655e+02 2.048e+02 2.440e+02 4.868e+02, threshold=4.096e+02, percent-clipped=4.0 +2023-04-27 07:38:17,994 INFO [finetune.py:976] (3/7) Epoch 15, batch 5100, loss[loss=0.1586, simple_loss=0.2262, pruned_loss=0.04553, over 4820.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2467, pruned_loss=0.05367, over 956388.44 frames. ], batch size: 41, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:38:18,686 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:38:41,086 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85307.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:39:06,777 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:39:07,341 INFO [finetune.py:976] (3/7) Epoch 15, batch 5150, loss[loss=0.1831, simple_loss=0.2436, pruned_loss=0.06129, over 4780.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2467, pruned_loss=0.05417, over 957536.87 frames. ], batch size: 29, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:39:20,282 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.606e+02 1.874e+02 2.297e+02 4.541e+02, threshold=3.747e+02, percent-clipped=1.0 +2023-04-27 07:39:37,948 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1893, 1.3462, 1.3106, 1.5758, 1.4581, 1.5621, 1.3109, 3.0703], + device='cuda:3'), covar=tensor([0.0697, 0.0952, 0.0952, 0.1472, 0.0773, 0.0592, 0.0898, 0.0218], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 07:39:44,103 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5998, 1.1018, 1.6826, 2.0532, 1.6768, 1.5559, 1.6419, 1.5837], + device='cuda:3'), covar=tensor([0.4700, 0.6900, 0.6104, 0.6002, 0.5802, 0.8148, 0.7729, 0.8096], + device='cuda:3'), in_proj_covar=tensor([0.0417, 0.0408, 0.0494, 0.0507, 0.0445, 0.0469, 0.0475, 0.0478], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:39:47,578 INFO [finetune.py:976] (3/7) Epoch 15, batch 5200, loss[loss=0.2151, simple_loss=0.2779, pruned_loss=0.07613, over 4897.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2512, pruned_loss=0.05643, over 957313.04 frames. ], batch size: 32, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:40:15,422 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85428.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:18,543 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8080, 2.2880, 1.7307, 1.7315, 1.2912, 1.3351, 1.8220, 1.2845], + device='cuda:3'), covar=tensor([0.1612, 0.1340, 0.1474, 0.1646, 0.2450, 0.1860, 0.1014, 0.1996], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0212, 0.0168, 0.0203, 0.0201, 0.0184, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 07:40:21,370 INFO [finetune.py:976] (3/7) Epoch 15, batch 5250, loss[loss=0.1585, simple_loss=0.2358, pruned_loss=0.04059, over 4812.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2527, pruned_loss=0.05641, over 957984.39 frames. ], batch size: 25, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:40:21,520 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5486, 1.8893, 2.3761, 2.9348, 2.3637, 1.8204, 1.7653, 2.2981], + device='cuda:3'), covar=tensor([0.3026, 0.3203, 0.1439, 0.2606, 0.2561, 0.2441, 0.3848, 0.2127], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0246, 0.0223, 0.0315, 0.0216, 0.0230, 0.0229, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 07:40:27,425 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.183e+02 1.584e+02 2.032e+02 2.482e+02 6.006e+02, threshold=4.063e+02, percent-clipped=4.0 +2023-04-27 07:40:33,765 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85456.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:36,111 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.7193, 4.6701, 3.0498, 5.3290, 4.7048, 4.6234, 2.1690, 4.5534], + device='cuda:3'), covar=tensor([0.1411, 0.0685, 0.2882, 0.0816, 0.2433, 0.1447, 0.5175, 0.2060], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0213, 0.0250, 0.0301, 0.0295, 0.0247, 0.0270, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:40:38,412 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85463.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:40:44,661 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:51,999 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:54,957 INFO [finetune.py:976] (3/7) Epoch 15, batch 5300, loss[loss=0.1729, simple_loss=0.24, pruned_loss=0.05292, over 4815.00 frames. ], tot_loss[loss=0.184, simple_loss=0.254, pruned_loss=0.057, over 955638.96 frames. ], batch size: 38, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:40:55,671 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85489.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:04,897 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85504.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:23,649 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85530.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:28,364 INFO [finetune.py:976] (3/7) Epoch 15, batch 5350, loss[loss=0.1647, simple_loss=0.2409, pruned_loss=0.04424, over 4882.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2534, pruned_loss=0.05623, over 956203.67 frames. ], batch size: 35, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:41:32,136 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85544.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:34,433 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.243e+02 1.620e+02 1.904e+02 2.320e+02 4.507e+02, threshold=3.809e+02, percent-clipped=1.0 +2023-04-27 07:41:44,559 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8773, 1.1616, 1.7067, 1.8424, 1.8060, 1.8397, 1.6919, 1.6953], + device='cuda:3'), covar=tensor([0.4170, 0.5317, 0.4718, 0.4769, 0.5558, 0.8166, 0.4882, 0.4592], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0374, 0.0320, 0.0333, 0.0344, 0.0399, 0.0354, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:42:02,177 INFO [finetune.py:976] (3/7) Epoch 15, batch 5400, loss[loss=0.1952, simple_loss=0.256, pruned_loss=0.06722, over 4747.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2514, pruned_loss=0.05621, over 957116.53 frames. ], batch size: 54, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:42:04,170 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85591.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:42:13,927 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85607.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:42:27,849 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:42:41,670 INFO [finetune.py:976] (3/7) Epoch 15, batch 5450, loss[loss=0.1673, simple_loss=0.2433, pruned_loss=0.04563, over 4915.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2484, pruned_loss=0.05536, over 956796.72 frames. ], batch size: 36, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:42:53,051 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.619e+02 1.980e+02 2.331e+02 6.511e+02, threshold=3.961e+02, percent-clipped=1.0 +2023-04-27 07:43:02,875 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85655.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:43:45,382 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85686.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:43:46,491 INFO [finetune.py:976] (3/7) Epoch 15, batch 5500, loss[loss=0.1621, simple_loss=0.2339, pruned_loss=0.04509, over 4834.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2457, pruned_loss=0.05452, over 957973.76 frames. ], batch size: 30, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:44:51,824 INFO [finetune.py:976] (3/7) Epoch 15, batch 5550, loss[loss=0.1431, simple_loss=0.2191, pruned_loss=0.03356, over 4777.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2472, pruned_loss=0.05533, over 957179.13 frames. ], batch size: 26, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:45:02,123 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.632e+02 1.848e+02 2.307e+02 4.705e+02, threshold=3.696e+02, percent-clipped=4.0 +2023-04-27 07:45:11,440 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85763.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:16,811 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85771.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:18,171 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-27 07:45:24,798 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:27,127 INFO [finetune.py:976] (3/7) Epoch 15, batch 5600, loss[loss=0.2034, simple_loss=0.2784, pruned_loss=0.06424, over 4815.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2519, pruned_loss=0.05677, over 955200.95 frames. ], batch size: 38, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:45:28,370 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9160, 1.8211, 1.6546, 1.5347, 1.9033, 1.5713, 2.3736, 1.4909], + device='cuda:3'), covar=tensor([0.4046, 0.1838, 0.4517, 0.3219, 0.1873, 0.2501, 0.1383, 0.4617], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0347, 0.0424, 0.0355, 0.0383, 0.0380, 0.0371, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:45:34,307 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6242, 1.4125, 1.7190, 1.7979, 1.4649, 1.3529, 1.5375, 0.9399], + device='cuda:3'), covar=tensor([0.0519, 0.0746, 0.0480, 0.0607, 0.0728, 0.1178, 0.0534, 0.0791], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0070, 0.0070, 0.0068, 0.0076, 0.0098, 0.0076, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:45:37,760 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3090, 1.2424, 1.5644, 1.4987, 1.1934, 1.0980, 1.3305, 0.9565], + device='cuda:3'), covar=tensor([0.0587, 0.0665, 0.0459, 0.0721, 0.0795, 0.1118, 0.0669, 0.0641], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0070, 0.0070, 0.0068, 0.0076, 0.0098, 0.0076, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:45:40,675 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85811.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:45,403 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:45,485 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6831, 2.0430, 1.8403, 2.0449, 1.7944, 2.0054, 1.9387, 1.8536], + device='cuda:3'), covar=tensor([0.4223, 0.6723, 0.6071, 0.5033, 0.6431, 0.8416, 0.7333, 0.6946], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0376, 0.0321, 0.0335, 0.0345, 0.0401, 0.0355, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:45:57,426 INFO [finetune.py:976] (3/7) Epoch 15, batch 5650, loss[loss=0.1961, simple_loss=0.2712, pruned_loss=0.06046, over 4909.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2542, pruned_loss=0.05662, over 954914.72 frames. ], batch size: 37, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:45:58,080 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85839.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:46:03,466 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.549e+02 1.787e+02 2.141e+02 3.216e+02, threshold=3.573e+02, percent-clipped=0.0 +2023-04-27 07:46:09,099 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5006, 1.9871, 2.0752, 2.3502, 2.2162, 2.2714, 1.8075, 4.7465], + device='cuda:3'), covar=tensor([0.0533, 0.0746, 0.0711, 0.1098, 0.0587, 0.0460, 0.0679, 0.0119], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 07:46:18,021 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8289, 2.1452, 0.7862, 1.1270, 1.5347, 1.1285, 2.3212, 1.3721], + device='cuda:3'), covar=tensor([0.0591, 0.0509, 0.0606, 0.1226, 0.0399, 0.0942, 0.0367, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 07:46:19,884 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85875.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:46:26,479 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85886.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:46:27,651 INFO [finetune.py:976] (3/7) Epoch 15, batch 5700, loss[loss=0.1642, simple_loss=0.2171, pruned_loss=0.05568, over 4437.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2498, pruned_loss=0.05541, over 938970.17 frames. ], batch size: 19, lr: 3.47e-03, grad_scale: 64.0 +2023-04-27 07:46:32,306 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 07:46:59,231 INFO [finetune.py:976] (3/7) Epoch 16, batch 0, loss[loss=0.1809, simple_loss=0.2522, pruned_loss=0.05482, over 4792.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2522, pruned_loss=0.05482, over 4792.00 frames. ], batch size: 29, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:46:59,231 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 07:47:06,809 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2024, 2.4845, 1.0330, 1.5005, 1.9641, 1.4217, 2.9918, 1.7694], + device='cuda:3'), covar=tensor([0.0541, 0.0588, 0.0671, 0.1117, 0.0386, 0.0819, 0.0254, 0.0529], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 07:47:15,737 INFO [finetune.py:1010] (3/7) Epoch 16, validation: loss=0.1534, simple_loss=0.2252, pruned_loss=0.04076, over 2265189.00 frames. +2023-04-27 07:47:15,738 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 07:47:30,502 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85930.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:47:34,159 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:47:41,352 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.546e+02 1.841e+02 2.213e+02 4.481e+02, threshold=3.682e+02, percent-clipped=3.0 +2023-04-27 07:47:43,313 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5183, 3.1644, 1.0543, 1.8536, 1.9017, 2.2799, 1.9296, 0.9399], + device='cuda:3'), covar=tensor([0.1196, 0.0791, 0.1767, 0.1069, 0.0832, 0.0904, 0.1229, 0.1956], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0244, 0.0136, 0.0121, 0.0131, 0.0152, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:47:46,981 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85957.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:47:53,857 INFO [finetune.py:976] (3/7) Epoch 16, batch 50, loss[loss=0.2123, simple_loss=0.285, pruned_loss=0.06983, over 4876.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2551, pruned_loss=0.05897, over 215022.11 frames. ], batch size: 35, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:48:04,121 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85981.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:48:10,278 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85991.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:48:23,358 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86002.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:48:44,015 INFO [finetune.py:976] (3/7) Epoch 16, batch 100, loss[loss=0.1713, simple_loss=0.234, pruned_loss=0.05433, over 4834.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.247, pruned_loss=0.05539, over 378728.18 frames. ], batch size: 30, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:48:45,854 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:49:24,770 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6035, 3.9280, 0.7238, 1.9833, 2.1448, 2.3637, 2.2833, 0.8574], + device='cuda:3'), covar=tensor([0.1411, 0.0730, 0.2203, 0.1262, 0.0987, 0.1168, 0.1357, 0.2247], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0244, 0.0136, 0.0121, 0.0131, 0.0152, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:49:27,105 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.526e+02 1.932e+02 2.355e+02 3.544e+02, threshold=3.863e+02, percent-clipped=0.0 +2023-04-27 07:49:34,864 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86052.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:49:47,408 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86063.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:49:49,096 INFO [finetune.py:976] (3/7) Epoch 16, batch 150, loss[loss=0.148, simple_loss=0.2178, pruned_loss=0.03905, over 4733.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2424, pruned_loss=0.05405, over 506763.90 frames. ], batch size: 59, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:49:52,567 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86069.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:49:57,386 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 07:50:02,657 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86084.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:50:37,995 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86113.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:50:45,303 INFO [finetune.py:976] (3/7) Epoch 16, batch 200, loss[loss=0.2297, simple_loss=0.2836, pruned_loss=0.08786, over 4822.00 frames. ], tot_loss[loss=0.177, simple_loss=0.243, pruned_loss=0.05551, over 606573.86 frames. ], batch size: 40, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:51:06,770 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86130.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:51:07,904 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86132.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:51:16,800 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86139.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:51:22,244 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.564e+02 1.820e+02 2.219e+02 4.567e+02, threshold=3.639e+02, percent-clipped=1.0 +2023-04-27 07:51:33,767 INFO [finetune.py:976] (3/7) Epoch 16, batch 250, loss[loss=0.2076, simple_loss=0.2772, pruned_loss=0.06897, over 4833.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2473, pruned_loss=0.05655, over 682410.15 frames. ], batch size: 40, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:51:48,354 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86186.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:51:48,949 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86187.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:51:54,413 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3110, 3.4455, 0.7409, 1.8132, 1.8050, 2.3339, 1.8993, 0.9473], + device='cuda:3'), covar=tensor([0.1520, 0.0711, 0.2164, 0.1314, 0.1140, 0.1092, 0.1542, 0.2137], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0245, 0.0137, 0.0121, 0.0131, 0.0153, 0.0119, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:52:06,460 INFO [finetune.py:976] (3/7) Epoch 16, batch 300, loss[loss=0.1917, simple_loss=0.2723, pruned_loss=0.05555, over 4841.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2529, pruned_loss=0.05714, over 744280.90 frames. ], batch size: 49, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:52:17,998 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86231.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:19,822 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86234.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:20,583 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.65 vs. limit=5.0 +2023-04-27 07:52:27,210 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86246.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:28,287 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.717e+02 1.966e+02 2.307e+02 3.761e+02, threshold=3.931e+02, percent-clipped=1.0 +2023-04-27 07:52:30,218 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.0473, 3.9230, 2.7570, 4.6327, 4.0178, 4.0287, 1.7067, 4.0089], + device='cuda:3'), covar=tensor([0.1333, 0.1087, 0.3074, 0.1203, 0.2446, 0.1501, 0.5383, 0.2171], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0214, 0.0251, 0.0304, 0.0297, 0.0247, 0.0271, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 07:52:39,295 INFO [finetune.py:976] (3/7) Epoch 16, batch 350, loss[loss=0.199, simple_loss=0.276, pruned_loss=0.06093, over 4869.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2542, pruned_loss=0.05709, over 791256.28 frames. ], batch size: 43, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:52:45,434 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8669, 1.3531, 1.4422, 1.6448, 2.1575, 1.5971, 1.3325, 1.3950], + device='cuda:3'), covar=tensor([0.1593, 0.1561, 0.1916, 0.1317, 0.0805, 0.1723, 0.2109, 0.2148], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0316, 0.0353, 0.0292, 0.0332, 0.0313, 0.0301, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.3701e-05, 6.6055e-05, 7.5404e-05, 5.9583e-05, 6.9171e-05, 6.6080e-05, + 6.3631e-05, 7.7644e-05], device='cuda:3') +2023-04-27 07:52:50,049 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86281.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:54,333 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86286.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:53:07,226 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86307.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:53:10,808 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86313.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:53:12,610 INFO [finetune.py:976] (3/7) Epoch 16, batch 400, loss[loss=0.1364, simple_loss=0.2047, pruned_loss=0.034, over 4798.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2534, pruned_loss=0.05648, over 827450.80 frames. ], batch size: 25, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:53:13,329 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2166, 1.3776, 1.2533, 1.7247, 1.4854, 1.5381, 1.2676, 3.0187], + device='cuda:3'), covar=tensor([0.0725, 0.1100, 0.1044, 0.1340, 0.0857, 0.0649, 0.1008, 0.0266], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 07:53:21,090 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:53:35,101 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8149, 1.3571, 1.4030, 1.5880, 2.0601, 1.6334, 1.3500, 1.3260], + device='cuda:3'), covar=tensor([0.1519, 0.1512, 0.1877, 0.1167, 0.0640, 0.1502, 0.2158, 0.2170], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0314, 0.0351, 0.0290, 0.0330, 0.0311, 0.0299, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.3224e-05, 6.5552e-05, 7.4937e-05, 5.9137e-05, 6.8735e-05, 6.5718e-05, + 6.3172e-05, 7.6970e-05], device='cuda:3') +2023-04-27 07:53:35,566 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.581e+02 1.858e+02 2.136e+02 3.923e+02, threshold=3.716e+02, percent-clipped=0.0 +2023-04-27 07:53:41,146 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86358.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:53:45,903 INFO [finetune.py:976] (3/7) Epoch 16, batch 450, loss[loss=0.1783, simple_loss=0.2488, pruned_loss=0.05387, over 4769.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2511, pruned_loss=0.05562, over 855715.44 frames. ], batch size: 26, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:54:14,885 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86408.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:54:19,720 INFO [finetune.py:976] (3/7) Epoch 16, batch 500, loss[loss=0.1763, simple_loss=0.2444, pruned_loss=0.05409, over 4729.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.249, pruned_loss=0.05475, over 878190.53 frames. ], batch size: 59, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:54:25,193 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:54:42,231 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86440.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:54:45,736 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86445.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:54:52,709 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.729e+01 1.556e+02 1.970e+02 2.447e+02 5.301e+02, threshold=3.940e+02, percent-clipped=3.0 +2023-04-27 07:55:08,755 INFO [finetune.py:976] (3/7) Epoch 16, batch 550, loss[loss=0.1829, simple_loss=0.2504, pruned_loss=0.05773, over 4812.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2454, pruned_loss=0.0539, over 895646.16 frames. ], batch size: 51, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:55:26,747 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1720, 2.2124, 1.8224, 1.7708, 2.2016, 1.7809, 2.5628, 1.6144], + device='cuda:3'), covar=tensor([0.3257, 0.1428, 0.3905, 0.2531, 0.1495, 0.2100, 0.1282, 0.3797], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0348, 0.0426, 0.0357, 0.0383, 0.0380, 0.0372, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 07:56:00,653 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86501.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:56:09,575 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86506.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:56:20,905 INFO [finetune.py:976] (3/7) Epoch 16, batch 600, loss[loss=0.2029, simple_loss=0.2614, pruned_loss=0.07216, over 4890.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2469, pruned_loss=0.0549, over 910585.08 frames. ], batch size: 32, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:56:35,190 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86531.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:56:53,559 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8020, 2.2120, 0.8774, 1.1664, 1.5824, 1.0469, 2.4469, 1.2602], + device='cuda:3'), covar=tensor([0.0725, 0.0658, 0.0718, 0.1262, 0.0467, 0.1063, 0.0317, 0.0744], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 07:56:54,677 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.714e+02 1.969e+02 2.421e+02 4.960e+02, threshold=3.938e+02, percent-clipped=3.0 +2023-04-27 07:57:04,546 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9378, 2.3879, 1.2880, 1.5553, 2.2914, 1.8161, 1.6850, 1.7825], + device='cuda:3'), covar=tensor([0.0488, 0.0332, 0.0303, 0.0552, 0.0253, 0.0506, 0.0514, 0.0563], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 07:57:05,050 INFO [finetune.py:976] (3/7) Epoch 16, batch 650, loss[loss=0.1268, simple_loss=0.2041, pruned_loss=0.0247, over 4810.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2499, pruned_loss=0.05524, over 920200.42 frames. ], batch size: 25, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:57:12,975 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86579.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:13,043 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5882, 1.4733, 1.8266, 1.7553, 1.3651, 1.2420, 1.6146, 1.0509], + device='cuda:3'), covar=tensor([0.0548, 0.0664, 0.0410, 0.0714, 0.0841, 0.1274, 0.0639, 0.0721], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0069], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 07:57:13,609 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86580.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:15,468 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7187, 2.2211, 1.6057, 1.4355, 1.2741, 1.2962, 1.5481, 1.2171], + device='cuda:3'), covar=tensor([0.1763, 0.1323, 0.1522, 0.1825, 0.2374, 0.2037, 0.1126, 0.2021], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0212, 0.0168, 0.0205, 0.0200, 0.0184, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 07:57:18,209 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86586.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:57:29,904 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86602.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:36,594 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:38,333 INFO [finetune.py:976] (3/7) Epoch 16, batch 700, loss[loss=0.1909, simple_loss=0.2361, pruned_loss=0.07288, over 4083.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2518, pruned_loss=0.05572, over 927179.64 frames. ], batch size: 17, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:57:49,363 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86634.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:57:54,605 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86641.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:00,292 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.513e+02 1.841e+02 2.164e+02 3.453e+02, threshold=3.681e+02, percent-clipped=0.0 +2023-04-27 07:58:06,422 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86658.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:08,177 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86661.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:11,206 INFO [finetune.py:976] (3/7) Epoch 16, batch 750, loss[loss=0.2021, simple_loss=0.271, pruned_loss=0.06662, over 4916.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2525, pruned_loss=0.05608, over 933446.48 frames. ], batch size: 38, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:58:38,830 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86706.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:40,090 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86708.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:44,816 INFO [finetune.py:976] (3/7) Epoch 16, batch 800, loss[loss=0.1906, simple_loss=0.2671, pruned_loss=0.05705, over 4810.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2534, pruned_loss=0.0568, over 936376.44 frames. ], batch size: 39, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:58:50,342 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86725.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:05,803 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.560e+02 1.911e+02 2.310e+02 3.714e+02, threshold=3.821e+02, percent-clipped=1.0 +2023-04-27 07:59:11,625 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86756.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:17,588 INFO [finetune.py:976] (3/7) Epoch 16, batch 850, loss[loss=0.1825, simple_loss=0.2567, pruned_loss=0.05411, over 4186.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2508, pruned_loss=0.05609, over 939930.25 frames. ], batch size: 65, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:59:21,922 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86773.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:24,487 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.66 vs. limit=5.0 +2023-04-27 07:59:36,382 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86796.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:39,984 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86801.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:59:49,824 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86814.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:50,969 INFO [finetune.py:976] (3/7) Epoch 16, batch 900, loss[loss=0.1598, simple_loss=0.23, pruned_loss=0.04479, over 4815.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2473, pruned_loss=0.05456, over 942255.91 frames. ], batch size: 41, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:00:18,131 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5935, 3.4217, 0.8332, 1.8826, 1.9676, 2.3827, 1.9596, 0.9241], + device='cuda:3'), covar=tensor([0.1273, 0.1098, 0.2000, 0.1241, 0.1023, 0.1053, 0.1421, 0.1953], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0245, 0.0137, 0.0121, 0.0131, 0.0152, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:00:18,159 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0736, 1.6886, 1.9918, 2.3151, 2.2999, 1.8915, 1.6360, 2.1737], + device='cuda:3'), covar=tensor([0.0861, 0.1260, 0.0656, 0.0646, 0.0666, 0.0879, 0.0841, 0.0538], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0206, 0.0184, 0.0175, 0.0179, 0.0186, 0.0156, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:00:22,180 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.518e+02 1.851e+02 2.220e+02 4.721e+02, threshold=3.701e+02, percent-clipped=1.0 +2023-04-27 08:00:52,314 INFO [finetune.py:976] (3/7) Epoch 16, batch 950, loss[loss=0.2069, simple_loss=0.2795, pruned_loss=0.06718, over 4855.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2447, pruned_loss=0.05322, over 947066.81 frames. ], batch size: 44, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:01:03,390 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86875.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:01:05,176 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86878.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:01:37,241 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86902.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:01:58,440 INFO [finetune.py:976] (3/7) Epoch 16, batch 1000, loss[loss=0.2498, simple_loss=0.3165, pruned_loss=0.09159, over 4763.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2471, pruned_loss=0.05409, over 950258.42 frames. ], batch size: 54, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:02:18,017 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:02:19,923 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86939.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:02:29,740 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3286, 1.2417, 1.5373, 1.5057, 1.1990, 1.1278, 1.2326, 0.7166], + device='cuda:3'), covar=tensor([0.0577, 0.0566, 0.0427, 0.0597, 0.0790, 0.1145, 0.0584, 0.0706], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0075, 0.0097, 0.0075, 0.0068], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:02:31,464 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.765e+02 1.981e+02 2.453e+02 9.285e+02, threshold=3.962e+02, percent-clipped=3.0 +2023-04-27 08:02:37,037 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:02:59,848 INFO [finetune.py:976] (3/7) Epoch 16, batch 1050, loss[loss=0.1972, simple_loss=0.2695, pruned_loss=0.06251, over 4819.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2495, pruned_loss=0.0538, over 953315.62 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:03:28,668 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87002.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:03:41,500 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-27 08:03:43,717 INFO [finetune.py:976] (3/7) Epoch 16, batch 1100, loss[loss=0.1756, simple_loss=0.2449, pruned_loss=0.05311, over 4920.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2507, pruned_loss=0.05441, over 953951.59 frames. ], batch size: 33, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:04:06,135 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.660e+02 1.994e+02 2.342e+02 3.600e+02, threshold=3.988e+02, percent-clipped=0.0 +2023-04-27 08:04:08,081 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7862, 2.4358, 1.7977, 1.7689, 1.2921, 1.2970, 1.9884, 1.2749], + device='cuda:3'), covar=tensor([0.1682, 0.1278, 0.1518, 0.1757, 0.2407, 0.1931, 0.0981, 0.2047], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0211, 0.0167, 0.0203, 0.0199, 0.0182, 0.0154, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 08:04:15,689 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87063.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:04:17,376 INFO [finetune.py:976] (3/7) Epoch 16, batch 1150, loss[loss=0.2067, simple_loss=0.2726, pruned_loss=0.07041, over 4818.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2525, pruned_loss=0.05545, over 954699.67 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:04:37,608 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87096.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:04:40,659 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87101.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:04:50,580 INFO [finetune.py:976] (3/7) Epoch 16, batch 1200, loss[loss=0.2027, simple_loss=0.2789, pruned_loss=0.06329, over 4788.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.251, pruned_loss=0.05468, over 953554.13 frames. ], batch size: 51, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:04:51,379 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 08:05:10,108 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:05:13,050 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.748e+02 2.017e+02 2.271e+02 5.228e+02, threshold=4.034e+02, percent-clipped=1.0 +2023-04-27 08:05:13,120 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87149.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:05:24,512 INFO [finetune.py:976] (3/7) Epoch 16, batch 1250, loss[loss=0.129, simple_loss=0.212, pruned_loss=0.02304, over 4788.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.248, pruned_loss=0.05339, over 954434.77 frames. ], batch size: 51, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:05:27,049 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87170.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:05:52,202 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6464, 1.3104, 1.2622, 1.4248, 1.8297, 1.5009, 1.2559, 1.2233], + device='cuda:3'), covar=tensor([0.1280, 0.1353, 0.1832, 0.1335, 0.0794, 0.1617, 0.1956, 0.2029], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0313, 0.0352, 0.0289, 0.0329, 0.0312, 0.0300, 0.0362], + device='cuda:3'), out_proj_covar=tensor([6.3079e-05, 6.5378e-05, 7.5182e-05, 5.8876e-05, 6.8631e-05, 6.5931e-05, + 6.3333e-05, 7.7220e-05], device='cuda:3') +2023-04-27 08:05:58,112 INFO [finetune.py:976] (3/7) Epoch 16, batch 1300, loss[loss=0.1785, simple_loss=0.2405, pruned_loss=0.05821, over 4818.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2455, pruned_loss=0.05307, over 955918.43 frames. ], batch size: 41, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:06:09,203 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87230.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:12,073 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87234.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:13,312 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87236.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:21,119 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.625e+02 1.883e+02 2.267e+02 3.477e+02, threshold=3.766e+02, percent-clipped=0.0 +2023-04-27 08:06:31,975 INFO [finetune.py:976] (3/7) Epoch 16, batch 1350, loss[loss=0.1742, simple_loss=0.2597, pruned_loss=0.04432, over 4859.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2448, pruned_loss=0.05279, over 957007.13 frames. ], batch size: 44, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:06:45,593 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87284.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:48,763 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 08:06:50,347 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87291.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:07:14,207 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-04-27 08:07:16,734 INFO [finetune.py:976] (3/7) Epoch 16, batch 1400, loss[loss=0.1563, simple_loss=0.2524, pruned_loss=0.03008, over 4913.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2484, pruned_loss=0.05391, over 955731.86 frames. ], batch size: 42, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:07:58,924 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.760e+02 2.100e+02 2.306e+02 6.240e+02, threshold=4.200e+02, percent-clipped=3.0 +2023-04-27 08:08:06,821 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5270, 2.6647, 2.3198, 2.3613, 2.7867, 2.3984, 3.6037, 1.9457], + device='cuda:3'), covar=tensor([0.4068, 0.2243, 0.4422, 0.3656, 0.1972, 0.2716, 0.1341, 0.4473], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0353, 0.0433, 0.0362, 0.0390, 0.0386, 0.0377, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:08:15,343 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87358.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:08:17,781 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87362.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:08:20,151 INFO [finetune.py:976] (3/7) Epoch 16, batch 1450, loss[loss=0.2241, simple_loss=0.2987, pruned_loss=0.07477, over 4811.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2511, pruned_loss=0.05512, over 953700.65 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:08:25,413 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 08:09:12,953 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1870, 1.4566, 1.3977, 1.7067, 1.6005, 2.0111, 1.3242, 3.6580], + device='cuda:3'), covar=tensor([0.0628, 0.0840, 0.0796, 0.1248, 0.0674, 0.0524, 0.0795, 0.0130], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 08:09:14,070 INFO [finetune.py:976] (3/7) Epoch 16, batch 1500, loss[loss=0.2012, simple_loss=0.2744, pruned_loss=0.06399, over 4810.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2527, pruned_loss=0.05584, over 954557.41 frames. ], batch size: 40, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:09:17,558 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.38 vs. limit=5.0 +2023-04-27 08:09:18,464 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87423.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:09:36,459 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.658e+02 1.977e+02 2.415e+02 4.579e+02, threshold=3.953e+02, percent-clipped=1.0 +2023-04-27 08:09:43,839 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87461.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:09:46,778 INFO [finetune.py:976] (3/7) Epoch 16, batch 1550, loss[loss=0.1917, simple_loss=0.2643, pruned_loss=0.0595, over 4915.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2523, pruned_loss=0.05509, over 954250.26 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:09:47,522 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7093, 1.8030, 1.8493, 2.4627, 2.6723, 2.1832, 2.0927, 1.9004], + device='cuda:3'), covar=tensor([0.1707, 0.1758, 0.2233, 0.1761, 0.1085, 0.1985, 0.2208, 0.2113], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0311, 0.0351, 0.0289, 0.0328, 0.0311, 0.0300, 0.0360], + device='cuda:3'), out_proj_covar=tensor([6.3134e-05, 6.5041e-05, 7.4977e-05, 5.8882e-05, 6.8367e-05, 6.5676e-05, + 6.3449e-05, 7.6851e-05], device='cuda:3') +2023-04-27 08:09:49,313 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87470.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:30,184 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87505.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:42,329 INFO [finetune.py:976] (3/7) Epoch 16, batch 1600, loss[loss=0.2, simple_loss=0.2688, pruned_loss=0.06557, over 4903.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2499, pruned_loss=0.05449, over 953874.90 frames. ], batch size: 36, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:10:43,626 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87518.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:46,107 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87522.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:54,391 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87534.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:05,886 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.660e+02 1.902e+02 2.277e+02 5.320e+02, threshold=3.803e+02, percent-clipped=1.0 +2023-04-27 08:11:16,216 INFO [finetune.py:976] (3/7) Epoch 16, batch 1650, loss[loss=0.1265, simple_loss=0.2009, pruned_loss=0.02603, over 4791.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2473, pruned_loss=0.05413, over 954208.49 frames. ], batch size: 29, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:11:16,349 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87566.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:11:26,482 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87582.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:28,254 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87584.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:29,408 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87586.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:29,483 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9588, 2.5216, 1.8798, 1.8388, 1.3241, 1.3814, 2.0545, 1.3307], + device='cuda:3'), covar=tensor([0.1573, 0.1217, 0.1437, 0.1689, 0.2175, 0.1935, 0.0944, 0.1992], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0213, 0.0168, 0.0205, 0.0200, 0.0184, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 08:11:31,292 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87589.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:43,133 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87605.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:49,704 INFO [finetune.py:976] (3/7) Epoch 16, batch 1700, loss[loss=0.1869, simple_loss=0.2651, pruned_loss=0.05431, over 4861.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2459, pruned_loss=0.0539, over 954383.83 frames. ], batch size: 44, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:12:08,466 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87645.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:12:12,238 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.609e+02 2.053e+02 2.465e+02 3.849e+02, threshold=4.106e+02, percent-clipped=1.0 +2023-04-27 08:12:13,477 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:12:18,811 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87658.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:12:23,601 INFO [finetune.py:976] (3/7) Epoch 16, batch 1750, loss[loss=0.1768, simple_loss=0.2626, pruned_loss=0.04545, over 4906.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2475, pruned_loss=0.05452, over 955234.02 frames. ], batch size: 37, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:12:23,730 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87666.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:12:25,108 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 08:12:55,812 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87706.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:13:07,614 INFO [finetune.py:976] (3/7) Epoch 16, batch 1800, loss[loss=0.2011, simple_loss=0.2643, pruned_loss=0.06896, over 4814.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.25, pruned_loss=0.05469, over 956482.84 frames. ], batch size: 51, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:13:14,246 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:13:28,229 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.22 vs. limit=5.0 +2023-04-27 08:13:49,467 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.657e+02 1.996e+02 2.458e+02 4.294e+02, threshold=3.992e+02, percent-clipped=1.0 +2023-04-27 08:14:00,439 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 08:14:12,970 INFO [finetune.py:976] (3/7) Epoch 16, batch 1850, loss[loss=0.1693, simple_loss=0.2336, pruned_loss=0.05252, over 4912.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2516, pruned_loss=0.05535, over 956552.85 frames. ], batch size: 36, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:14:52,536 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5779, 3.0399, 1.0607, 1.7718, 2.5249, 1.6405, 4.2816, 1.9968], + device='cuda:3'), covar=tensor([0.0615, 0.0758, 0.0865, 0.1288, 0.0462, 0.0947, 0.0265, 0.0646], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 08:15:07,316 INFO [finetune.py:976] (3/7) Epoch 16, batch 1900, loss[loss=0.1786, simple_loss=0.2496, pruned_loss=0.05381, over 4768.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2528, pruned_loss=0.05498, over 958787.82 frames. ], batch size: 26, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:15:08,023 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87817.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:15:28,298 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.628e+02 1.844e+02 2.203e+02 4.331e+02, threshold=3.688e+02, percent-clipped=1.0 +2023-04-27 08:15:37,113 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87861.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:15:40,074 INFO [finetune.py:976] (3/7) Epoch 16, batch 1950, loss[loss=0.1821, simple_loss=0.2532, pruned_loss=0.05552, over 4742.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2516, pruned_loss=0.05471, over 957324.48 frames. ], batch size: 54, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:16:10,452 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87886.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:16:47,482 INFO [finetune.py:976] (3/7) Epoch 16, batch 2000, loss[loss=0.1492, simple_loss=0.2212, pruned_loss=0.0386, over 4816.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2492, pruned_loss=0.05439, over 957236.24 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:16:49,919 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5120, 3.4636, 0.8723, 1.9174, 1.9234, 2.5646, 1.9692, 1.0143], + device='cuda:3'), covar=tensor([0.1447, 0.0974, 0.2009, 0.1267, 0.1083, 0.0992, 0.1511, 0.1970], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0245, 0.0137, 0.0121, 0.0130, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:16:58,619 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 08:16:58,997 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87934.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:17:02,681 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87940.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:17:05,747 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:17:05,858 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-27 08:17:08,487 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.621e+02 1.906e+02 2.380e+02 5.072e+02, threshold=3.811e+02, percent-clipped=4.0 +2023-04-27 08:17:17,311 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87961.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:17:21,295 INFO [finetune.py:976] (3/7) Epoch 16, batch 2050, loss[loss=0.183, simple_loss=0.2525, pruned_loss=0.05676, over 4753.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2457, pruned_loss=0.05376, over 956732.31 frames. ], batch size: 26, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:17:55,289 INFO [finetune.py:976] (3/7) Epoch 16, batch 2100, loss[loss=0.1903, simple_loss=0.255, pruned_loss=0.06283, over 4919.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2448, pruned_loss=0.05336, over 958250.01 frames. ], batch size: 37, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:17:57,082 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:18:10,277 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.1642, 3.1662, 2.4433, 3.7627, 3.2124, 3.2040, 1.4490, 3.1707], + device='cuda:3'), covar=tensor([0.2225, 0.1435, 0.3986, 0.2322, 0.3607, 0.2046, 0.6254, 0.2870], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0214, 0.0251, 0.0303, 0.0300, 0.0249, 0.0271, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 08:18:14,643 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.66 vs. limit=5.0 +2023-04-27 08:18:16,283 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.262e+02 1.786e+02 2.026e+02 2.569e+02 5.831e+02, threshold=4.052e+02, percent-clipped=4.0 +2023-04-27 08:18:28,508 INFO [finetune.py:976] (3/7) Epoch 16, batch 2150, loss[loss=0.2003, simple_loss=0.2736, pruned_loss=0.0635, over 4275.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2483, pruned_loss=0.05487, over 955549.86 frames. ], batch size: 66, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:18:28,582 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88066.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:18:50,553 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88100.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:01,076 INFO [finetune.py:976] (3/7) Epoch 16, batch 2200, loss[loss=0.1701, simple_loss=0.2547, pruned_loss=0.04275, over 4814.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2503, pruned_loss=0.05512, over 955164.01 frames. ], batch size: 40, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:19:02,289 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88117.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:08,632 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4344, 1.3260, 4.3117, 4.0478, 3.8434, 4.1036, 4.0548, 3.7755], + device='cuda:3'), covar=tensor([0.7795, 0.6283, 0.1084, 0.1785, 0.1177, 0.1628, 0.1283, 0.1784], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0304, 0.0403, 0.0406, 0.0348, 0.0407, 0.0310, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:19:34,550 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.597e+02 1.844e+02 2.202e+02 4.029e+02, threshold=3.688e+02, percent-clipped=0.0 +2023-04-27 08:19:47,594 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88161.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:19:47,638 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88161.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:55,584 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:56,117 INFO [finetune.py:976] (3/7) Epoch 16, batch 2250, loss[loss=0.1665, simple_loss=0.2464, pruned_loss=0.04331, over 4903.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.252, pruned_loss=0.0555, over 952288.67 frames. ], batch size: 46, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:20:21,181 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9663, 1.7453, 2.1739, 2.4701, 2.0445, 1.9504, 2.0489, 2.0040], + device='cuda:3'), covar=tensor([0.5279, 0.7466, 0.7488, 0.6404, 0.6368, 0.8950, 0.9665, 0.9318], + device='cuda:3'), in_proj_covar=tensor([0.0417, 0.0406, 0.0494, 0.0504, 0.0445, 0.0470, 0.0476, 0.0478], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:20:50,811 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88209.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:21:01,644 INFO [finetune.py:976] (3/7) Epoch 16, batch 2300, loss[loss=0.2088, simple_loss=0.2748, pruned_loss=0.07142, over 4886.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2513, pruned_loss=0.05495, over 951983.12 frames. ], batch size: 35, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:21:11,034 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6683, 1.3491, 4.2218, 3.9775, 3.7341, 3.9629, 3.8328, 3.7346], + device='cuda:3'), covar=tensor([0.6998, 0.5640, 0.0952, 0.1426, 0.1087, 0.2080, 0.2167, 0.1450], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0303, 0.0401, 0.0405, 0.0347, 0.0405, 0.0310, 0.0362], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:21:18,881 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88240.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:21:21,959 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88245.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:21:24,250 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.059e+01 1.606e+02 1.877e+02 2.265e+02 8.419e+02, threshold=3.754e+02, percent-clipped=4.0 +2023-04-27 08:21:32,156 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88261.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:21:41,586 INFO [finetune.py:976] (3/7) Epoch 16, batch 2350, loss[loss=0.1451, simple_loss=0.2169, pruned_loss=0.03667, over 4765.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.249, pruned_loss=0.05476, over 953388.73 frames. ], batch size: 28, lr: 3.44e-03, grad_scale: 64.0 +2023-04-27 08:22:14,477 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88288.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:22:16,365 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88291.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:22:17,545 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:22:38,197 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88309.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:22:48,484 INFO [finetune.py:976] (3/7) Epoch 16, batch 2400, loss[loss=0.1597, simple_loss=0.2297, pruned_loss=0.04487, over 4938.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2472, pruned_loss=0.05437, over 954728.96 frames. ], batch size: 33, lr: 3.44e-03, grad_scale: 64.0 +2023-04-27 08:23:23,052 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.498e+02 1.832e+02 2.156e+02 3.450e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-04-27 08:23:25,038 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88352.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:23:33,494 INFO [finetune.py:976] (3/7) Epoch 16, batch 2450, loss[loss=0.1474, simple_loss=0.2186, pruned_loss=0.0381, over 4886.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2449, pruned_loss=0.05353, over 954700.84 frames. ], batch size: 32, lr: 3.44e-03, grad_scale: 64.0 +2023-04-27 08:24:30,566 INFO [finetune.py:976] (3/7) Epoch 16, batch 2500, loss[loss=0.166, simple_loss=0.2369, pruned_loss=0.04755, over 4830.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2465, pruned_loss=0.05425, over 954833.83 frames. ], batch size: 33, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:24:54,797 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.230e+02 1.712e+02 2.054e+02 2.521e+02 4.262e+02, threshold=4.109e+02, percent-clipped=2.0 +2023-04-27 08:24:58,640 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88456.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:25:04,746 INFO [finetune.py:976] (3/7) Epoch 16, batch 2550, loss[loss=0.1648, simple_loss=0.2438, pruned_loss=0.04292, over 4864.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2485, pruned_loss=0.05415, over 955560.72 frames. ], batch size: 31, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:25:19,565 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5753, 1.7186, 0.6626, 1.2965, 1.9084, 1.4562, 1.3370, 1.4199], + device='cuda:3'), covar=tensor([0.0502, 0.0350, 0.0367, 0.0543, 0.0272, 0.0512, 0.0494, 0.0566], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 08:25:36,246 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7705, 1.8606, 0.9190, 1.4289, 2.0979, 1.5819, 1.4985, 1.5477], + device='cuda:3'), covar=tensor([0.0465, 0.0325, 0.0313, 0.0523, 0.0256, 0.0487, 0.0469, 0.0505], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 08:25:38,575 INFO [finetune.py:976] (3/7) Epoch 16, batch 2600, loss[loss=0.1714, simple_loss=0.2355, pruned_loss=0.05361, over 4873.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2502, pruned_loss=0.05441, over 955202.26 frames. ], batch size: 34, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:25:41,392 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.29 vs. limit=5.0 +2023-04-27 08:25:47,740 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 08:26:18,046 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.329e+02 1.735e+02 2.047e+02 2.467e+02 4.454e+02, threshold=4.094e+02, percent-clipped=2.0 +2023-04-27 08:26:39,231 INFO [finetune.py:976] (3/7) Epoch 16, batch 2650, loss[loss=0.1258, simple_loss=0.1834, pruned_loss=0.03412, over 3403.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.251, pruned_loss=0.05477, over 952724.44 frames. ], batch size: 14, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:26:42,527 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 08:27:29,040 INFO [finetune.py:976] (3/7) Epoch 16, batch 2700, loss[loss=0.1915, simple_loss=0.2497, pruned_loss=0.06663, over 4241.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2506, pruned_loss=0.05416, over 953983.80 frames. ], batch size: 65, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:28:03,360 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2753, 1.2976, 1.3881, 1.5611, 1.6329, 1.3341, 0.8961, 1.4811], + device='cuda:3'), covar=tensor([0.0873, 0.1347, 0.0816, 0.0592, 0.0697, 0.0859, 0.0897, 0.0584], + device='cuda:3'), in_proj_covar=tensor([0.0192, 0.0204, 0.0185, 0.0175, 0.0178, 0.0184, 0.0156, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:28:11,106 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88647.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:28:13,347 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.464e+02 1.764e+02 2.078e+02 3.620e+02, threshold=3.528e+02, percent-clipped=0.0 +2023-04-27 08:28:36,532 INFO [finetune.py:976] (3/7) Epoch 16, batch 2750, loss[loss=0.1575, simple_loss=0.2233, pruned_loss=0.04583, over 4905.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2485, pruned_loss=0.05378, over 955301.47 frames. ], batch size: 43, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:29:21,107 INFO [finetune.py:976] (3/7) Epoch 16, batch 2800, loss[loss=0.1474, simple_loss=0.2267, pruned_loss=0.034, over 4911.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2451, pruned_loss=0.05281, over 954796.63 frames. ], batch size: 36, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:29:29,077 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6871, 1.6462, 1.6282, 1.2922, 1.7727, 1.4968, 2.2269, 1.4818], + device='cuda:3'), covar=tensor([0.3727, 0.1913, 0.5003, 0.2970, 0.1794, 0.2230, 0.1470, 0.4721], + device='cuda:3'), in_proj_covar=tensor([0.0348, 0.0353, 0.0433, 0.0363, 0.0390, 0.0388, 0.0378, 0.0426], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:29:42,737 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.582e+02 1.928e+02 2.328e+02 4.353e+02, threshold=3.856e+02, percent-clipped=4.0 +2023-04-27 08:29:47,477 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88756.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:29:54,459 INFO [finetune.py:976] (3/7) Epoch 16, batch 2850, loss[loss=0.1872, simple_loss=0.2289, pruned_loss=0.0728, over 3962.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2434, pruned_loss=0.05226, over 952920.62 frames. ], batch size: 17, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:30:01,876 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-04-27 08:30:20,171 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88804.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:30:28,459 INFO [finetune.py:976] (3/7) Epoch 16, batch 2900, loss[loss=0.205, simple_loss=0.2835, pruned_loss=0.06327, over 4812.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2462, pruned_loss=0.05331, over 950905.41 frames. ], batch size: 40, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:30:49,013 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9968, 1.0671, 1.1693, 1.1836, 1.0239, 0.8952, 0.9677, 0.5737], + device='cuda:3'), covar=tensor([0.0548, 0.0530, 0.0487, 0.0515, 0.0715, 0.1233, 0.0454, 0.0673], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0076, 0.0097, 0.0075, 0.0068], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:30:50,707 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.633e+02 2.026e+02 2.312e+02 4.286e+02, threshold=4.053e+02, percent-clipped=1.0 +2023-04-27 08:31:02,488 INFO [finetune.py:976] (3/7) Epoch 16, batch 2950, loss[loss=0.2091, simple_loss=0.2812, pruned_loss=0.06857, over 4817.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2486, pruned_loss=0.05396, over 950761.95 frames. ], batch size: 39, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:31:08,635 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4276, 1.7325, 1.6189, 1.8188, 1.8089, 1.8730, 1.5597, 3.1026], + device='cuda:3'), covar=tensor([0.0522, 0.0604, 0.0628, 0.0965, 0.0488, 0.0522, 0.0646, 0.0184], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 08:31:58,220 INFO [finetune.py:976] (3/7) Epoch 16, batch 3000, loss[loss=0.1636, simple_loss=0.228, pruned_loss=0.04958, over 4760.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.25, pruned_loss=0.05475, over 951242.15 frames. ], batch size: 23, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:31:58,220 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 08:32:04,794 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4327, 1.2955, 1.5542, 1.6168, 1.3092, 1.2275, 1.3496, 0.8506], + device='cuda:3'), covar=tensor([0.0603, 0.0672, 0.0489, 0.0646, 0.0944, 0.1258, 0.0579, 0.0695], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0068], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:32:07,776 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1898, 1.6486, 2.0297, 2.2284, 1.9778, 1.6293, 1.1895, 1.6778], + device='cuda:3'), covar=tensor([0.3234, 0.3422, 0.1729, 0.2277, 0.2792, 0.2898, 0.4067, 0.2172], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0245, 0.0223, 0.0313, 0.0215, 0.0229, 0.0226, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 08:32:14,556 INFO [finetune.py:1010] (3/7) Epoch 16, validation: loss=0.1523, simple_loss=0.2234, pruned_loss=0.04062, over 2265189.00 frames. +2023-04-27 08:32:14,557 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 08:32:29,680 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6529, 1.9589, 2.1084, 2.2111, 2.1001, 2.2035, 2.2140, 2.1681], + device='cuda:3'), covar=tensor([0.4376, 0.5467, 0.4550, 0.4384, 0.5240, 0.6778, 0.5096, 0.4894], + device='cuda:3'), in_proj_covar=tensor([0.0331, 0.0370, 0.0317, 0.0332, 0.0343, 0.0394, 0.0351, 0.0323], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 08:32:46,743 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4025, 1.6727, 1.7434, 1.8842, 1.7071, 1.8280, 1.8323, 1.8113], + device='cuda:3'), covar=tensor([0.4265, 0.6337, 0.4970, 0.5031, 0.6065, 0.7555, 0.6216, 0.5775], + device='cuda:3'), in_proj_covar=tensor([0.0330, 0.0371, 0.0317, 0.0332, 0.0343, 0.0393, 0.0351, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 08:32:47,288 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88947.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:32:49,015 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.654e+02 1.995e+02 2.342e+02 3.743e+02, threshold=3.990e+02, percent-clipped=0.0 +2023-04-27 08:32:59,789 INFO [finetune.py:976] (3/7) Epoch 16, batch 3050, loss[loss=0.1778, simple_loss=0.2544, pruned_loss=0.05066, over 4924.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2514, pruned_loss=0.05575, over 950612.92 frames. ], batch size: 42, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:33:30,461 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88995.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:33:31,731 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4426, 1.3252, 1.6437, 1.6764, 1.3558, 1.1424, 1.4235, 0.9267], + device='cuda:3'), covar=tensor([0.0517, 0.0672, 0.0388, 0.0689, 0.0771, 0.1106, 0.0657, 0.0619], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0076, 0.0097, 0.0075, 0.0068], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:33:55,682 INFO [finetune.py:976] (3/7) Epoch 16, batch 3100, loss[loss=0.1342, simple_loss=0.2109, pruned_loss=0.02875, over 4796.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2503, pruned_loss=0.05524, over 951030.34 frames. ], batch size: 29, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:34:13,208 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:34:14,135 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 08:34:38,477 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.571e+02 1.837e+02 2.281e+02 4.804e+02, threshold=3.674e+02, percent-clipped=2.0 +2023-04-27 08:34:59,019 INFO [finetune.py:976] (3/7) Epoch 16, batch 3150, loss[loss=0.2064, simple_loss=0.2599, pruned_loss=0.07648, over 4869.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2473, pruned_loss=0.05416, over 953290.77 frames. ], batch size: 31, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:35:14,721 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:35:37,803 INFO [finetune.py:976] (3/7) Epoch 16, batch 3200, loss[loss=0.1891, simple_loss=0.2612, pruned_loss=0.05844, over 4876.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2441, pruned_loss=0.05326, over 952596.12 frames. ], batch size: 34, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:35:49,528 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 08:36:22,223 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8842, 1.9831, 1.7925, 1.5386, 2.1600, 1.6645, 2.5274, 1.5251], + device='cuda:3'), covar=tensor([0.3702, 0.1868, 0.4474, 0.3138, 0.1615, 0.2394, 0.1317, 0.4753], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0350, 0.0432, 0.0359, 0.0386, 0.0385, 0.0373, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:36:22,290 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 08:36:23,958 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.556e+02 1.903e+02 2.206e+02 4.255e+02, threshold=3.807e+02, percent-clipped=1.0 +2023-04-27 08:36:31,290 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9014, 1.2522, 1.4766, 1.5608, 1.9634, 1.5958, 1.3467, 1.3585], + device='cuda:3'), covar=tensor([0.1889, 0.1862, 0.2226, 0.1529, 0.1048, 0.1777, 0.2834, 0.2302], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0313, 0.0352, 0.0291, 0.0330, 0.0312, 0.0302, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.3884e-05, 6.5268e-05, 7.5271e-05, 5.9269e-05, 6.8703e-05, 6.5894e-05, + 6.3875e-05, 7.7822e-05], device='cuda:3') +2023-04-27 08:36:42,649 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5057, 1.7605, 1.3821, 1.1805, 1.1679, 1.1456, 1.3712, 1.1055], + device='cuda:3'), covar=tensor([0.1703, 0.1331, 0.1470, 0.1747, 0.2306, 0.2007, 0.1071, 0.2064], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0213, 0.0168, 0.0205, 0.0200, 0.0184, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 08:36:44,942 INFO [finetune.py:976] (3/7) Epoch 16, batch 3250, loss[loss=0.1826, simple_loss=0.2274, pruned_loss=0.06884, over 4311.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.245, pruned_loss=0.05335, over 952301.29 frames. ], batch size: 18, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:36:46,907 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89169.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:37:21,517 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0210, 2.5814, 0.9874, 1.3921, 1.8471, 1.3125, 3.4625, 1.5895], + device='cuda:3'), covar=tensor([0.0680, 0.0683, 0.0824, 0.1208, 0.0519, 0.0946, 0.0246, 0.0632], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 08:37:33,749 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2297, 2.7739, 0.9729, 1.5556, 2.0356, 1.4500, 3.6935, 1.9283], + device='cuda:3'), covar=tensor([0.0686, 0.0647, 0.0855, 0.1252, 0.0549, 0.0986, 0.0412, 0.0620], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 08:37:46,007 INFO [finetune.py:976] (3/7) Epoch 16, batch 3300, loss[loss=0.1765, simple_loss=0.2492, pruned_loss=0.05192, over 4765.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2489, pruned_loss=0.05475, over 950608.62 frames. ], batch size: 28, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:38:07,677 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89230.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:38:21,747 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.680e+02 1.944e+02 2.414e+02 4.518e+02, threshold=3.887e+02, percent-clipped=2.0 +2023-04-27 08:38:43,925 INFO [finetune.py:976] (3/7) Epoch 16, batch 3350, loss[loss=0.1697, simple_loss=0.2375, pruned_loss=0.0509, over 4754.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2496, pruned_loss=0.05482, over 949835.52 frames. ], batch size: 27, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:39:01,224 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2943, 3.2963, 2.5535, 3.8632, 3.3092, 3.3255, 1.4832, 3.2541], + device='cuda:3'), covar=tensor([0.2018, 0.1333, 0.3115, 0.2143, 0.3173, 0.2110, 0.5989, 0.2597], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0212, 0.0251, 0.0302, 0.0296, 0.0247, 0.0270, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 08:39:17,177 INFO [finetune.py:976] (3/7) Epoch 16, batch 3400, loss[loss=0.1555, simple_loss=0.2278, pruned_loss=0.04155, over 4905.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2509, pruned_loss=0.05534, over 950061.94 frames. ], batch size: 36, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:39:29,735 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5953, 2.2929, 1.0242, 1.3345, 2.2116, 1.5224, 1.4674, 1.4929], + device='cuda:3'), covar=tensor([0.0611, 0.0296, 0.0320, 0.0619, 0.0236, 0.0643, 0.0648, 0.0671], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 08:39:40,163 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.554e+02 1.949e+02 2.299e+02 3.513e+02, threshold=3.898e+02, percent-clipped=0.0 +2023-04-27 08:39:53,662 INFO [finetune.py:976] (3/7) Epoch 16, batch 3450, loss[loss=0.1404, simple_loss=0.2119, pruned_loss=0.03446, over 4771.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2501, pruned_loss=0.05452, over 951209.43 frames. ], batch size: 27, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:40:04,308 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:40:16,196 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89383.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:40:39,300 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7724, 3.7243, 2.8213, 4.3944, 3.8465, 3.7454, 1.8526, 3.6686], + device='cuda:3'), covar=tensor([0.1702, 0.1214, 0.3058, 0.1631, 0.2610, 0.1778, 0.5853, 0.2547], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0213, 0.0252, 0.0303, 0.0298, 0.0249, 0.0272, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 08:40:54,851 INFO [finetune.py:976] (3/7) Epoch 16, batch 3500, loss[loss=0.145, simple_loss=0.2029, pruned_loss=0.04357, over 4926.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2482, pruned_loss=0.05425, over 954719.59 frames. ], batch size: 33, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:41:00,462 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:41:07,521 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89435.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:41:09,947 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89439.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:41:18,078 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.586e+02 1.877e+02 2.284e+02 3.930e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-04-27 08:41:34,093 INFO [finetune.py:976] (3/7) Epoch 16, batch 3550, loss[loss=0.1671, simple_loss=0.2387, pruned_loss=0.04772, over 4917.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2464, pruned_loss=0.05405, over 954444.39 frames. ], batch size: 37, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:41:55,221 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9147, 1.5255, 1.4284, 1.7807, 2.0923, 1.7429, 1.4220, 1.3734], + device='cuda:3'), covar=tensor([0.1752, 0.1501, 0.2216, 0.1357, 0.1076, 0.1665, 0.2195, 0.2332], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0313, 0.0352, 0.0291, 0.0330, 0.0312, 0.0301, 0.0365], + device='cuda:3'), out_proj_covar=tensor([6.3890e-05, 6.5258e-05, 7.5163e-05, 5.9343e-05, 6.8670e-05, 6.5888e-05, + 6.3691e-05, 7.7978e-05], device='cuda:3') +2023-04-27 08:41:58,287 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89486.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:42:08,143 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89500.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:42:20,188 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89509.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:42:22,381 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-27 08:42:29,612 INFO [finetune.py:976] (3/7) Epoch 16, batch 3600, loss[loss=0.1789, simple_loss=0.2503, pruned_loss=0.0537, over 4741.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2443, pruned_loss=0.05342, over 953623.34 frames. ], batch size: 59, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:42:41,040 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89525.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:43:13,814 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.211e+02 1.668e+02 1.907e+02 2.305e+02 5.911e+02, threshold=3.815e+02, percent-clipped=3.0 +2023-04-27 08:43:27,144 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9750, 2.1476, 1.8326, 1.6793, 2.1429, 1.7488, 2.7551, 1.4413], + device='cuda:3'), covar=tensor([0.3899, 0.1825, 0.4891, 0.3024, 0.1674, 0.2711, 0.1372, 0.5112], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0350, 0.0432, 0.0358, 0.0387, 0.0385, 0.0373, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:43:36,162 INFO [finetune.py:976] (3/7) Epoch 16, batch 3650, loss[loss=0.1899, simple_loss=0.27, pruned_loss=0.05486, over 4799.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2471, pruned_loss=0.05496, over 951547.28 frames. ], batch size: 41, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:43:38,737 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89570.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:44:42,831 INFO [finetune.py:976] (3/7) Epoch 16, batch 3700, loss[loss=0.1997, simple_loss=0.2753, pruned_loss=0.06203, over 4929.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2497, pruned_loss=0.05541, over 952547.73 frames. ], batch size: 38, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:45:03,500 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2678, 2.9450, 2.4657, 2.7569, 2.1043, 2.5148, 2.7407, 2.0067], + device='cuda:3'), covar=tensor([0.2109, 0.1001, 0.0837, 0.1095, 0.2940, 0.1122, 0.1846, 0.2702], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0311, 0.0222, 0.0281, 0.0313, 0.0264, 0.0252, 0.0267], + device='cuda:3'), out_proj_covar=tensor([1.1545e-04, 1.2372e-04, 8.8248e-05, 1.1181e-04, 1.2747e-04, 1.0500e-04, + 1.0194e-04, 1.0639e-04], device='cuda:3') +2023-04-27 08:45:06,550 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9620, 1.3698, 1.2794, 1.6157, 1.4725, 1.4837, 1.3279, 2.4391], + device='cuda:3'), covar=tensor([0.0619, 0.0794, 0.0786, 0.1196, 0.0640, 0.0473, 0.0748, 0.0234], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0057], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 08:45:17,914 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 08:45:26,132 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.724e+02 2.041e+02 2.428e+02 5.061e+02, threshold=4.082e+02, percent-clipped=2.0 +2023-04-27 08:45:29,269 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2028, 2.8978, 2.4495, 2.7119, 2.0768, 2.5511, 2.6964, 1.9739], + device='cuda:3'), covar=tensor([0.2105, 0.1276, 0.0800, 0.1195, 0.2818, 0.1100, 0.1850, 0.2408], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0311, 0.0222, 0.0282, 0.0313, 0.0264, 0.0252, 0.0268], + device='cuda:3'), out_proj_covar=tensor([1.1562e-04, 1.2382e-04, 8.8386e-05, 1.1188e-04, 1.2758e-04, 1.0508e-04, + 1.0201e-04, 1.0649e-04], device='cuda:3') +2023-04-27 08:45:36,266 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9078, 1.6259, 1.8109, 2.2241, 2.2344, 1.7579, 1.5313, 1.9499], + device='cuda:3'), covar=tensor([0.0894, 0.1114, 0.0765, 0.0575, 0.0635, 0.0896, 0.0823, 0.0650], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0202, 0.0182, 0.0174, 0.0179, 0.0182, 0.0154, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:45:48,269 INFO [finetune.py:976] (3/7) Epoch 16, batch 3750, loss[loss=0.1789, simple_loss=0.2549, pruned_loss=0.0514, over 4788.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2521, pruned_loss=0.05633, over 953757.41 frames. ], batch size: 29, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:46:02,525 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:46:10,268 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89683.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:46:33,855 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-27 08:46:39,320 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9374, 2.4567, 2.1340, 2.3351, 1.7022, 2.0880, 2.1544, 1.5957], + device='cuda:3'), covar=tensor([0.1923, 0.0955, 0.0719, 0.1082, 0.3280, 0.1047, 0.1770, 0.2496], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0311, 0.0222, 0.0281, 0.0313, 0.0263, 0.0252, 0.0267], + device='cuda:3'), out_proj_covar=tensor([1.1538e-04, 1.2364e-04, 8.8360e-05, 1.1175e-04, 1.2749e-04, 1.0490e-04, + 1.0185e-04, 1.0631e-04], device='cuda:3') +2023-04-27 08:46:49,772 INFO [finetune.py:976] (3/7) Epoch 16, batch 3800, loss[loss=0.1614, simple_loss=0.2427, pruned_loss=0.04002, over 4886.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.253, pruned_loss=0.05626, over 954488.07 frames. ], batch size: 32, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:47:09,440 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89730.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:47:10,030 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89731.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:47:21,767 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89741.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:47:22,341 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3384, 1.7914, 1.8230, 2.1000, 1.9619, 2.0586, 1.6604, 4.4364], + device='cuda:3'), covar=tensor([0.0552, 0.0747, 0.0703, 0.1094, 0.0579, 0.0477, 0.0686, 0.0104], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 08:47:31,968 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.563e+02 1.915e+02 2.320e+02 3.976e+02, threshold=3.830e+02, percent-clipped=0.0 +2023-04-27 08:47:43,649 INFO [finetune.py:976] (3/7) Epoch 16, batch 3850, loss[loss=0.1844, simple_loss=0.2482, pruned_loss=0.06029, over 4899.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2522, pruned_loss=0.05615, over 955512.89 frames. ], batch size: 43, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:47:58,312 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89781.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:48:12,386 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89795.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:48:12,411 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4160, 3.0582, 0.8639, 1.6622, 1.8161, 2.2485, 1.8319, 0.9438], + device='cuda:3'), covar=tensor([0.1444, 0.1220, 0.2006, 0.1307, 0.1086, 0.0992, 0.1511, 0.1874], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0245, 0.0138, 0.0121, 0.0131, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:48:13,050 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8540, 2.4604, 1.8544, 1.7116, 1.4033, 1.4418, 1.9456, 1.3455], + device='cuda:3'), covar=tensor([0.1417, 0.1267, 0.1458, 0.1731, 0.2164, 0.1805, 0.0837, 0.1904], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0212, 0.0167, 0.0204, 0.0199, 0.0183, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 08:48:13,680 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0425, 2.1450, 1.8803, 1.8221, 2.3607, 1.8343, 2.8388, 1.5847], + device='cuda:3'), covar=tensor([0.3682, 0.1892, 0.4288, 0.2917, 0.1688, 0.2608, 0.1254, 0.4334], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0350, 0.0433, 0.0359, 0.0387, 0.0386, 0.0375, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:48:19,671 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3834, 1.4132, 4.0602, 3.7604, 3.5363, 3.8478, 3.7627, 3.5723], + device='cuda:3'), covar=tensor([0.7415, 0.5806, 0.1215, 0.1851, 0.1322, 0.1449, 0.1594, 0.1555], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0305, 0.0402, 0.0405, 0.0347, 0.0410, 0.0310, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:48:26,722 INFO [finetune.py:976] (3/7) Epoch 16, batch 3900, loss[loss=0.1504, simple_loss=0.2222, pruned_loss=0.03926, over 4874.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2502, pruned_loss=0.05575, over 955875.99 frames. ], batch size: 31, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:48:33,315 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89825.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:48:48,868 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.776e+01 1.757e+02 2.140e+02 2.503e+02 6.967e+02, threshold=4.281e+02, percent-clipped=3.0 +2023-04-27 08:48:50,219 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8241, 1.8418, 2.0880, 2.2650, 1.7133, 1.4701, 1.9409, 1.0810], + device='cuda:3'), covar=tensor([0.0830, 0.0967, 0.0609, 0.1008, 0.1013, 0.1263, 0.0820, 0.0860], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0076, 0.0097, 0.0075, 0.0068], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:48:59,088 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:48:59,123 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:48:59,622 INFO [finetune.py:976] (3/7) Epoch 16, batch 3950, loss[loss=0.1538, simple_loss=0.2238, pruned_loss=0.04193, over 4908.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.247, pruned_loss=0.05471, over 956579.21 frames. ], batch size: 35, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:49:05,394 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89873.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:49:06,553 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1948, 1.7021, 2.0675, 2.4269, 2.0497, 1.6637, 1.2320, 1.7016], + device='cuda:3'), covar=tensor([0.3373, 0.3346, 0.1724, 0.2346, 0.2665, 0.2831, 0.4443, 0.2096], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0247, 0.0225, 0.0315, 0.0217, 0.0230, 0.0228, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 08:49:19,519 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-04-27 08:49:33,477 INFO [finetune.py:976] (3/7) Epoch 16, batch 4000, loss[loss=0.1597, simple_loss=0.2212, pruned_loss=0.04913, over 4298.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2452, pruned_loss=0.05405, over 953169.64 frames. ], batch size: 18, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:49:47,327 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89926.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:50:09,494 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2616, 1.7343, 1.5823, 1.9909, 1.9806, 2.0469, 1.5560, 4.3406], + device='cuda:3'), covar=tensor([0.0573, 0.0784, 0.0795, 0.1214, 0.0622, 0.0575, 0.0756, 0.0108], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 08:50:11,835 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.722e+02 2.012e+02 2.405e+02 4.800e+02, threshold=4.024e+02, percent-clipped=2.0 +2023-04-27 08:50:33,227 INFO [finetune.py:976] (3/7) Epoch 16, batch 4050, loss[loss=0.1896, simple_loss=0.2638, pruned_loss=0.05767, over 4789.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2483, pruned_loss=0.05547, over 950289.21 frames. ], batch size: 29, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:51:08,133 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-27 08:51:40,297 INFO [finetune.py:976] (3/7) Epoch 16, batch 4100, loss[loss=0.2456, simple_loss=0.3, pruned_loss=0.09561, over 4746.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.251, pruned_loss=0.05572, over 951926.42 frames. ], batch size: 59, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:51:41,021 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0197, 2.5694, 2.1442, 2.2763, 1.7300, 2.1337, 2.2198, 1.6772], + device='cuda:3'), covar=tensor([0.1952, 0.1005, 0.0776, 0.1334, 0.3051, 0.1114, 0.1834, 0.2635], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0311, 0.0223, 0.0283, 0.0314, 0.0264, 0.0253, 0.0268], + device='cuda:3'), out_proj_covar=tensor([1.1587e-04, 1.2376e-04, 8.8943e-05, 1.1241e-04, 1.2783e-04, 1.0515e-04, + 1.0212e-04, 1.0648e-04], device='cuda:3') +2023-04-27 08:51:49,668 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7174, 2.0885, 1.7228, 1.4854, 1.2879, 1.3450, 1.7481, 1.2391], + device='cuda:3'), covar=tensor([0.1739, 0.1288, 0.1488, 0.1812, 0.2422, 0.2058, 0.1046, 0.2112], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0214, 0.0169, 0.0206, 0.0202, 0.0185, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 08:52:02,658 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90030.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:52:12,535 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:52:22,279 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0212, 2.5426, 0.8147, 1.5134, 1.4916, 1.9148, 1.5703, 0.8648], + device='cuda:3'), covar=tensor([0.1536, 0.1064, 0.1776, 0.1259, 0.1129, 0.0897, 0.1580, 0.1662], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0246, 0.0139, 0.0121, 0.0132, 0.0154, 0.0119, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 08:52:27,086 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.746e+02 2.004e+02 2.346e+02 5.282e+02, threshold=4.007e+02, percent-clipped=2.0 +2023-04-27 08:52:34,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7392, 2.1491, 1.9035, 1.9810, 1.5992, 1.7895, 1.8114, 1.3988], + device='cuda:3'), covar=tensor([0.1848, 0.1097, 0.0762, 0.1256, 0.3697, 0.1073, 0.1854, 0.2679], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0313, 0.0225, 0.0285, 0.0316, 0.0266, 0.0254, 0.0269], + device='cuda:3'), out_proj_covar=tensor([1.1654e-04, 1.2457e-04, 8.9455e-05, 1.1311e-04, 1.2875e-04, 1.0570e-04, + 1.0277e-04, 1.0716e-04], device='cuda:3') +2023-04-27 08:52:48,656 INFO [finetune.py:976] (3/7) Epoch 16, batch 4150, loss[loss=0.2341, simple_loss=0.3009, pruned_loss=0.08362, over 4818.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2526, pruned_loss=0.05662, over 951409.44 frames. ], batch size: 38, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:53:06,919 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:53:09,316 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90081.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:53:26,045 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90095.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:53:50,593 INFO [finetune.py:976] (3/7) Epoch 16, batch 4200, loss[loss=0.1578, simple_loss=0.2348, pruned_loss=0.04038, over 4767.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2525, pruned_loss=0.05627, over 950915.37 frames. ], batch size: 51, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:53:59,899 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:54:00,556 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90130.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:54:13,671 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8454, 1.1150, 3.2692, 3.0351, 2.9544, 3.2104, 3.1600, 2.8614], + device='cuda:3'), covar=tensor([0.7021, 0.5239, 0.1348, 0.2095, 0.1431, 0.1878, 0.1625, 0.1619], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0301, 0.0398, 0.0398, 0.0344, 0.0405, 0.0306, 0.0360], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:54:14,874 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90143.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:54:24,333 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.851e+01 1.635e+02 1.941e+02 2.339e+02 3.883e+02, threshold=3.882e+02, percent-clipped=0.0 +2023-04-27 08:54:34,088 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:54:34,189 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 08:54:34,610 INFO [finetune.py:976] (3/7) Epoch 16, batch 4250, loss[loss=0.1592, simple_loss=0.2297, pruned_loss=0.04438, over 4765.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2497, pruned_loss=0.05489, over 951653.85 frames. ], batch size: 28, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:54:52,884 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90191.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:55:06,210 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90213.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:55:08,443 INFO [finetune.py:976] (3/7) Epoch 16, batch 4300, loss[loss=0.1737, simple_loss=0.2368, pruned_loss=0.05533, over 4833.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2468, pruned_loss=0.05371, over 953211.41 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:55:11,525 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90221.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:55:19,275 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:55:22,818 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2569, 2.6359, 0.9783, 1.5801, 2.1334, 1.4649, 3.6380, 1.8764], + device='cuda:3'), covar=tensor([0.0627, 0.0676, 0.0848, 0.1246, 0.0499, 0.0940, 0.0191, 0.0608], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 08:55:32,139 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.571e+02 1.935e+02 2.216e+02 3.879e+02, threshold=3.870e+02, percent-clipped=0.0 +2023-04-27 08:55:41,849 INFO [finetune.py:976] (3/7) Epoch 16, batch 4350, loss[loss=0.1613, simple_loss=0.2267, pruned_loss=0.04791, over 4829.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.243, pruned_loss=0.05235, over 955060.14 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:56:00,593 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:56:11,324 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2009, 1.1916, 3.8610, 3.5647, 3.3929, 3.6925, 3.6789, 3.3404], + device='cuda:3'), covar=tensor([0.7474, 0.6022, 0.1187, 0.1899, 0.1295, 0.2097, 0.1567, 0.1636], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0300, 0.0398, 0.0398, 0.0343, 0.0403, 0.0305, 0.0359], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 08:56:15,530 INFO [finetune.py:976] (3/7) Epoch 16, batch 4400, loss[loss=0.1747, simple_loss=0.2557, pruned_loss=0.04689, over 4849.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2446, pruned_loss=0.05297, over 954361.84 frames. ], batch size: 47, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:56:16,818 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8710, 2.1747, 1.3936, 1.5671, 2.4550, 1.7511, 1.6412, 1.7837], + device='cuda:3'), covar=tensor([0.0495, 0.0347, 0.0286, 0.0558, 0.0227, 0.0511, 0.0511, 0.0543], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 08:56:34,938 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:56:55,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.655e+02 1.875e+02 2.353e+02 3.798e+02, threshold=3.751e+02, percent-clipped=0.0 +2023-04-27 08:56:56,116 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90351.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:57:10,217 INFO [finetune.py:976] (3/7) Epoch 16, batch 4450, loss[loss=0.2378, simple_loss=0.3011, pruned_loss=0.08723, over 4806.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2487, pruned_loss=0.05411, over 951917.05 frames. ], batch size: 45, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:57:28,322 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5177, 1.5356, 1.4198, 1.7853, 1.6959, 1.7314, 1.4175, 2.9242], + device='cuda:3'), covar=tensor([0.0466, 0.0641, 0.0695, 0.0997, 0.0514, 0.0377, 0.0632, 0.0204], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 08:57:31,125 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 08:57:33,655 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:57:39,600 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1420, 0.6583, 0.8982, 0.7606, 1.2623, 0.9452, 0.7949, 0.9586], + device='cuda:3'), covar=tensor([0.1799, 0.1780, 0.2240, 0.1724, 0.1246, 0.1510, 0.2043, 0.2556], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0311, 0.0351, 0.0290, 0.0327, 0.0310, 0.0301, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.3284e-05, 6.4892e-05, 7.4908e-05, 5.9133e-05, 6.8152e-05, 6.5423e-05, + 6.3633e-05, 7.7923e-05], device='cuda:3') +2023-04-27 08:58:04,796 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90412.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:58:07,139 INFO [finetune.py:976] (3/7) Epoch 16, batch 4500, loss[loss=0.1702, simple_loss=0.2472, pruned_loss=0.04655, over 4777.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2494, pruned_loss=0.05468, over 950171.78 frames. ], batch size: 51, lr: 3.43e-03, grad_scale: 64.0 +2023-04-27 08:58:52,084 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.589e+02 1.862e+02 2.271e+02 3.876e+02, threshold=3.723e+02, percent-clipped=1.0 +2023-04-27 08:59:04,453 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:59:14,175 INFO [finetune.py:976] (3/7) Epoch 16, batch 4550, loss[loss=0.1432, simple_loss=0.1982, pruned_loss=0.04414, over 4195.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2503, pruned_loss=0.05474, over 949549.06 frames. ], batch size: 18, lr: 3.43e-03, grad_scale: 64.0 +2023-04-27 08:59:37,652 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90486.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:59:55,974 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90496.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:00:15,724 INFO [finetune.py:976] (3/7) Epoch 16, batch 4600, loss[loss=0.1627, simple_loss=0.2344, pruned_loss=0.04553, over 4898.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2511, pruned_loss=0.05438, over 951674.82 frames. ], batch size: 43, lr: 3.43e-03, grad_scale: 64.0 +2023-04-27 09:00:18,339 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:00:18,936 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90521.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:00:37,739 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.675e+02 1.952e+02 2.329e+02 6.121e+02, threshold=3.905e+02, percent-clipped=3.0 +2023-04-27 09:00:43,135 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90557.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:00:49,342 INFO [finetune.py:976] (3/7) Epoch 16, batch 4650, loss[loss=0.2147, simple_loss=0.2618, pruned_loss=0.0838, over 4897.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.248, pruned_loss=0.05359, over 952524.95 frames. ], batch size: 32, lr: 3.42e-03, grad_scale: 64.0 +2023-04-27 09:00:51,267 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90569.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:01:03,241 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90588.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:01:23,252 INFO [finetune.py:976] (3/7) Epoch 16, batch 4700, loss[loss=0.163, simple_loss=0.2319, pruned_loss=0.04699, over 4819.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2465, pruned_loss=0.0539, over 953482.59 frames. ], batch size: 25, lr: 3.42e-03, grad_scale: 64.0 +2023-04-27 09:01:45,413 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.583e+02 1.845e+02 2.338e+02 5.124e+02, threshold=3.690e+02, percent-clipped=1.0 +2023-04-27 09:02:01,940 INFO [finetune.py:976] (3/7) Epoch 16, batch 4750, loss[loss=0.1744, simple_loss=0.2215, pruned_loss=0.06364, over 4737.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2458, pruned_loss=0.0541, over 955010.73 frames. ], batch size: 23, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:02:51,158 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:03:07,871 INFO [finetune.py:976] (3/7) Epoch 16, batch 4800, loss[loss=0.1725, simple_loss=0.2426, pruned_loss=0.05117, over 4895.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2479, pruned_loss=0.0551, over 953131.50 frames. ], batch size: 32, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:03:36,199 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.708e+02 2.005e+02 2.486e+02 4.014e+02, threshold=4.009e+02, percent-clipped=3.0 +2023-04-27 09:03:47,215 INFO [finetune.py:976] (3/7) Epoch 16, batch 4850, loss[loss=0.2719, simple_loss=0.3293, pruned_loss=0.1073, over 4197.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2505, pruned_loss=0.05544, over 953747.55 frames. ], batch size: 65, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:03:55,774 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.19 vs. limit=5.0 +2023-04-27 09:04:00,345 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90786.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:04:36,231 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90815.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:04:36,789 INFO [finetune.py:976] (3/7) Epoch 16, batch 4900, loss[loss=0.177, simple_loss=0.2544, pruned_loss=0.04976, over 4817.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2524, pruned_loss=0.05618, over 954946.55 frames. ], batch size: 38, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:05:01,144 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90834.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:05:21,475 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90849.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:05:22,606 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.194e+02 1.643e+02 1.978e+02 2.332e+02 3.633e+02, threshold=3.955e+02, percent-clipped=0.0 +2023-04-27 09:05:23,822 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:05:38,745 INFO [finetune.py:976] (3/7) Epoch 16, batch 4950, loss[loss=0.1865, simple_loss=0.2605, pruned_loss=0.05625, over 4783.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2533, pruned_loss=0.05643, over 953605.25 frames. ], batch size: 51, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:05:40,805 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-04-27 09:05:54,249 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:06:05,350 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 09:06:08,380 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90910.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:06:12,377 INFO [finetune.py:976] (3/7) Epoch 16, batch 5000, loss[loss=0.1507, simple_loss=0.2212, pruned_loss=0.04005, over 4717.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2504, pruned_loss=0.05496, over 954199.52 frames. ], batch size: 54, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:06:27,004 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:06:30,767 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4007, 1.2383, 1.5585, 1.5248, 1.2565, 1.1617, 1.2116, 0.9019], + device='cuda:3'), covar=tensor([0.0517, 0.0639, 0.0406, 0.0663, 0.0712, 0.1059, 0.0588, 0.0619], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0069, 0.0069, 0.0067, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:06:33,891 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.76 vs. limit=5.0 +2023-04-27 09:06:36,101 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.563e+02 1.794e+02 2.191e+02 3.401e+02, threshold=3.587e+02, percent-clipped=0.0 +2023-04-27 09:06:46,207 INFO [finetune.py:976] (3/7) Epoch 16, batch 5050, loss[loss=0.1518, simple_loss=0.2272, pruned_loss=0.03817, over 4823.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2479, pruned_loss=0.05413, over 954063.28 frames. ], batch size: 25, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:07:03,642 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-27 09:07:08,619 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 09:07:11,441 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5328, 2.1840, 2.4508, 2.9630, 2.8410, 2.3298, 2.0106, 2.6027], + device='cuda:3'), covar=tensor([0.0804, 0.1034, 0.0592, 0.0502, 0.0529, 0.0901, 0.0804, 0.0536], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0201, 0.0182, 0.0173, 0.0178, 0.0182, 0.0153, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:07:13,884 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91007.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:07:19,798 INFO [finetune.py:976] (3/7) Epoch 16, batch 5100, loss[loss=0.1497, simple_loss=0.22, pruned_loss=0.0397, over 4905.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2435, pruned_loss=0.05248, over 953881.56 frames. ], batch size: 32, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:07:21,162 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7039, 0.7377, 1.4859, 2.0424, 1.7395, 1.5592, 1.5435, 1.5912], + device='cuda:3'), covar=tensor([0.4468, 0.6428, 0.6278, 0.6190, 0.5981, 0.7146, 0.7403, 0.8375], + device='cuda:3'), in_proj_covar=tensor([0.0420, 0.0405, 0.0494, 0.0506, 0.0447, 0.0472, 0.0478, 0.0483], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:07:43,900 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.518e+02 1.785e+02 2.206e+02 4.900e+02, threshold=3.570e+02, percent-clipped=2.0 +2023-04-27 09:07:46,424 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:07:50,143 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3147, 2.1093, 1.8457, 1.8669, 2.1700, 1.8868, 2.4255, 1.6362], + device='cuda:3'), covar=tensor([0.2937, 0.1316, 0.3422, 0.2284, 0.1487, 0.1981, 0.1433, 0.3782], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0345, 0.0428, 0.0353, 0.0381, 0.0378, 0.0369, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:07:53,572 INFO [finetune.py:976] (3/7) Epoch 16, batch 5150, loss[loss=0.2153, simple_loss=0.2796, pruned_loss=0.07549, over 4923.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2457, pruned_loss=0.05368, over 954010.01 frames. ], batch size: 38, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:08:12,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8107, 1.4592, 1.9544, 2.2208, 1.8973, 1.7828, 1.8862, 1.8234], + device='cuda:3'), covar=tensor([0.4732, 0.6702, 0.6357, 0.5991, 0.5724, 0.7896, 0.8013, 0.9041], + device='cuda:3'), in_proj_covar=tensor([0.0419, 0.0404, 0.0492, 0.0505, 0.0447, 0.0470, 0.0477, 0.0482], + device='cuda:3'), out_proj_covar=tensor([1.0098e-04, 9.9966e-05, 1.1088e-04, 1.2034e-04, 1.0744e-04, 1.1317e-04, + 1.1318e-04, 1.1406e-04], device='cuda:3') +2023-04-27 09:08:46,667 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6752, 1.2125, 1.7344, 2.1047, 1.7563, 1.6229, 1.6950, 1.6403], + device='cuda:3'), covar=tensor([0.4762, 0.7078, 0.6910, 0.6598, 0.6286, 0.8547, 0.8845, 0.9109], + device='cuda:3'), in_proj_covar=tensor([0.0420, 0.0405, 0.0493, 0.0506, 0.0447, 0.0471, 0.0477, 0.0482], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:08:48,452 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:08:48,975 INFO [finetune.py:976] (3/7) Epoch 16, batch 5200, loss[loss=0.1652, simple_loss=0.2465, pruned_loss=0.04192, over 4897.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2491, pruned_loss=0.05482, over 954961.38 frames. ], batch size: 35, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:09:38,602 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.654e+02 2.065e+02 2.493e+02 5.806e+02, threshold=4.130e+02, percent-clipped=4.0 +2023-04-27 09:09:39,326 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91152.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:09:51,505 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91163.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:09:53,288 INFO [finetune.py:976] (3/7) Epoch 16, batch 5250, loss[loss=0.1508, simple_loss=0.2178, pruned_loss=0.04196, over 4698.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2526, pruned_loss=0.05626, over 954432.45 frames. ], batch size: 23, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:10:44,542 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91200.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:10:47,634 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91205.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:11:00,424 INFO [finetune.py:976] (3/7) Epoch 16, batch 5300, loss[loss=0.2053, simple_loss=0.2647, pruned_loss=0.07295, over 4892.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2539, pruned_loss=0.05681, over 954775.07 frames. ], batch size: 32, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:11:10,772 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8501, 1.4410, 1.3956, 1.6357, 2.0632, 1.6608, 1.3768, 1.3539], + device='cuda:3'), covar=tensor([0.1572, 0.1419, 0.1647, 0.1285, 0.0718, 0.1483, 0.1804, 0.1940], + device='cuda:3'), in_proj_covar=tensor([0.0304, 0.0309, 0.0348, 0.0285, 0.0324, 0.0310, 0.0299, 0.0361], + device='cuda:3'), out_proj_covar=tensor([6.2771e-05, 6.4377e-05, 7.4233e-05, 5.7977e-05, 6.7381e-05, 6.5304e-05, + 6.3173e-05, 7.7096e-05], device='cuda:3') +2023-04-27 09:11:25,281 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.661e+02 1.899e+02 2.252e+02 3.532e+02, threshold=3.799e+02, percent-clipped=0.0 +2023-04-27 09:11:34,433 INFO [finetune.py:976] (3/7) Epoch 16, batch 5350, loss[loss=0.1757, simple_loss=0.2485, pruned_loss=0.05142, over 4814.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2535, pruned_loss=0.05641, over 954425.57 frames. ], batch size: 39, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:11:38,910 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-27 09:11:45,881 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6355, 1.7193, 0.7239, 1.3043, 1.8403, 1.5131, 1.4037, 1.4247], + device='cuda:3'), covar=tensor([0.0476, 0.0363, 0.0372, 0.0550, 0.0289, 0.0505, 0.0490, 0.0534], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 09:11:47,122 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91285.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:12:08,307 INFO [finetune.py:976] (3/7) Epoch 16, batch 5400, loss[loss=0.1927, simple_loss=0.2643, pruned_loss=0.06048, over 4929.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2517, pruned_loss=0.05551, over 955172.10 frames. ], batch size: 38, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:12:28,660 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91346.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:12:32,094 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.605e+02 2.003e+02 2.320e+02 6.253e+02, threshold=4.007e+02, percent-clipped=1.0 +2023-04-27 09:12:42,145 INFO [finetune.py:976] (3/7) Epoch 16, batch 5450, loss[loss=0.2022, simple_loss=0.2666, pruned_loss=0.06888, over 4826.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2487, pruned_loss=0.05493, over 955750.77 frames. ], batch size: 40, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:12:51,534 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 09:13:15,970 INFO [finetune.py:976] (3/7) Epoch 16, batch 5500, loss[loss=0.1748, simple_loss=0.2372, pruned_loss=0.05619, over 4908.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2455, pruned_loss=0.05396, over 953149.96 frames. ], batch size: 32, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:13:33,045 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1733, 1.6899, 2.0506, 2.5623, 2.0670, 1.5718, 1.4440, 1.8720], + device='cuda:3'), covar=tensor([0.3206, 0.3119, 0.1681, 0.2026, 0.2511, 0.2600, 0.4100, 0.1952], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0248, 0.0228, 0.0316, 0.0218, 0.0232, 0.0229, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 09:13:37,833 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8451, 1.3069, 1.8728, 2.3078, 1.9443, 1.7774, 1.7914, 1.7868], + device='cuda:3'), covar=tensor([0.4877, 0.7011, 0.6684, 0.6109, 0.6114, 0.7953, 0.8056, 0.9008], + device='cuda:3'), in_proj_covar=tensor([0.0419, 0.0407, 0.0494, 0.0505, 0.0448, 0.0471, 0.0477, 0.0483], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:13:38,282 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.525e+01 1.617e+02 1.882e+02 2.275e+02 4.330e+02, threshold=3.765e+02, percent-clipped=1.0 +2023-04-27 09:13:45,258 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9289, 1.8946, 2.3214, 2.4642, 1.8046, 1.6403, 2.0124, 1.1650], + device='cuda:3'), covar=tensor([0.0603, 0.0624, 0.0469, 0.0561, 0.0720, 0.1105, 0.0608, 0.0768], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0075, 0.0097, 0.0075, 0.0068], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:13:49,959 INFO [finetune.py:976] (3/7) Epoch 16, batch 5550, loss[loss=0.1776, simple_loss=0.2483, pruned_loss=0.05343, over 4897.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2467, pruned_loss=0.05441, over 953655.95 frames. ], batch size: 43, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:13:50,095 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2575, 1.5655, 1.4402, 2.0964, 2.3829, 1.8149, 1.8489, 1.5872], + device='cuda:3'), covar=tensor([0.1712, 0.1989, 0.2085, 0.1719, 0.1079, 0.2234, 0.2276, 0.2426], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0312, 0.0351, 0.0287, 0.0326, 0.0311, 0.0301, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.3286e-05, 6.5018e-05, 7.4823e-05, 5.8437e-05, 6.7914e-05, 6.5688e-05, + 6.3647e-05, 7.7774e-05], device='cuda:3') +2023-04-27 09:14:10,030 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2746, 1.1737, 1.3183, 1.5475, 1.5338, 1.2523, 0.9143, 1.3941], + device='cuda:3'), covar=tensor([0.0909, 0.1352, 0.0845, 0.0650, 0.0672, 0.0813, 0.0889, 0.0636], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0201, 0.0180, 0.0172, 0.0176, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:14:20,588 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91505.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:14:32,820 INFO [finetune.py:976] (3/7) Epoch 16, batch 5600, loss[loss=0.1969, simple_loss=0.2611, pruned_loss=0.06633, over 4934.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2499, pruned_loss=0.05486, over 954111.87 frames. ], batch size: 33, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:15:04,841 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91542.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:15:15,398 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.634e+02 1.984e+02 2.410e+02 6.572e+02, threshold=3.968e+02, percent-clipped=2.0 +2023-04-27 09:15:16,632 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91553.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:15:29,904 INFO [finetune.py:976] (3/7) Epoch 16, batch 5650, loss[loss=0.1552, simple_loss=0.2289, pruned_loss=0.04074, over 4775.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2517, pruned_loss=0.05529, over 952316.96 frames. ], batch size: 26, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:16:14,756 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5790, 1.0628, 1.3854, 1.1779, 1.7130, 1.4714, 1.2057, 1.3392], + device='cuda:3'), covar=tensor([0.2328, 0.2095, 0.2771, 0.1939, 0.1129, 0.1622, 0.2650, 0.3295], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0312, 0.0349, 0.0287, 0.0324, 0.0310, 0.0301, 0.0363], + device='cuda:3'), out_proj_covar=tensor([6.3186e-05, 6.5033e-05, 7.4583e-05, 5.8310e-05, 6.7405e-05, 6.5329e-05, + 6.3511e-05, 7.7607e-05], device='cuda:3') +2023-04-27 09:16:20,689 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91603.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:16:34,226 INFO [finetune.py:976] (3/7) Epoch 16, batch 5700, loss[loss=0.1481, simple_loss=0.2072, pruned_loss=0.04448, over 4191.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2478, pruned_loss=0.0544, over 933848.48 frames. ], batch size: 18, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:16:46,633 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8586, 1.1717, 1.7368, 2.2295, 1.9294, 1.7712, 1.7819, 1.7775], + device='cuda:3'), covar=tensor([0.4312, 0.6339, 0.6095, 0.5789, 0.5886, 0.7597, 0.7928, 0.7350], + device='cuda:3'), in_proj_covar=tensor([0.0420, 0.0409, 0.0496, 0.0507, 0.0449, 0.0472, 0.0479, 0.0485], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:17:06,021 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91641.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:17:19,303 INFO [finetune.py:976] (3/7) Epoch 17, batch 0, loss[loss=0.2368, simple_loss=0.3017, pruned_loss=0.08599, over 4922.00 frames. ], tot_loss[loss=0.2368, simple_loss=0.3017, pruned_loss=0.08599, over 4922.00 frames. ], batch size: 42, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:17:19,303 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 09:17:40,815 INFO [finetune.py:1010] (3/7) Epoch 17, validation: loss=0.1535, simple_loss=0.2247, pruned_loss=0.04111, over 2265189.00 frames. +2023-04-27 09:17:40,815 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 09:17:45,698 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.575e+02 1.838e+02 2.202e+02 3.811e+02, threshold=3.676e+02, percent-clipped=0.0 +2023-04-27 09:18:17,722 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7271, 2.0773, 1.1805, 1.4679, 2.0367, 1.6524, 1.5107, 1.6618], + device='cuda:3'), covar=tensor([0.0480, 0.0332, 0.0301, 0.0538, 0.0253, 0.0506, 0.0513, 0.0531], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 09:18:19,483 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 09:18:24,712 INFO [finetune.py:976] (3/7) Epoch 17, batch 50, loss[loss=0.2018, simple_loss=0.2603, pruned_loss=0.07159, over 4923.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.253, pruned_loss=0.05643, over 216128.56 frames. ], batch size: 38, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:18:38,157 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3039, 2.8773, 2.3944, 2.7041, 2.0389, 2.3962, 2.6653, 1.7574], + device='cuda:3'), covar=tensor([0.2213, 0.1393, 0.0844, 0.1404, 0.3171, 0.1325, 0.2134, 0.3082], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0309, 0.0221, 0.0280, 0.0314, 0.0261, 0.0252, 0.0266], + device='cuda:3'), out_proj_covar=tensor([1.1527e-04, 1.2286e-04, 8.8165e-05, 1.1105e-04, 1.2781e-04, 1.0379e-04, + 1.0174e-04, 1.0593e-04], device='cuda:3') +2023-04-27 09:18:57,080 INFO [finetune.py:976] (3/7) Epoch 17, batch 100, loss[loss=0.1652, simple_loss=0.2326, pruned_loss=0.04891, over 4891.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2448, pruned_loss=0.05416, over 379846.52 frames. ], batch size: 35, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:19:02,435 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.653e+02 1.974e+02 2.303e+02 4.426e+02, threshold=3.948e+02, percent-clipped=2.0 +2023-04-27 09:19:30,053 INFO [finetune.py:976] (3/7) Epoch 17, batch 150, loss[loss=0.195, simple_loss=0.2602, pruned_loss=0.06492, over 4816.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2415, pruned_loss=0.05235, over 508303.73 frames. ], batch size: 41, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:20:03,533 INFO [finetune.py:976] (3/7) Epoch 17, batch 200, loss[loss=0.1839, simple_loss=0.2471, pruned_loss=0.06034, over 4771.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2405, pruned_loss=0.05203, over 607960.41 frames. ], batch size: 29, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:20:04,225 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4803, 1.7084, 1.6072, 1.8831, 1.7807, 2.0618, 1.5560, 3.2327], + device='cuda:3'), covar=tensor([0.0517, 0.0621, 0.0698, 0.1036, 0.0560, 0.0565, 0.0700, 0.0173], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 09:20:08,957 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.596e+01 1.677e+02 1.899e+02 2.270e+02 6.599e+02, threshold=3.797e+02, percent-clipped=2.0 +2023-04-27 09:20:35,886 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91891.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:20:36,989 INFO [finetune.py:976] (3/7) Epoch 17, batch 250, loss[loss=0.2189, simple_loss=0.2856, pruned_loss=0.07614, over 4903.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2443, pruned_loss=0.05374, over 682419.00 frames. ], batch size: 37, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:20:40,159 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91898.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 09:20:41,410 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3809, 1.3167, 1.7547, 1.6647, 1.2776, 1.1744, 1.3067, 0.7979], + device='cuda:3'), covar=tensor([0.0560, 0.0687, 0.0370, 0.0609, 0.0745, 0.1108, 0.0653, 0.0695], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0069, 0.0068, 0.0067, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:21:05,107 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4630, 2.0349, 2.4467, 2.7859, 2.8187, 2.3286, 2.0140, 2.5115], + device='cuda:3'), covar=tensor([0.0880, 0.1181, 0.0669, 0.0644, 0.0629, 0.0956, 0.0812, 0.0648], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0200, 0.0181, 0.0172, 0.0176, 0.0181, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:21:08,452 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:21:10,057 INFO [finetune.py:976] (3/7) Epoch 17, batch 300, loss[loss=0.205, simple_loss=0.2725, pruned_loss=0.06874, over 4777.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2484, pruned_loss=0.05466, over 743383.75 frames. ], batch size: 29, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:21:14,187 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 09:21:15,846 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.651e+02 1.943e+02 2.275e+02 4.588e+02, threshold=3.886e+02, percent-clipped=2.0 +2023-04-27 09:21:16,619 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91952.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:21:17,783 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5593, 3.4806, 2.7069, 4.1619, 3.6123, 3.6115, 1.4517, 3.5718], + device='cuda:3'), covar=tensor([0.1973, 0.1320, 0.3227, 0.1864, 0.3553, 0.1876, 0.6081, 0.2644], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0214, 0.0253, 0.0306, 0.0300, 0.0250, 0.0275, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 09:21:18,452 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91955.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:21:30,543 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7399, 1.9680, 0.9239, 1.4279, 1.8951, 1.6252, 1.4905, 1.5781], + device='cuda:3'), covar=tensor([0.0505, 0.0351, 0.0359, 0.0570, 0.0261, 0.0536, 0.0510, 0.0589], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 09:21:41,951 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4974, 1.3623, 1.8732, 1.7603, 1.3364, 1.2026, 1.4756, 0.9589], + device='cuda:3'), covar=tensor([0.0492, 0.0679, 0.0365, 0.0550, 0.0732, 0.1192, 0.0535, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0069, 0.0068, 0.0067, 0.0074, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:21:43,852 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-04-27 09:21:51,281 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91981.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:21:56,126 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:22:03,159 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 09:22:04,086 INFO [finetune.py:976] (3/7) Epoch 17, batch 350, loss[loss=0.1913, simple_loss=0.2569, pruned_loss=0.06284, over 4808.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2501, pruned_loss=0.05553, over 791176.14 frames. ], batch size: 39, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:22:36,922 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92016.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:22:57,034 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8553, 3.7069, 2.8809, 4.4680, 3.9027, 3.8499, 1.5309, 3.8349], + device='cuda:3'), covar=tensor([0.1709, 0.1361, 0.3208, 0.1371, 0.2940, 0.1746, 0.5973, 0.2489], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0213, 0.0252, 0.0306, 0.0299, 0.0249, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 09:23:09,174 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92042.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 09:23:09,650 INFO [finetune.py:976] (3/7) Epoch 17, batch 400, loss[loss=0.1553, simple_loss=0.2347, pruned_loss=0.03801, over 4751.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2503, pruned_loss=0.05479, over 827837.18 frames. ], batch size: 54, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:23:21,384 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.614e+02 1.906e+02 2.282e+02 3.471e+02, threshold=3.811e+02, percent-clipped=0.0 +2023-04-27 09:23:31,663 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 09:24:03,171 INFO [finetune.py:976] (3/7) Epoch 17, batch 450, loss[loss=0.188, simple_loss=0.2546, pruned_loss=0.06071, over 4859.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2491, pruned_loss=0.05384, over 857750.14 frames. ], batch size: 31, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:24:36,847 INFO [finetune.py:976] (3/7) Epoch 17, batch 500, loss[loss=0.1709, simple_loss=0.2322, pruned_loss=0.05482, over 4389.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2476, pruned_loss=0.05372, over 880696.35 frames. ], batch size: 65, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:24:42,152 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.524e+02 1.904e+02 2.252e+02 3.809e+02, threshold=3.808e+02, percent-clipped=0.0 +2023-04-27 09:25:10,282 INFO [finetune.py:976] (3/7) Epoch 17, batch 550, loss[loss=0.2023, simple_loss=0.2727, pruned_loss=0.06595, over 4813.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2459, pruned_loss=0.05368, over 897664.79 frames. ], batch size: 41, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:25:13,460 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92198.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:25:44,127 INFO [finetune.py:976] (3/7) Epoch 17, batch 600, loss[loss=0.1879, simple_loss=0.2619, pruned_loss=0.05691, over 4917.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2473, pruned_loss=0.05449, over 910426.54 frames. ], batch size: 43, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:25:46,054 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92246.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:25:46,683 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92247.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:25:49,033 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.221e+02 1.690e+02 1.979e+02 2.477e+02 6.011e+02, threshold=3.959e+02, percent-clipped=1.0 +2023-04-27 09:26:17,389 INFO [finetune.py:976] (3/7) Epoch 17, batch 650, loss[loss=0.1283, simple_loss=0.2094, pruned_loss=0.0236, over 4776.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2485, pruned_loss=0.05444, over 917262.33 frames. ], batch size: 26, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:26:29,965 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92311.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:26:56,503 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92337.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 09:27:00,059 INFO [finetune.py:976] (3/7) Epoch 17, batch 700, loss[loss=0.1516, simple_loss=0.2274, pruned_loss=0.03791, over 4856.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2491, pruned_loss=0.0538, over 925060.67 frames. ], batch size: 31, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:27:10,670 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.019e+02 1.654e+02 1.859e+02 2.105e+02 3.715e+02, threshold=3.718e+02, percent-clipped=1.0 +2023-04-27 09:27:49,356 INFO [finetune.py:976] (3/7) Epoch 17, batch 750, loss[loss=0.2199, simple_loss=0.299, pruned_loss=0.07038, over 4813.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2501, pruned_loss=0.05362, over 931569.01 frames. ], batch size: 39, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:28:44,568 INFO [finetune.py:976] (3/7) Epoch 17, batch 800, loss[loss=0.1333, simple_loss=0.1969, pruned_loss=0.03486, over 4047.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2496, pruned_loss=0.05327, over 937227.95 frames. ], batch size: 17, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:28:54,788 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.529e+02 1.784e+02 2.077e+02 4.006e+02, threshold=3.568e+02, percent-clipped=1.0 +2023-04-27 09:29:31,910 INFO [finetune.py:976] (3/7) Epoch 17, batch 850, loss[loss=0.1807, simple_loss=0.2453, pruned_loss=0.05807, over 4822.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2482, pruned_loss=0.05336, over 940722.53 frames. ], batch size: 40, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:30:05,439 INFO [finetune.py:976] (3/7) Epoch 17, batch 900, loss[loss=0.1354, simple_loss=0.2131, pruned_loss=0.0289, over 4765.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2453, pruned_loss=0.05249, over 943795.98 frames. ], batch size: 26, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:30:07,970 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92547.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:30:10,296 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.562e+01 1.556e+02 1.849e+02 2.275e+02 6.056e+02, threshold=3.698e+02, percent-clipped=3.0 +2023-04-27 09:30:38,497 INFO [finetune.py:976] (3/7) Epoch 17, batch 950, loss[loss=0.1849, simple_loss=0.2681, pruned_loss=0.0509, over 4824.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2446, pruned_loss=0.05252, over 946907.24 frames. ], batch size: 40, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:30:39,830 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92595.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:30:49,712 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92611.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:30:59,799 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:31:07,676 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92637.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 09:31:11,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9729, 1.9302, 2.2974, 2.4977, 1.8440, 1.6359, 2.0949, 1.1145], + device='cuda:3'), covar=tensor([0.0641, 0.0802, 0.0529, 0.0692, 0.0695, 0.1151, 0.0640, 0.0852], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:31:12,293 INFO [finetune.py:976] (3/7) Epoch 17, batch 1000, loss[loss=0.1897, simple_loss=0.2588, pruned_loss=0.0603, over 4905.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2462, pruned_loss=0.05289, over 948566.64 frames. ], batch size: 36, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:31:17,228 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.750e+01 1.761e+02 2.088e+02 2.403e+02 5.880e+02, threshold=4.175e+02, percent-clipped=3.0 +2023-04-27 09:31:22,263 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:31:27,272 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0214, 2.5786, 1.0906, 1.4235, 1.9989, 1.2489, 3.5559, 1.7841], + device='cuda:3'), covar=tensor([0.0749, 0.0674, 0.0738, 0.1311, 0.0527, 0.1071, 0.0243, 0.0649], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 09:31:40,697 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92685.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:31:41,330 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92686.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:31:45,966 INFO [finetune.py:976] (3/7) Epoch 17, batch 1050, loss[loss=0.2084, simple_loss=0.2722, pruned_loss=0.07225, over 4899.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2499, pruned_loss=0.05362, over 950636.44 frames. ], batch size: 35, lr: 3.40e-03, grad_scale: 64.0 +2023-04-27 09:31:47,252 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-27 09:32:03,096 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:32:26,710 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8985, 1.1068, 3.2443, 3.0067, 2.9344, 3.1685, 3.1601, 2.8682], + device='cuda:3'), covar=tensor([0.7024, 0.5586, 0.1504, 0.2258, 0.1508, 0.1623, 0.1910, 0.1670], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0304, 0.0399, 0.0402, 0.0346, 0.0406, 0.0307, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:32:30,324 INFO [finetune.py:976] (3/7) Epoch 17, batch 1100, loss[loss=0.1953, simple_loss=0.2657, pruned_loss=0.06244, over 4778.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2508, pruned_loss=0.05396, over 951737.65 frames. ], batch size: 51, lr: 3.40e-03, grad_scale: 64.0 +2023-04-27 09:32:36,189 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.706e+02 1.976e+02 2.311e+02 4.775e+02, threshold=3.952e+02, percent-clipped=2.0 +2023-04-27 09:32:47,765 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 09:33:03,800 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 09:33:11,524 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92781.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:33:25,331 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-27 09:33:32,087 INFO [finetune.py:976] (3/7) Epoch 17, batch 1150, loss[loss=0.1685, simple_loss=0.2387, pruned_loss=0.04918, over 4813.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2517, pruned_loss=0.05393, over 953577.44 frames. ], batch size: 25, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:34:20,061 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92831.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:34:39,530 INFO [finetune.py:976] (3/7) Epoch 17, batch 1200, loss[loss=0.1643, simple_loss=0.234, pruned_loss=0.04732, over 4786.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2501, pruned_loss=0.05389, over 954304.81 frames. ], batch size: 51, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:34:50,260 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.173e+01 1.694e+02 1.853e+02 2.185e+02 5.052e+02, threshold=3.707e+02, percent-clipped=2.0 +2023-04-27 09:35:44,786 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92892.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:35:45,254 INFO [finetune.py:976] (3/7) Epoch 17, batch 1250, loss[loss=0.2139, simple_loss=0.2781, pruned_loss=0.0749, over 4918.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2474, pruned_loss=0.05286, over 954303.55 frames. ], batch size: 43, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:35:48,975 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92899.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:36:06,395 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6442, 1.2607, 4.5029, 4.2094, 3.9742, 4.3465, 4.2475, 3.9371], + device='cuda:3'), covar=tensor([0.6973, 0.6793, 0.1149, 0.1790, 0.1090, 0.2125, 0.1360, 0.1779], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0304, 0.0400, 0.0401, 0.0346, 0.0405, 0.0308, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:36:25,937 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92931.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:36:45,900 INFO [finetune.py:976] (3/7) Epoch 17, batch 1300, loss[loss=0.1755, simple_loss=0.238, pruned_loss=0.05646, over 4908.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2452, pruned_loss=0.05284, over 954702.37 frames. ], batch size: 43, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:36:46,011 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2129, 2.1557, 1.8828, 1.9038, 2.2369, 1.8294, 2.8338, 1.6075], + device='cuda:3'), covar=tensor([0.3627, 0.2229, 0.4571, 0.2951, 0.1714, 0.2486, 0.1144, 0.4602], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0352, 0.0437, 0.0362, 0.0389, 0.0387, 0.0376, 0.0430], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:36:57,406 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 1.510e+02 1.807e+02 2.085e+02 3.413e+02, threshold=3.613e+02, percent-clipped=0.0 +2023-04-27 09:37:08,826 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92960.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:11,893 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6931, 1.5149, 2.0240, 2.0534, 1.5236, 1.3400, 1.7270, 1.1060], + device='cuda:3'), covar=tensor([0.0489, 0.0834, 0.0359, 0.0536, 0.0716, 0.1143, 0.0640, 0.0749], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0069, 0.0068, 0.0067, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:37:23,988 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92975.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:33,569 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:40,825 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92992.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:41,317 INFO [finetune.py:976] (3/7) Epoch 17, batch 1350, loss[loss=0.1699, simple_loss=0.2315, pruned_loss=0.05422, over 4899.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2453, pruned_loss=0.05312, over 954463.85 frames. ], batch size: 32, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:37:54,394 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93010.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:38:13,981 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 09:38:16,120 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:38:16,134 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:38:26,347 INFO [finetune.py:976] (3/7) Epoch 17, batch 1400, loss[loss=0.1474, simple_loss=0.2098, pruned_loss=0.04252, over 4231.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2502, pruned_loss=0.05479, over 955205.17 frames. ], batch size: 18, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:38:44,541 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.602e+02 1.954e+02 2.219e+02 5.640e+02, threshold=3.909e+02, percent-clipped=3.0 +2023-04-27 09:38:46,592 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-04-27 09:39:06,942 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6105, 1.3686, 0.5950, 1.2606, 1.4345, 1.4627, 1.3535, 1.3745], + device='cuda:3'), covar=tensor([0.0504, 0.0406, 0.0420, 0.0603, 0.0314, 0.0572, 0.0578, 0.0573], + device='cuda:3'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0051], + device='cuda:3') +2023-04-27 09:39:08,301 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 09:39:08,813 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93071.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:39:12,769 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93076.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:39:28,195 INFO [finetune.py:976] (3/7) Epoch 17, batch 1450, loss[loss=0.1902, simple_loss=0.2541, pruned_loss=0.06313, over 4916.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.252, pruned_loss=0.05493, over 954407.49 frames. ], batch size: 37, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:39:37,784 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93097.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:40:01,749 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:40:20,814 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 09:40:24,592 INFO [finetune.py:976] (3/7) Epoch 17, batch 1500, loss[loss=0.1883, simple_loss=0.2776, pruned_loss=0.04954, over 4901.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2521, pruned_loss=0.05453, over 955172.45 frames. ], batch size: 37, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:40:31,494 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.667e+02 1.959e+02 2.406e+02 7.498e+02, threshold=3.919e+02, percent-clipped=4.0 +2023-04-27 09:40:54,018 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:41:06,568 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93187.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:41:15,565 INFO [finetune.py:976] (3/7) Epoch 17, batch 1550, loss[loss=0.2026, simple_loss=0.2721, pruned_loss=0.06654, over 4828.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.252, pruned_loss=0.0543, over 955756.70 frames. ], batch size: 39, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:41:49,255 INFO [finetune.py:976] (3/7) Epoch 17, batch 1600, loss[loss=0.1641, simple_loss=0.2315, pruned_loss=0.04837, over 4860.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2486, pruned_loss=0.0533, over 956100.04 frames. ], batch size: 31, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:41:54,716 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.590e+02 1.909e+02 2.207e+02 5.345e+02, threshold=3.818e+02, percent-clipped=1.0 +2023-04-27 09:41:57,605 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93255.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:42:15,827 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93281.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:42:19,487 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93287.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:42:23,068 INFO [finetune.py:976] (3/7) Epoch 17, batch 1650, loss[loss=0.1669, simple_loss=0.2361, pruned_loss=0.04886, over 4902.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2447, pruned_loss=0.05178, over 957866.31 frames. ], batch size: 32, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:42:27,559 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 09:42:41,750 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7868, 1.2567, 3.2709, 3.0229, 2.9533, 3.2427, 3.2152, 2.8674], + device='cuda:3'), covar=tensor([0.7358, 0.5205, 0.1432, 0.2130, 0.1454, 0.2020, 0.1768, 0.1743], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0304, 0.0400, 0.0402, 0.0347, 0.0406, 0.0309, 0.0362], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:43:03,620 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:04,871 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93331.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:14,707 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1102, 1.6409, 1.9954, 2.4497, 2.0272, 1.5818, 1.4006, 1.8286], + device='cuda:3'), covar=tensor([0.3185, 0.3131, 0.1654, 0.2251, 0.2508, 0.2691, 0.4332, 0.2155], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0245, 0.0224, 0.0313, 0.0216, 0.0230, 0.0228, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 09:43:17,624 INFO [finetune.py:976] (3/7) Epoch 17, batch 1700, loss[loss=0.1813, simple_loss=0.2427, pruned_loss=0.05999, over 4848.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2432, pruned_loss=0.05145, over 958006.16 frames. ], batch size: 49, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:43:23,104 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.559e+02 1.851e+02 2.220e+02 4.230e+02, threshold=3.703e+02, percent-clipped=1.0 +2023-04-27 09:43:33,691 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93366.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:41,258 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93376.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:44,303 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6548, 2.7027, 2.2422, 2.4576, 2.7044, 2.3860, 3.6643, 2.0171], + device='cuda:3'), covar=tensor([0.3774, 0.2374, 0.4542, 0.3380, 0.1934, 0.2624, 0.1813, 0.4533], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0347, 0.0431, 0.0357, 0.0383, 0.0384, 0.0370, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:43:51,042 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93392.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:43:51,577 INFO [finetune.py:976] (3/7) Epoch 17, batch 1750, loss[loss=0.2257, simple_loss=0.3041, pruned_loss=0.07359, over 4028.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2458, pruned_loss=0.05295, over 955088.40 frames. ], batch size: 65, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:44:13,255 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93424.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:44:34,423 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3663, 1.5349, 1.4185, 1.8576, 1.7298, 2.0264, 1.3854, 3.7805], + device='cuda:3'), covar=tensor([0.0576, 0.0821, 0.0847, 0.1155, 0.0658, 0.0480, 0.0817, 0.0131], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0037, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 09:44:36,171 INFO [finetune.py:976] (3/7) Epoch 17, batch 1800, loss[loss=0.1952, simple_loss=0.2637, pruned_loss=0.06336, over 4789.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2484, pruned_loss=0.05346, over 954182.98 frames. ], batch size: 29, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:44:47,780 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.257e+02 1.710e+02 1.900e+02 2.209e+02 3.616e+02, threshold=3.799e+02, percent-clipped=0.0 +2023-04-27 09:44:50,281 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9573, 1.3289, 4.9380, 4.6345, 4.3227, 4.6294, 4.4393, 4.3806], + device='cuda:3'), covar=tensor([0.6748, 0.6105, 0.0924, 0.1730, 0.1063, 0.1386, 0.1319, 0.1452], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0305, 0.0401, 0.0405, 0.0349, 0.0408, 0.0310, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:44:57,294 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1676, 2.4826, 1.0294, 1.3567, 1.9377, 1.3126, 3.4725, 1.6754], + device='cuda:3'), covar=tensor([0.0696, 0.0666, 0.0786, 0.1275, 0.0540, 0.1001, 0.0290, 0.0679], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0049, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 09:45:05,932 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:45:22,739 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93487.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:45:26,250 INFO [finetune.py:976] (3/7) Epoch 17, batch 1850, loss[loss=0.1442, simple_loss=0.2093, pruned_loss=0.03956, over 4722.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2487, pruned_loss=0.05327, over 954092.84 frames. ], batch size: 23, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:46:26,180 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8551, 1.5623, 3.9801, 3.7301, 3.5707, 3.6742, 3.6212, 3.5688], + device='cuda:3'), covar=tensor([0.6271, 0.5413, 0.1075, 0.1705, 0.1127, 0.1603, 0.3091, 0.1443], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0306, 0.0401, 0.0405, 0.0348, 0.0408, 0.0310, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:46:27,191 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93535.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:46:37,236 INFO [finetune.py:976] (3/7) Epoch 17, batch 1900, loss[loss=0.2336, simple_loss=0.2973, pruned_loss=0.08499, over 4138.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2497, pruned_loss=0.05341, over 952816.33 frames. ], batch size: 65, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:46:42,806 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.573e+02 1.862e+02 2.216e+02 4.322e+02, threshold=3.725e+02, percent-clipped=2.0 +2023-04-27 09:46:44,714 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:06,560 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93587.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:10,125 INFO [finetune.py:976] (3/7) Epoch 17, batch 1950, loss[loss=0.1498, simple_loss=0.2283, pruned_loss=0.03559, over 4786.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2489, pruned_loss=0.0532, over 953010.75 frames. ], batch size: 26, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:47:16,346 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93603.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:35,237 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93631.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:37,612 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93635.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:48,925 INFO [finetune.py:976] (3/7) Epoch 17, batch 2000, loss[loss=0.1702, simple_loss=0.2391, pruned_loss=0.05062, over 4926.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2463, pruned_loss=0.05268, over 953848.69 frames. ], batch size: 33, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:47:54,466 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.579e+02 1.885e+02 2.263e+02 4.038e+02, threshold=3.769e+02, percent-clipped=1.0 +2023-04-27 09:47:58,873 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6943, 4.1085, 0.7212, 2.2877, 2.1711, 2.4882, 2.4698, 1.0308], + device='cuda:3'), covar=tensor([0.1405, 0.0878, 0.2160, 0.1306, 0.1063, 0.1304, 0.1376, 0.2083], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0242, 0.0137, 0.0121, 0.0132, 0.0153, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:48:03,150 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93666.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:11,986 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:16,588 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4757, 1.3276, 4.0442, 3.8008, 3.5368, 3.7708, 3.7673, 3.5899], + device='cuda:3'), covar=tensor([0.6708, 0.5563, 0.0978, 0.1443, 0.1182, 0.1823, 0.1713, 0.1468], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0302, 0.0395, 0.0400, 0.0343, 0.0403, 0.0306, 0.0360], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:48:21,377 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:21,911 INFO [finetune.py:976] (3/7) Epoch 17, batch 2050, loss[loss=0.1906, simple_loss=0.2568, pruned_loss=0.0622, over 4229.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2427, pruned_loss=0.05165, over 953805.05 frames. ], batch size: 65, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:48:24,847 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 09:48:35,566 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93714.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:44,211 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 09:48:53,240 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93740.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:55,037 INFO [finetune.py:976] (3/7) Epoch 17, batch 2100, loss[loss=0.1547, simple_loss=0.2305, pruned_loss=0.03945, over 4901.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2437, pruned_loss=0.05221, over 955884.01 frames. ], batch size: 36, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:49:01,813 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.279e+02 1.585e+02 1.847e+02 2.242e+02 6.268e+02, threshold=3.694e+02, percent-clipped=2.0 +2023-04-27 09:49:12,979 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93770.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:49:13,592 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93771.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:49:32,633 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2410, 1.2760, 4.1158, 3.8490, 3.5464, 3.8922, 3.9074, 3.5793], + device='cuda:3'), covar=tensor([0.7870, 0.6330, 0.1208, 0.2028, 0.1311, 0.2257, 0.1573, 0.1746], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0303, 0.0397, 0.0401, 0.0344, 0.0404, 0.0308, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:49:33,700 INFO [finetune.py:976] (3/7) Epoch 17, batch 2150, loss[loss=0.1905, simple_loss=0.2536, pruned_loss=0.06376, over 4896.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2482, pruned_loss=0.05408, over 956251.55 frames. ], batch size: 32, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:49:55,754 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 09:50:02,759 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5164, 3.3671, 0.8924, 1.8336, 2.0366, 2.3347, 2.0550, 1.1347], + device='cuda:3'), covar=tensor([0.1616, 0.1346, 0.2196, 0.1403, 0.1142, 0.1245, 0.1521, 0.1857], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0241, 0.0137, 0.0120, 0.0130, 0.0152, 0.0116, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 09:50:13,738 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:50:26,749 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93831.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:50:47,863 INFO [finetune.py:976] (3/7) Epoch 17, batch 2200, loss[loss=0.2144, simple_loss=0.2721, pruned_loss=0.07838, over 4194.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2502, pruned_loss=0.0549, over 953760.62 frames. ], batch size: 66, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:50:59,348 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.629e+02 1.916e+02 2.338e+02 4.475e+02, threshold=3.833e+02, percent-clipped=3.0 +2023-04-27 09:51:12,106 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2956, 1.3907, 1.3168, 1.6068, 1.5033, 1.8077, 1.2586, 3.0848], + device='cuda:3'), covar=tensor([0.0623, 0.0846, 0.0843, 0.1348, 0.0706, 0.0456, 0.0812, 0.0184], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 09:51:41,208 INFO [finetune.py:976] (3/7) Epoch 17, batch 2250, loss[loss=0.1886, simple_loss=0.2585, pruned_loss=0.0593, over 4876.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2508, pruned_loss=0.05502, over 952521.48 frames. ], batch size: 35, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:52:03,863 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1047, 0.8004, 0.9257, 0.8386, 1.2405, 0.9888, 0.8592, 0.9605], + device='cuda:3'), covar=tensor([0.1537, 0.1287, 0.1959, 0.1407, 0.0869, 0.1497, 0.1704, 0.2034], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0312, 0.0350, 0.0290, 0.0328, 0.0311, 0.0300, 0.0366], + device='cuda:3'), out_proj_covar=tensor([6.3586e-05, 6.5103e-05, 7.4575e-05, 5.8928e-05, 6.8305e-05, 6.5456e-05, + 6.3401e-05, 7.8041e-05], device='cuda:3') +2023-04-27 09:52:14,472 INFO [finetune.py:976] (3/7) Epoch 17, batch 2300, loss[loss=0.1611, simple_loss=0.2324, pruned_loss=0.04488, over 4850.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.251, pruned_loss=0.05434, over 953377.06 frames. ], batch size: 44, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:52:15,227 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4625, 1.9503, 2.2807, 2.9112, 2.3766, 1.8675, 1.9143, 2.2506], + device='cuda:3'), covar=tensor([0.3033, 0.3091, 0.1569, 0.2411, 0.2724, 0.2566, 0.3670, 0.2225], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0249, 0.0229, 0.0319, 0.0220, 0.0233, 0.0231, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 09:52:20,961 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.592e+02 1.958e+02 2.327e+02 4.753e+02, threshold=3.916e+02, percent-clipped=2.0 +2023-04-27 09:52:40,196 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:52:47,881 INFO [finetune.py:976] (3/7) Epoch 17, batch 2350, loss[loss=0.1673, simple_loss=0.2375, pruned_loss=0.04856, over 4825.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.249, pruned_loss=0.05395, over 952751.91 frames. ], batch size: 39, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:53:46,197 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8201, 1.4568, 1.8748, 2.3744, 1.9838, 1.7637, 1.8449, 1.8267], + device='cuda:3'), covar=tensor([0.4488, 0.6555, 0.6070, 0.5232, 0.5380, 0.7981, 0.7681, 0.8532], + device='cuda:3'), in_proj_covar=tensor([0.0421, 0.0406, 0.0497, 0.0506, 0.0450, 0.0474, 0.0479, 0.0484], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:53:48,617 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94042.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:53:49,095 INFO [finetune.py:976] (3/7) Epoch 17, batch 2400, loss[loss=0.175, simple_loss=0.243, pruned_loss=0.05346, over 4832.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2464, pruned_loss=0.05324, over 953956.04 frames. ], batch size: 47, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:53:56,111 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.528e+02 1.774e+02 2.113e+02 4.800e+02, threshold=3.549e+02, percent-clipped=1.0 +2023-04-27 09:54:18,339 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1386, 1.4265, 1.3006, 1.6338, 1.4883, 1.7015, 1.3148, 2.5265], + device='cuda:3'), covar=tensor([0.0585, 0.0766, 0.0745, 0.1150, 0.0633, 0.0454, 0.0734, 0.0214], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 09:54:23,069 INFO [finetune.py:976] (3/7) Epoch 17, batch 2450, loss[loss=0.156, simple_loss=0.2282, pruned_loss=0.04185, over 4764.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2438, pruned_loss=0.05213, over 952993.18 frames. ], batch size: 26, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:54:42,353 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 09:54:46,623 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94126.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:54:48,812 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 09:54:54,296 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 09:54:57,072 INFO [finetune.py:976] (3/7) Epoch 17, batch 2500, loss[loss=0.2199, simple_loss=0.2876, pruned_loss=0.07609, over 4126.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2453, pruned_loss=0.05314, over 952889.70 frames. ], batch size: 65, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:55:03,648 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.665e+02 1.997e+02 2.427e+02 4.291e+02, threshold=3.995e+02, percent-clipped=3.0 +2023-04-27 09:55:21,897 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3079, 1.3373, 3.7360, 3.4605, 3.3272, 3.5628, 3.4533, 3.2545], + device='cuda:3'), covar=tensor([0.7104, 0.5810, 0.1189, 0.1900, 0.1253, 0.1789, 0.2903, 0.1594], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0303, 0.0398, 0.0401, 0.0344, 0.0403, 0.0308, 0.0362], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 09:55:31,032 INFO [finetune.py:976] (3/7) Epoch 17, batch 2550, loss[loss=0.207, simple_loss=0.2634, pruned_loss=0.07534, over 4227.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2484, pruned_loss=0.05402, over 949486.01 frames. ], batch size: 18, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:55:52,768 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94223.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:56:09,849 INFO [finetune.py:976] (3/7) Epoch 17, batch 2600, loss[loss=0.2485, simple_loss=0.3154, pruned_loss=0.09083, over 4781.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2489, pruned_loss=0.05398, over 949377.71 frames. ], batch size: 51, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:56:21,228 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.596e+02 1.895e+02 2.398e+02 4.293e+02, threshold=3.790e+02, percent-clipped=2.0 +2023-04-27 09:57:00,077 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94284.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 09:57:05,478 INFO [finetune.py:976] (3/7) Epoch 17, batch 2650, loss[loss=0.1459, simple_loss=0.2181, pruned_loss=0.03679, over 4791.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2494, pruned_loss=0.05402, over 950458.21 frames. ], batch size: 25, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:57:34,893 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:57:38,467 INFO [finetune.py:976] (3/7) Epoch 17, batch 2700, loss[loss=0.153, simple_loss=0.226, pruned_loss=0.03998, over 4781.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2481, pruned_loss=0.05361, over 950769.65 frames. ], batch size: 29, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:57:41,627 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94348.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:57:43,963 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.591e+02 1.889e+02 2.175e+02 3.860e+02, threshold=3.779e+02, percent-clipped=1.0 +2023-04-27 09:57:44,664 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5575, 1.9078, 1.8109, 2.2689, 2.0576, 2.2435, 1.7507, 4.5898], + device='cuda:3'), covar=tensor([0.0530, 0.0745, 0.0764, 0.1124, 0.0610, 0.0474, 0.0697, 0.0094], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 09:58:17,893 INFO [finetune.py:976] (3/7) Epoch 17, batch 2750, loss[loss=0.165, simple_loss=0.2292, pruned_loss=0.05039, over 4833.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.245, pruned_loss=0.05222, over 949583.70 frames. ], batch size: 33, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:58:30,399 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 09:58:38,899 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94409.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 09:58:52,832 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-27 09:59:01,612 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94426.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:59:24,238 INFO [finetune.py:976] (3/7) Epoch 17, batch 2800, loss[loss=0.1623, simple_loss=0.2307, pruned_loss=0.04701, over 4907.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2428, pruned_loss=0.05135, over 950765.35 frames. ], batch size: 35, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:59:35,021 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.496e+02 1.786e+02 2.106e+02 4.249e+02, threshold=3.571e+02, percent-clipped=1.0 +2023-04-27 09:59:54,409 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94474.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:00:06,867 INFO [finetune.py:976] (3/7) Epoch 17, batch 2850, loss[loss=0.1747, simple_loss=0.2498, pruned_loss=0.04977, over 4223.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2416, pruned_loss=0.05105, over 947893.57 frames. ], batch size: 65, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:00:14,375 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4604, 3.5196, 1.0818, 1.8061, 1.9896, 2.5603, 1.9690, 0.9959], + device='cuda:3'), covar=tensor([0.1382, 0.0940, 0.1878, 0.1245, 0.1046, 0.0941, 0.1487, 0.2035], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0245, 0.0138, 0.0121, 0.0133, 0.0154, 0.0118, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:00:30,646 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-04-27 10:00:33,675 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1743, 1.6483, 2.1373, 2.6821, 2.1535, 1.6796, 1.5554, 1.9170], + device='cuda:3'), covar=tensor([0.3198, 0.3257, 0.1606, 0.2373, 0.2622, 0.2720, 0.4066, 0.2202], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0249, 0.0229, 0.0318, 0.0220, 0.0233, 0.0232, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 10:00:40,995 INFO [finetune.py:976] (3/7) Epoch 17, batch 2900, loss[loss=0.168, simple_loss=0.2468, pruned_loss=0.04456, over 4841.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2453, pruned_loss=0.05255, over 949825.37 frames. ], batch size: 49, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:00:46,390 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.635e+02 2.008e+02 2.460e+02 5.439e+02, threshold=4.016e+02, percent-clipped=3.0 +2023-04-27 10:01:03,847 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94579.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 10:01:09,675 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8925, 2.4683, 0.9709, 1.2770, 1.8148, 1.1598, 3.2224, 1.7413], + device='cuda:3'), covar=tensor([0.0783, 0.0659, 0.0841, 0.1318, 0.0567, 0.1074, 0.0243, 0.0631], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 10:01:14,240 INFO [finetune.py:976] (3/7) Epoch 17, batch 2950, loss[loss=0.2121, simple_loss=0.2821, pruned_loss=0.07105, over 4828.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2479, pruned_loss=0.05343, over 951148.54 frames. ], batch size: 51, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:01:27,157 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7194, 1.5632, 2.0291, 2.1159, 1.5524, 1.4289, 1.7533, 1.0436], + device='cuda:3'), covar=tensor([0.0542, 0.0865, 0.0404, 0.0565, 0.0758, 0.1267, 0.0612, 0.0677], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:01:48,054 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 10:02:00,087 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94637.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:02:04,040 INFO [finetune.py:976] (3/7) Epoch 17, batch 3000, loss[loss=0.1619, simple_loss=0.2444, pruned_loss=0.03967, over 4841.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2499, pruned_loss=0.05414, over 952192.81 frames. ], batch size: 44, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:02:04,040 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 10:02:12,449 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3754, 1.3033, 1.6404, 1.6011, 1.2858, 1.2023, 1.3891, 0.8200], + device='cuda:3'), covar=tensor([0.0614, 0.0772, 0.0478, 0.0577, 0.0810, 0.1127, 0.0568, 0.0585], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0068, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:02:27,439 INFO [finetune.py:1010] (3/7) Epoch 17, validation: loss=0.1526, simple_loss=0.2233, pruned_loss=0.04089, over 2265189.00 frames. +2023-04-27 10:02:27,439 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6344MB +2023-04-27 10:02:39,392 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.688e+02 2.099e+02 2.458e+02 3.567e+02, threshold=4.198e+02, percent-clipped=0.0 +2023-04-27 10:03:10,437 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:03:15,761 INFO [finetune.py:976] (3/7) Epoch 17, batch 3050, loss[loss=0.1524, simple_loss=0.2345, pruned_loss=0.03512, over 4905.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2511, pruned_loss=0.05468, over 950221.25 frames. ], batch size: 37, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:03:24,051 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94704.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:03:47,902 INFO [finetune.py:976] (3/7) Epoch 17, batch 3100, loss[loss=0.2379, simple_loss=0.2908, pruned_loss=0.0925, over 4167.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2506, pruned_loss=0.05489, over 947282.66 frames. ], batch size: 65, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:03:55,871 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.556e+02 1.873e+02 2.227e+02 4.960e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-04-27 10:03:58,295 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0308, 1.3937, 5.1554, 4.8074, 4.5075, 4.9743, 4.5561, 4.5697], + device='cuda:3'), covar=tensor([0.6630, 0.6309, 0.1057, 0.1750, 0.1139, 0.1382, 0.1388, 0.1537], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0304, 0.0401, 0.0402, 0.0346, 0.0404, 0.0308, 0.0363], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:04:21,247 INFO [finetune.py:976] (3/7) Epoch 17, batch 3150, loss[loss=0.2112, simple_loss=0.2722, pruned_loss=0.0751, over 4858.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2477, pruned_loss=0.05412, over 948922.18 frames. ], batch size: 47, lr: 3.39e-03, grad_scale: 64.0 +2023-04-27 10:05:23,151 INFO [finetune.py:976] (3/7) Epoch 17, batch 3200, loss[loss=0.196, simple_loss=0.2534, pruned_loss=0.06929, over 4121.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2448, pruned_loss=0.05299, over 950221.94 frames. ], batch size: 65, lr: 3.39e-03, grad_scale: 64.0 +2023-04-27 10:05:34,598 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.148e+01 1.533e+02 1.770e+02 2.163e+02 3.313e+02, threshold=3.540e+02, percent-clipped=0.0 +2023-04-27 10:05:53,657 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94879.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:06:02,206 INFO [finetune.py:976] (3/7) Epoch 17, batch 3250, loss[loss=0.2219, simple_loss=0.2951, pruned_loss=0.07436, over 4825.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2445, pruned_loss=0.05269, over 953120.28 frames. ], batch size: 39, lr: 3.39e-03, grad_scale: 64.0 +2023-04-27 10:06:04,538 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-27 10:06:19,240 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5025, 3.1441, 0.8440, 1.7680, 1.8103, 2.2471, 1.7282, 0.9538], + device='cuda:3'), covar=tensor([0.1392, 0.0949, 0.2143, 0.1205, 0.1100, 0.1070, 0.1619, 0.2132], + device='cuda:3'), in_proj_covar=tensor([0.0119, 0.0247, 0.0140, 0.0122, 0.0134, 0.0155, 0.0120, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:06:26,489 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94927.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:06:36,232 INFO [finetune.py:976] (3/7) Epoch 17, batch 3300, loss[loss=0.19, simple_loss=0.2597, pruned_loss=0.06015, over 4799.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2473, pruned_loss=0.05357, over 954580.47 frames. ], batch size: 29, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:06:47,632 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.257e+02 1.598e+02 1.896e+02 2.226e+02 3.987e+02, threshold=3.793e+02, percent-clipped=1.0 +2023-04-27 10:07:30,062 INFO [finetune.py:976] (3/7) Epoch 17, batch 3350, loss[loss=0.206, simple_loss=0.2818, pruned_loss=0.06507, over 4816.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2508, pruned_loss=0.05467, over 956166.64 frames. ], batch size: 39, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:07:36,977 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95004.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:08:03,816 INFO [finetune.py:976] (3/7) Epoch 17, batch 3400, loss[loss=0.2169, simple_loss=0.2822, pruned_loss=0.07581, over 4104.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.252, pruned_loss=0.05509, over 956934.57 frames. ], batch size: 65, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:08:09,293 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.633e+02 1.895e+02 2.474e+02 5.536e+02, threshold=3.790e+02, percent-clipped=1.0 +2023-04-27 10:08:09,365 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95052.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:08:33,732 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4873, 1.9163, 2.4072, 2.9865, 2.3671, 1.9246, 1.7462, 2.2080], + device='cuda:3'), covar=tensor([0.3333, 0.3398, 0.1583, 0.2520, 0.2966, 0.2580, 0.4113, 0.2268], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0247, 0.0226, 0.0316, 0.0218, 0.0231, 0.0229, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 10:08:37,251 INFO [finetune.py:976] (3/7) Epoch 17, batch 3450, loss[loss=0.166, simple_loss=0.2357, pruned_loss=0.04816, over 4845.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2515, pruned_loss=0.05479, over 955895.82 frames. ], batch size: 49, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:08:44,644 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95105.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:09:04,855 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.65 vs. limit=2.0 +2023-04-27 10:09:11,016 INFO [finetune.py:976] (3/7) Epoch 17, batch 3500, loss[loss=0.1472, simple_loss=0.2259, pruned_loss=0.03428, over 4794.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.249, pruned_loss=0.05384, over 955688.25 frames. ], batch size: 29, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:09:15,405 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 10:09:16,406 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.623e+02 1.952e+02 2.265e+02 3.860e+02, threshold=3.904e+02, percent-clipped=1.0 +2023-04-27 10:09:19,038 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 10:09:26,083 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95166.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:09:44,913 INFO [finetune.py:976] (3/7) Epoch 17, batch 3550, loss[loss=0.1885, simple_loss=0.2462, pruned_loss=0.06536, over 4868.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2465, pruned_loss=0.05317, over 955582.25 frames. ], batch size: 34, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:09:46,889 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5550, 2.4774, 1.9479, 2.2092, 2.3511, 1.9801, 3.0824, 1.8486], + device='cuda:3'), covar=tensor([0.3567, 0.1901, 0.4451, 0.3246, 0.2025, 0.2567, 0.2103, 0.4486], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0346, 0.0427, 0.0354, 0.0383, 0.0379, 0.0372, 0.0421], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:10:42,815 INFO [finetune.py:976] (3/7) Epoch 17, batch 3600, loss[loss=0.1836, simple_loss=0.2735, pruned_loss=0.04684, over 4851.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2444, pruned_loss=0.05293, over 954953.92 frames. ], batch size: 49, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:10:54,068 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.577e+02 1.958e+02 2.437e+02 7.407e+02, threshold=3.916e+02, percent-clipped=3.0 +2023-04-27 10:10:59,896 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:11:10,483 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95268.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:11:32,404 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0004, 1.5183, 1.9086, 2.1316, 1.8649, 1.5065, 1.1440, 1.6010], + device='cuda:3'), covar=tensor([0.3294, 0.3076, 0.1565, 0.2292, 0.2518, 0.2636, 0.4229, 0.1985], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0246, 0.0227, 0.0315, 0.0218, 0.0230, 0.0229, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 10:11:34,140 INFO [finetune.py:976] (3/7) Epoch 17, batch 3650, loss[loss=0.1985, simple_loss=0.277, pruned_loss=0.06006, over 4820.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2459, pruned_loss=0.05335, over 954170.12 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:11:51,463 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:11:57,806 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:12:05,340 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95339.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:12:07,680 INFO [finetune.py:976] (3/7) Epoch 17, batch 3700, loss[loss=0.1859, simple_loss=0.2649, pruned_loss=0.05349, over 4824.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2509, pruned_loss=0.05453, over 954140.39 frames. ], batch size: 49, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:12:13,171 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.977e+01 1.742e+02 1.979e+02 2.292e+02 4.077e+02, threshold=3.957e+02, percent-clipped=1.0 +2023-04-27 10:12:41,445 INFO [finetune.py:976] (3/7) Epoch 17, batch 3750, loss[loss=0.165, simple_loss=0.2381, pruned_loss=0.04598, over 4821.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2522, pruned_loss=0.05555, over 953133.19 frames. ], batch size: 33, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:12:46,015 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95400.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:12:52,193 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 10:13:14,401 INFO [finetune.py:976] (3/7) Epoch 17, batch 3800, loss[loss=0.1538, simple_loss=0.2238, pruned_loss=0.04189, over 4746.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2516, pruned_loss=0.05467, over 954948.96 frames. ], batch size: 27, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:13:18,151 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 10:13:20,914 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.628e+02 1.968e+02 2.270e+02 5.105e+02, threshold=3.935e+02, percent-clipped=1.0 +2023-04-27 10:13:25,884 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95461.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:13:48,140 INFO [finetune.py:976] (3/7) Epoch 17, batch 3850, loss[loss=0.1833, simple_loss=0.2469, pruned_loss=0.05982, over 4809.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2504, pruned_loss=0.05398, over 955886.86 frames. ], batch size: 30, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:13:49,762 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 10:13:55,158 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 10:14:20,429 INFO [finetune.py:976] (3/7) Epoch 17, batch 3900, loss[loss=0.1554, simple_loss=0.2249, pruned_loss=0.04294, over 4807.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.248, pruned_loss=0.0537, over 956658.47 frames. ], batch size: 25, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:14:27,544 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.967e+01 1.562e+02 1.786e+02 2.211e+02 6.075e+02, threshold=3.573e+02, percent-clipped=1.0 +2023-04-27 10:14:52,352 INFO [finetune.py:976] (3/7) Epoch 17, batch 3950, loss[loss=0.1577, simple_loss=0.2299, pruned_loss=0.04275, over 4895.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2445, pruned_loss=0.05217, over 956643.58 frames. ], batch size: 32, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:15:08,540 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:15:09,191 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:15:13,457 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95624.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:15:16,788 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-04-27 10:15:31,265 INFO [finetune.py:976] (3/7) Epoch 17, batch 4000, loss[loss=0.1107, simple_loss=0.172, pruned_loss=0.02473, over 3837.00 frames. ], tot_loss[loss=0.172, simple_loss=0.242, pruned_loss=0.05104, over 955169.64 frames. ], batch size: 16, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:15:43,873 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.207e+01 1.538e+02 1.849e+02 2.258e+02 4.125e+02, threshold=3.697e+02, percent-clipped=1.0 +2023-04-27 10:16:16,180 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5330, 1.1562, 1.2533, 1.2794, 1.6852, 1.3960, 1.1154, 1.1656], + device='cuda:3'), covar=tensor([0.1816, 0.1355, 0.1988, 0.1328, 0.0802, 0.1326, 0.1893, 0.2163], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0312, 0.0349, 0.0287, 0.0325, 0.0308, 0.0299, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.3366e-05, 6.5154e-05, 7.4384e-05, 5.8322e-05, 6.7570e-05, 6.4944e-05, + 6.3062e-05, 7.7648e-05], device='cuda:3') +2023-04-27 10:16:16,183 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95678.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:16:36,027 INFO [finetune.py:976] (3/7) Epoch 17, batch 4050, loss[loss=0.1881, simple_loss=0.259, pruned_loss=0.05859, over 4932.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2443, pruned_loss=0.05216, over 954711.76 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:16:37,794 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95695.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:17:36,636 INFO [finetune.py:976] (3/7) Epoch 17, batch 4100, loss[loss=0.2062, simple_loss=0.2786, pruned_loss=0.06695, over 4821.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2468, pruned_loss=0.05294, over 952979.93 frames. ], batch size: 40, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:17:43,691 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.341e+02 1.651e+02 1.954e+02 2.364e+02 4.876e+02, threshold=3.909e+02, percent-clipped=1.0 +2023-04-27 10:17:50,090 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95761.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:17:57,951 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1292, 0.6365, 0.8880, 0.7975, 1.2513, 0.9521, 0.8391, 0.9234], + device='cuda:3'), covar=tensor([0.1577, 0.1778, 0.2017, 0.1515, 0.1028, 0.1444, 0.1624, 0.2455], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0312, 0.0349, 0.0288, 0.0326, 0.0309, 0.0299, 0.0365], + device='cuda:3'), out_proj_covar=tensor([6.3549e-05, 6.5189e-05, 7.4364e-05, 5.8463e-05, 6.7820e-05, 6.5149e-05, + 6.2964e-05, 7.7793e-05], device='cuda:3') +2023-04-27 10:18:06,460 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1482, 1.5984, 2.0780, 2.3799, 2.1290, 1.6031, 1.2745, 1.8163], + device='cuda:3'), covar=tensor([0.3745, 0.3385, 0.1767, 0.2456, 0.2732, 0.2780, 0.4484, 0.2141], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0246, 0.0227, 0.0316, 0.0218, 0.0231, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 10:18:09,896 INFO [finetune.py:976] (3/7) Epoch 17, batch 4150, loss[loss=0.1749, simple_loss=0.2618, pruned_loss=0.04406, over 4812.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2485, pruned_loss=0.05381, over 951055.23 frames. ], batch size: 39, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:18:21,840 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95809.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:18:24,179 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95812.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:18:24,989 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.79 vs. limit=5.0 +2023-04-27 10:18:25,715 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 10:18:43,506 INFO [finetune.py:976] (3/7) Epoch 17, batch 4200, loss[loss=0.1558, simple_loss=0.2349, pruned_loss=0.03836, over 4842.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2504, pruned_loss=0.05431, over 952685.79 frames. ], batch size: 49, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:18:46,045 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95847.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:18:49,670 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.908e+01 1.602e+02 1.963e+02 2.412e+02 3.992e+02, threshold=3.927e+02, percent-clipped=2.0 +2023-04-27 10:18:58,345 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2863, 2.0838, 2.5586, 2.8243, 2.7346, 2.3258, 1.9773, 2.4980], + device='cuda:3'), covar=tensor([0.0850, 0.0969, 0.0544, 0.0551, 0.0554, 0.0750, 0.0775, 0.0557], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0200, 0.0181, 0.0171, 0.0175, 0.0179, 0.0151, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:19:04,205 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95873.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:19:05,980 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5875, 3.2969, 0.7689, 1.9565, 2.0212, 2.2675, 1.9401, 0.9962], + device='cuda:3'), covar=tensor([0.1325, 0.0813, 0.2087, 0.1081, 0.0922, 0.1051, 0.1444, 0.1838], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0244, 0.0138, 0.0120, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:19:16,330 INFO [finetune.py:976] (3/7) Epoch 17, batch 4250, loss[loss=0.1791, simple_loss=0.2426, pruned_loss=0.05783, over 4811.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2484, pruned_loss=0.05349, over 953197.49 frames. ], batch size: 30, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:19:26,045 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95908.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:19:31,926 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:19:36,857 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95924.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:19:48,683 INFO [finetune.py:976] (3/7) Epoch 17, batch 4300, loss[loss=0.1538, simple_loss=0.22, pruned_loss=0.04378, over 4901.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2456, pruned_loss=0.05267, over 952966.64 frames. ], batch size: 35, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:19:53,135 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8853, 2.4143, 1.7853, 1.7018, 1.3410, 1.4064, 1.8879, 1.2856], + device='cuda:3'), covar=tensor([0.1721, 0.1342, 0.1546, 0.1834, 0.2474, 0.2002, 0.1034, 0.2097], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0212, 0.0167, 0.0203, 0.0200, 0.0183, 0.0155, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 10:19:54,839 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.623e+02 1.872e+02 2.208e+02 5.069e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-04-27 10:20:03,627 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95964.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:20:09,301 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:20:09,894 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95973.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:20:14,628 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:20:22,489 INFO [finetune.py:976] (3/7) Epoch 17, batch 4350, loss[loss=0.1931, simple_loss=0.2575, pruned_loss=0.0643, over 4813.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.242, pruned_loss=0.05158, over 952984.98 frames. ], batch size: 39, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:20:23,791 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95995.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:20:30,537 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4822, 3.4450, 2.5403, 4.1030, 3.5124, 3.5234, 1.5073, 3.4929], + device='cuda:3'), covar=tensor([0.1834, 0.1292, 0.3454, 0.1881, 0.3589, 0.1880, 0.5797, 0.2539], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0213, 0.0249, 0.0304, 0.0297, 0.0249, 0.0272, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:21:07,889 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96041.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:21:08,983 INFO [finetune.py:976] (3/7) Epoch 17, batch 4400, loss[loss=0.153, simple_loss=0.2254, pruned_loss=0.04028, over 4934.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2418, pruned_loss=0.05147, over 954596.87 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:21:09,042 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96043.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:21:15,038 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.713e+01 1.567e+02 1.808e+02 2.181e+02 3.991e+02, threshold=3.617e+02, percent-clipped=1.0 +2023-04-27 10:21:34,792 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:21:48,840 INFO [finetune.py:976] (3/7) Epoch 17, batch 4450, loss[loss=0.1698, simple_loss=0.2498, pruned_loss=0.04488, over 4317.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2468, pruned_loss=0.05328, over 952674.12 frames. ], batch size: 19, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:22:11,119 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6083, 3.5785, 2.6393, 4.2730, 3.6897, 3.6238, 1.5783, 3.7091], + device='cuda:3'), covar=tensor([0.1920, 0.1354, 0.3598, 0.1726, 0.4094, 0.1945, 0.6017, 0.2542], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0215, 0.0251, 0.0307, 0.0300, 0.0252, 0.0275, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:22:11,136 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5798, 1.3279, 4.2445, 3.9632, 3.7263, 4.0283, 3.8894, 3.7511], + device='cuda:3'), covar=tensor([0.7070, 0.5937, 0.0962, 0.1567, 0.1180, 0.2032, 0.1751, 0.1368], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0305, 0.0400, 0.0404, 0.0347, 0.0403, 0.0309, 0.0363], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:22:53,812 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96141.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:22:54,919 INFO [finetune.py:976] (3/7) Epoch 17, batch 4500, loss[loss=0.1674, simple_loss=0.2395, pruned_loss=0.04764, over 4889.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2493, pruned_loss=0.05382, over 955711.84 frames. ], batch size: 35, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:23:02,750 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.688e+02 1.990e+02 2.416e+02 4.917e+02, threshold=3.981e+02, percent-clipped=4.0 +2023-04-27 10:23:16,184 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:23:17,979 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96168.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:23:57,685 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0064, 1.5297, 1.5432, 1.7715, 2.1513, 1.8060, 1.4982, 1.4805], + device='cuda:3'), covar=tensor([0.1283, 0.1504, 0.1829, 0.1259, 0.0665, 0.1544, 0.1869, 0.1973], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0314, 0.0352, 0.0288, 0.0328, 0.0310, 0.0300, 0.0366], + device='cuda:3'), out_proj_covar=tensor([6.3983e-05, 6.5380e-05, 7.5071e-05, 5.8557e-05, 6.8114e-05, 6.5251e-05, + 6.3273e-05, 7.8075e-05], device='cuda:3') +2023-04-27 10:23:58,145 INFO [finetune.py:976] (3/7) Epoch 17, batch 4550, loss[loss=0.1549, simple_loss=0.2258, pruned_loss=0.04198, over 4766.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2508, pruned_loss=0.05412, over 957171.50 frames. ], batch size: 26, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:24:00,066 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9154, 2.5868, 2.1685, 2.1598, 1.3735, 1.4595, 2.2572, 1.4085], + device='cuda:3'), covar=tensor([0.1802, 0.1651, 0.1425, 0.1754, 0.2465, 0.2104, 0.0986, 0.2185], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0167, 0.0203, 0.0200, 0.0184, 0.0155, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 10:24:10,602 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96203.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:24:42,303 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96226.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:24:54,978 INFO [finetune.py:976] (3/7) Epoch 17, batch 4600, loss[loss=0.1204, simple_loss=0.1922, pruned_loss=0.0243, over 4787.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2503, pruned_loss=0.0539, over 958528.55 frames. ], batch size: 25, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:25:01,049 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.538e+02 1.819e+02 2.307e+02 5.540e+02, threshold=3.639e+02, percent-clipped=2.0 +2023-04-27 10:25:13,340 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:25:27,833 INFO [finetune.py:976] (3/7) Epoch 17, batch 4650, loss[loss=0.191, simple_loss=0.2566, pruned_loss=0.0627, over 4865.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2481, pruned_loss=0.05344, over 960127.45 frames. ], batch size: 34, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:25:30,978 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 10:25:37,140 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 10:25:45,392 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:25:47,234 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96324.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:25:49,105 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7178, 2.0827, 1.8159, 1.4155, 1.3115, 1.3465, 1.8334, 1.2887], + device='cuda:3'), covar=tensor([0.1620, 0.1289, 0.1284, 0.1732, 0.2240, 0.1903, 0.0910, 0.1906], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0212, 0.0167, 0.0203, 0.0200, 0.0183, 0.0155, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 10:25:56,371 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:26:01,179 INFO [finetune.py:976] (3/7) Epoch 17, batch 4700, loss[loss=0.1422, simple_loss=0.217, pruned_loss=0.03366, over 4918.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2457, pruned_loss=0.05289, over 958743.96 frames. ], batch size: 37, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:26:05,543 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.78 vs. limit=5.0 +2023-04-27 10:26:07,579 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.173e+02 1.665e+02 1.923e+02 2.344e+02 4.487e+02, threshold=3.847e+02, percent-clipped=4.0 +2023-04-27 10:26:43,975 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:26:48,549 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:26:55,844 INFO [finetune.py:976] (3/7) Epoch 17, batch 4750, loss[loss=0.1616, simple_loss=0.2335, pruned_loss=0.04489, over 4748.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2445, pruned_loss=0.05283, over 958055.22 frames. ], batch size: 27, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:27:18,041 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2220, 3.1892, 2.4350, 3.7591, 3.2118, 3.2336, 1.5137, 3.2477], + device='cuda:3'), covar=tensor([0.1873, 0.1445, 0.3802, 0.2275, 0.3287, 0.1886, 0.5292, 0.2515], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0215, 0.0251, 0.0308, 0.0300, 0.0250, 0.0274, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:27:45,322 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96436.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:27:50,990 INFO [finetune.py:976] (3/7) Epoch 17, batch 4800, loss[loss=0.1945, simple_loss=0.2646, pruned_loss=0.06219, over 4804.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2464, pruned_loss=0.05309, over 958185.68 frames. ], batch size: 41, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:27:53,563 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-27 10:27:56,400 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96451.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:27:57,982 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.553e+02 2.009e+02 2.369e+02 4.346e+02, threshold=4.018e+02, percent-clipped=1.0 +2023-04-27 10:28:03,810 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 10:28:07,160 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96468.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:28:34,487 INFO [finetune.py:976] (3/7) Epoch 17, batch 4850, loss[loss=0.217, simple_loss=0.2894, pruned_loss=0.07234, over 4809.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2491, pruned_loss=0.05336, over 956084.10 frames. ], batch size: 45, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:28:41,601 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96503.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:28:49,380 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5438, 1.3911, 1.8458, 1.8065, 1.3946, 1.2369, 1.5604, 0.8982], + device='cuda:3'), covar=tensor([0.0563, 0.0773, 0.0392, 0.0721, 0.0896, 0.1265, 0.0591, 0.0700], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0074, 0.0095, 0.0073, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:28:49,916 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96516.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:28:52,925 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96521.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:29:07,104 INFO [finetune.py:976] (3/7) Epoch 17, batch 4900, loss[loss=0.3166, simple_loss=0.3507, pruned_loss=0.1412, over 4160.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2501, pruned_loss=0.05402, over 955502.01 frames. ], batch size: 65, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:29:11,192 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.84 vs. limit=5.0 +2023-04-27 10:29:13,476 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96551.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:29:15,083 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.788e+01 1.547e+02 1.882e+02 2.260e+02 4.956e+02, threshold=3.763e+02, percent-clipped=3.0 +2023-04-27 10:29:21,096 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.0307, 3.9269, 3.0048, 4.6159, 4.0966, 3.9527, 1.8685, 3.9696], + device='cuda:3'), covar=tensor([0.1606, 0.1411, 0.2973, 0.1270, 0.2238, 0.1697, 0.5469, 0.2268], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0214, 0.0249, 0.0305, 0.0297, 0.0248, 0.0272, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:29:39,872 INFO [finetune.py:976] (3/7) Epoch 17, batch 4950, loss[loss=0.1479, simple_loss=0.217, pruned_loss=0.0394, over 4812.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2514, pruned_loss=0.05436, over 955955.76 frames. ], batch size: 25, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:30:09,373 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96636.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:30:13,520 INFO [finetune.py:976] (3/7) Epoch 17, batch 5000, loss[loss=0.1713, simple_loss=0.2387, pruned_loss=0.05201, over 4306.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2492, pruned_loss=0.05322, over 955254.34 frames. ], batch size: 66, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:30:21,475 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.257e+01 1.582e+02 1.781e+02 2.065e+02 3.108e+02, threshold=3.561e+02, percent-clipped=0.0 +2023-04-27 10:30:31,579 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2546, 1.2972, 1.4203, 1.5831, 1.6763, 1.2647, 0.8366, 1.4272], + device='cuda:3'), covar=tensor([0.0891, 0.1296, 0.0927, 0.0656, 0.0679, 0.0875, 0.0952, 0.0642], + device='cuda:3'), in_proj_covar=tensor([0.0190, 0.0202, 0.0182, 0.0172, 0.0177, 0.0181, 0.0153, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:30:38,900 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:30:41,300 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96684.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:30:46,664 INFO [finetune.py:976] (3/7) Epoch 17, batch 5050, loss[loss=0.194, simple_loss=0.2554, pruned_loss=0.06626, over 4900.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2463, pruned_loss=0.05253, over 956362.36 frames. ], batch size: 43, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:31:00,174 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3438, 1.8299, 2.3222, 2.6523, 2.2663, 1.8195, 1.4404, 1.9950], + device='cuda:3'), covar=tensor([0.3191, 0.3155, 0.1607, 0.2185, 0.2556, 0.2576, 0.4229, 0.2044], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0247, 0.0227, 0.0316, 0.0218, 0.0231, 0.0229, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 10:31:12,143 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.87 vs. limit=5.0 +2023-04-27 10:31:15,662 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:31:19,857 INFO [finetune.py:976] (3/7) Epoch 17, batch 5100, loss[loss=0.1853, simple_loss=0.2523, pruned_loss=0.05914, over 4815.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2428, pruned_loss=0.05161, over 956753.06 frames. ], batch size: 40, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:31:21,739 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96746.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:31:24,903 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 10:31:25,903 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.614e+02 1.893e+02 2.231e+02 4.752e+02, threshold=3.786e+02, percent-clipped=4.0 +2023-04-27 10:31:54,176 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96776.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:32:04,478 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:32:15,344 INFO [finetune.py:976] (3/7) Epoch 17, batch 5150, loss[loss=0.1989, simple_loss=0.2676, pruned_loss=0.06508, over 4814.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.244, pruned_loss=0.05322, over 953005.87 frames. ], batch size: 41, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:32:46,405 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:33:03,948 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96837.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:33:12,846 INFO [finetune.py:976] (3/7) Epoch 17, batch 5200, loss[loss=0.2206, simple_loss=0.2997, pruned_loss=0.07078, over 4862.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2485, pruned_loss=0.05398, over 955514.42 frames. ], batch size: 34, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:33:24,748 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.802e+01 1.740e+02 2.121e+02 2.596e+02 7.000e+02, threshold=4.242e+02, percent-clipped=4.0 +2023-04-27 10:33:47,097 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96869.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:34:02,258 INFO [finetune.py:976] (3/7) Epoch 17, batch 5250, loss[loss=0.1938, simple_loss=0.2612, pruned_loss=0.06319, over 4755.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2505, pruned_loss=0.05459, over 954467.70 frames. ], batch size: 27, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:34:12,066 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6842, 1.1146, 1.7440, 2.1067, 1.7533, 1.6894, 1.7322, 1.7464], + device='cuda:3'), covar=tensor([0.4812, 0.6798, 0.6719, 0.6089, 0.6077, 0.7818, 0.7687, 0.8206], + device='cuda:3'), in_proj_covar=tensor([0.0422, 0.0409, 0.0499, 0.0506, 0.0451, 0.0477, 0.0483, 0.0485], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:34:35,558 INFO [finetune.py:976] (3/7) Epoch 17, batch 5300, loss[loss=0.1816, simple_loss=0.2657, pruned_loss=0.04876, over 4790.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2516, pruned_loss=0.0552, over 950879.79 frames. ], batch size: 51, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:34:36,347 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.55 vs. limit=5.0 +2023-04-27 10:34:41,676 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.524e+02 1.845e+02 2.298e+02 4.151e+02, threshold=3.690e+02, percent-clipped=0.0 +2023-04-27 10:34:45,483 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9825, 1.6825, 1.9611, 2.2318, 2.2991, 1.8107, 1.4505, 2.0530], + device='cuda:3'), covar=tensor([0.0793, 0.1192, 0.0708, 0.0629, 0.0570, 0.0821, 0.0837, 0.0554], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0200, 0.0180, 0.0170, 0.0175, 0.0179, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:35:00,590 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:35:08,783 INFO [finetune.py:976] (3/7) Epoch 17, batch 5350, loss[loss=0.1841, simple_loss=0.2563, pruned_loss=0.05592, over 4727.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2519, pruned_loss=0.05473, over 951439.59 frames. ], batch size: 54, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:35:13,333 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97000.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:35:33,057 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97028.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:35:42,645 INFO [finetune.py:976] (3/7) Epoch 17, batch 5400, loss[loss=0.1503, simple_loss=0.2219, pruned_loss=0.03931, over 4752.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2492, pruned_loss=0.05399, over 951035.92 frames. ], batch size: 28, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:35:44,573 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97046.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:35:48,748 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.600e+02 1.805e+02 2.105e+02 4.244e+02, threshold=3.609e+02, percent-clipped=1.0 +2023-04-27 10:35:51,366 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4271, 1.7783, 1.7939, 1.9111, 1.8181, 1.8956, 1.8685, 1.8341], + device='cuda:3'), covar=tensor([0.3830, 0.4931, 0.4377, 0.4305, 0.5127, 0.6954, 0.5270, 0.4819], + device='cuda:3'), in_proj_covar=tensor([0.0331, 0.0371, 0.0317, 0.0331, 0.0342, 0.0392, 0.0354, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:35:53,812 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97061.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:36:06,706 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2253, 1.4079, 1.2585, 1.7168, 1.5806, 1.5888, 1.3429, 2.4761], + device='cuda:3'), covar=tensor([0.0567, 0.0828, 0.0866, 0.1186, 0.0642, 0.0455, 0.0769, 0.0252], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 10:36:15,421 INFO [finetune.py:976] (3/7) Epoch 17, batch 5450, loss[loss=0.1262, simple_loss=0.2028, pruned_loss=0.02476, over 4761.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2455, pruned_loss=0.0526, over 951526.17 frames. ], batch size: 54, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:36:16,112 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97094.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:36:28,417 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-27 10:36:31,401 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5745, 2.5960, 2.3509, 2.2845, 2.7035, 2.3395, 3.2440, 2.1264], + device='cuda:3'), covar=tensor([0.2832, 0.1491, 0.3182, 0.2288, 0.1191, 0.2053, 0.1115, 0.3238], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0341, 0.0419, 0.0350, 0.0376, 0.0375, 0.0365, 0.0413], + device='cuda:3'), out_proj_covar=tensor([9.9045e-05, 1.0264e-04, 1.2772e-04, 1.0584e-04, 1.1223e-04, 1.1233e-04, + 1.0752e-04, 1.2526e-04], device='cuda:3') +2023-04-27 10:36:40,613 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97132.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:36:47,755 INFO [finetune.py:976] (3/7) Epoch 17, batch 5500, loss[loss=0.1657, simple_loss=0.2379, pruned_loss=0.04673, over 4827.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2424, pruned_loss=0.05184, over 953176.38 frames. ], batch size: 39, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:36:54,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.627e+02 1.973e+02 2.284e+02 4.082e+02, threshold=3.945e+02, percent-clipped=1.0 +2023-04-27 10:37:37,878 INFO [finetune.py:976] (3/7) Epoch 17, batch 5550, loss[loss=0.1935, simple_loss=0.2528, pruned_loss=0.06714, over 4906.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2447, pruned_loss=0.05261, over 952943.63 frames. ], batch size: 32, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:37:59,576 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4448, 2.3053, 1.8031, 2.0954, 2.3314, 1.9406, 3.0472, 1.6557], + device='cuda:3'), covar=tensor([0.3863, 0.2298, 0.4589, 0.3537, 0.1961, 0.2831, 0.2189, 0.4678], + device='cuda:3'), in_proj_covar=tensor([0.0335, 0.0343, 0.0422, 0.0352, 0.0378, 0.0377, 0.0368, 0.0415], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:38:15,510 INFO [finetune.py:976] (3/7) Epoch 17, batch 5600, loss[loss=0.2028, simple_loss=0.2865, pruned_loss=0.05959, over 4869.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2489, pruned_loss=0.05373, over 953174.71 frames. ], batch size: 34, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:38:26,566 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.758e+02 2.119e+02 2.585e+02 7.806e+02, threshold=4.239e+02, percent-clipped=5.0 +2023-04-27 10:38:28,378 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-27 10:38:37,077 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:39:14,493 INFO [finetune.py:976] (3/7) Epoch 17, batch 5650, loss[loss=0.224, simple_loss=0.2984, pruned_loss=0.07479, over 4770.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2513, pruned_loss=0.05426, over 951047.49 frames. ], batch size: 54, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:39:22,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8499, 1.4426, 1.4614, 1.6755, 2.1480, 1.6616, 1.3941, 1.3696], + device='cuda:3'), covar=tensor([0.1506, 0.1531, 0.2009, 0.1345, 0.0739, 0.1899, 0.2255, 0.2280], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0314, 0.0354, 0.0290, 0.0330, 0.0311, 0.0302, 0.0370], + device='cuda:3'), out_proj_covar=tensor([6.4019e-05, 6.5461e-05, 7.5287e-05, 5.8838e-05, 6.8608e-05, 6.5489e-05, + 6.3582e-05, 7.8932e-05], device='cuda:3') +2023-04-27 10:39:45,144 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.1419, 2.8035, 3.1098, 3.5165, 3.3041, 3.0098, 2.5084, 3.2044], + device='cuda:3'), covar=tensor([0.0727, 0.0832, 0.0488, 0.0550, 0.0562, 0.0789, 0.0714, 0.0490], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0198, 0.0179, 0.0169, 0.0174, 0.0178, 0.0149, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:39:47,529 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:40:16,392 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97342.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:40:16,917 INFO [finetune.py:976] (3/7) Epoch 17, batch 5700, loss[loss=0.1976, simple_loss=0.2522, pruned_loss=0.07152, over 3989.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2472, pruned_loss=0.05346, over 933740.14 frames. ], batch size: 17, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:40:19,357 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97347.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:40:22,861 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.291e+01 1.370e+02 1.630e+02 1.988e+02 3.241e+02, threshold=3.261e+02, percent-clipped=0.0 +2023-04-27 10:40:24,680 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:40:46,576 INFO [finetune.py:976] (3/7) Epoch 18, batch 0, loss[loss=0.1849, simple_loss=0.2525, pruned_loss=0.05864, over 4930.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2525, pruned_loss=0.05864, over 4930.00 frames. ], batch size: 42, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:40:46,576 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 10:41:03,166 INFO [finetune.py:1010] (3/7) Epoch 18, validation: loss=0.1537, simple_loss=0.225, pruned_loss=0.04121, over 2265189.00 frames. +2023-04-27 10:41:03,167 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 10:41:05,477 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:41:28,942 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97403.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:41:32,028 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97408.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:41:36,757 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9161, 4.0969, 0.8818, 2.3118, 2.3181, 2.5925, 2.4371, 0.9766], + device='cuda:3'), covar=tensor([0.1381, 0.0944, 0.2205, 0.1133, 0.1027, 0.1125, 0.1350, 0.2211], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0243, 0.0137, 0.0121, 0.0133, 0.0153, 0.0118, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:41:40,834 INFO [finetune.py:976] (3/7) Epoch 18, batch 50, loss[loss=0.1845, simple_loss=0.2614, pruned_loss=0.05378, over 4832.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2531, pruned_loss=0.05633, over 216825.98 frames. ], batch size: 49, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:41:49,629 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:41:51,454 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97435.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:42:02,222 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.680e+01 1.493e+02 1.785e+02 2.115e+02 3.636e+02, threshold=3.569e+02, percent-clipped=3.0 +2023-04-27 10:42:14,117 INFO [finetune.py:976] (3/7) Epoch 18, batch 100, loss[loss=0.1715, simple_loss=0.2395, pruned_loss=0.05174, over 4888.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2489, pruned_loss=0.05618, over 381668.27 frames. ], batch size: 32, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:42:22,015 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97480.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:42:43,241 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9311, 1.6572, 1.9444, 2.3517, 2.2824, 1.8122, 1.5243, 2.0699], + device='cuda:3'), covar=tensor([0.0820, 0.1121, 0.0691, 0.0507, 0.0566, 0.0879, 0.0763, 0.0531], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0197, 0.0178, 0.0167, 0.0173, 0.0176, 0.0148, 0.0175], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:42:47,722 INFO [finetune.py:976] (3/7) Epoch 18, batch 150, loss[loss=0.1537, simple_loss=0.2305, pruned_loss=0.03839, over 4873.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2432, pruned_loss=0.05336, over 510424.70 frames. ], batch size: 34, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:43:08,514 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.039e+02 1.668e+02 1.928e+02 2.305e+02 4.422e+02, threshold=3.856e+02, percent-clipped=2.0 +2023-04-27 10:43:09,238 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1656, 2.6080, 1.0243, 1.6918, 2.2379, 1.3579, 3.6147, 2.0097], + device='cuda:3'), covar=tensor([0.0632, 0.0661, 0.0823, 0.1198, 0.0441, 0.1004, 0.0304, 0.0587], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0074, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 10:43:09,873 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:43:14,709 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9553, 1.1645, 1.6124, 1.7452, 1.6785, 1.7810, 1.6364, 1.6115], + device='cuda:3'), covar=tensor([0.3599, 0.4898, 0.4163, 0.4186, 0.5292, 0.7075, 0.4608, 0.4556], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0373, 0.0318, 0.0333, 0.0344, 0.0395, 0.0355, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:43:19,855 INFO [finetune.py:976] (3/7) Epoch 18, batch 200, loss[loss=0.1566, simple_loss=0.2318, pruned_loss=0.04074, over 4783.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2433, pruned_loss=0.05417, over 608303.55 frames. ], batch size: 29, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:43:55,466 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:43:55,523 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:43:57,822 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6823, 1.3547, 1.7651, 2.1700, 1.7899, 1.6352, 1.6743, 1.6709], + device='cuda:3'), covar=tensor([0.4775, 0.7093, 0.6369, 0.5636, 0.6091, 0.7502, 0.8052, 0.9202], + device='cuda:3'), in_proj_covar=tensor([0.0421, 0.0408, 0.0498, 0.0505, 0.0451, 0.0477, 0.0484, 0.0487], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:43:58,890 INFO [finetune.py:976] (3/7) Epoch 18, batch 250, loss[loss=0.1444, simple_loss=0.2223, pruned_loss=0.03324, over 4827.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2448, pruned_loss=0.05362, over 686238.20 frames. ], batch size: 30, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:44:42,422 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.696e+02 2.035e+02 2.379e+02 5.416e+02, threshold=4.070e+02, percent-clipped=1.0 +2023-04-27 10:44:49,660 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97656.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:45:03,744 INFO [finetune.py:976] (3/7) Epoch 18, batch 300, loss[loss=0.2324, simple_loss=0.3027, pruned_loss=0.08103, over 4820.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2497, pruned_loss=0.05478, over 746241.92 frames. ], batch size: 38, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:45:10,004 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97674.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:45:43,987 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97698.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 10:45:47,049 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97703.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:45:47,658 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:46:04,758 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0457, 1.6788, 2.1439, 2.5295, 2.1014, 1.9792, 2.0665, 1.9683], + device='cuda:3'), covar=tensor([0.4859, 0.6437, 0.6792, 0.5862, 0.6427, 0.8128, 0.8836, 0.8403], + device='cuda:3'), in_proj_covar=tensor([0.0421, 0.0407, 0.0497, 0.0504, 0.0450, 0.0477, 0.0483, 0.0487], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:46:07,660 INFO [finetune.py:976] (3/7) Epoch 18, batch 350, loss[loss=0.1969, simple_loss=0.2665, pruned_loss=0.06366, over 4906.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2524, pruned_loss=0.05631, over 792622.36 frames. ], batch size: 37, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:46:18,774 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97730.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:46:28,136 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97735.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:46:51,633 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.687e+02 1.979e+02 2.374e+02 3.417e+02, threshold=3.957e+02, percent-clipped=0.0 +2023-04-27 10:47:08,350 INFO [finetune.py:976] (3/7) Epoch 18, batch 400, loss[loss=0.1767, simple_loss=0.2487, pruned_loss=0.05235, over 4722.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.252, pruned_loss=0.05477, over 829124.03 frames. ], batch size: 54, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:47:32,322 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.5329, 4.3644, 3.0819, 5.2250, 4.5455, 4.4821, 1.7723, 4.5079], + device='cuda:3'), covar=tensor([0.1381, 0.0982, 0.2733, 0.0730, 0.2379, 0.1508, 0.5362, 0.1850], + device='cuda:3'), in_proj_covar=tensor([0.0236, 0.0207, 0.0242, 0.0296, 0.0290, 0.0243, 0.0265, 0.0264], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:47:42,004 INFO [finetune.py:976] (3/7) Epoch 18, batch 450, loss[loss=0.1838, simple_loss=0.2545, pruned_loss=0.05654, over 4888.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2493, pruned_loss=0.05355, over 859011.57 frames. ], batch size: 43, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:47:42,724 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7802, 3.0363, 0.9637, 1.8700, 2.3183, 1.6876, 4.4783, 2.6011], + device='cuda:3'), covar=tensor([0.0568, 0.0698, 0.0899, 0.1304, 0.0503, 0.0940, 0.0186, 0.0530], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0074, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 10:48:05,039 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.551e+02 1.835e+02 2.204e+02 3.781e+02, threshold=3.670e+02, percent-clipped=0.0 +2023-04-27 10:48:15,398 INFO [finetune.py:976] (3/7) Epoch 18, batch 500, loss[loss=0.1579, simple_loss=0.234, pruned_loss=0.04092, over 4906.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2469, pruned_loss=0.05288, over 881414.64 frames. ], batch size: 37, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:48:26,743 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:48:42,846 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97911.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:48:45,963 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:48:48,919 INFO [finetune.py:976] (3/7) Epoch 18, batch 550, loss[loss=0.1622, simple_loss=0.2459, pruned_loss=0.03927, over 4896.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2434, pruned_loss=0.05174, over 899082.97 frames. ], batch size: 35, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:48:59,973 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1127, 2.0796, 1.7722, 1.8488, 2.0816, 1.6859, 2.6160, 1.4089], + device='cuda:3'), covar=tensor([0.3795, 0.1855, 0.4867, 0.3030, 0.1802, 0.2674, 0.1655, 0.5174], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0344, 0.0424, 0.0353, 0.0379, 0.0377, 0.0369, 0.0415], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:49:09,113 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7675, 1.9329, 1.2020, 1.5913, 2.0327, 1.6924, 1.5948, 1.6818], + device='cuda:3'), covar=tensor([0.0401, 0.0287, 0.0319, 0.0450, 0.0267, 0.0415, 0.0382, 0.0457], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 10:49:09,140 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97949.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:12,032 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.221e+01 1.571e+02 1.914e+02 2.299e+02 4.493e+02, threshold=3.828e+02, percent-clipped=2.0 +2023-04-27 10:49:18,170 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97964.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:22,482 INFO [finetune.py:976] (3/7) Epoch 18, batch 600, loss[loss=0.1623, simple_loss=0.2305, pruned_loss=0.04704, over 4906.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2448, pruned_loss=0.05239, over 911262.11 frames. ], batch size: 37, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:49:23,145 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6514, 3.8631, 0.8802, 2.1039, 1.8548, 2.4732, 2.2658, 0.9436], + device='cuda:3'), covar=tensor([0.1344, 0.0841, 0.1978, 0.1192, 0.1181, 0.1151, 0.1441, 0.2187], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0242, 0.0137, 0.0120, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:49:35,801 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4891, 2.4910, 2.1311, 2.2473, 2.6503, 2.1920, 3.4321, 1.8877], + device='cuda:3'), covar=tensor([0.3770, 0.2266, 0.4794, 0.3290, 0.1607, 0.2715, 0.1617, 0.4331], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0346, 0.0427, 0.0356, 0.0381, 0.0379, 0.0371, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:49:41,026 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:46,922 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98003.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:57,857 INFO [finetune.py:976] (3/7) Epoch 18, batch 650, loss[loss=0.2577, simple_loss=0.3253, pruned_loss=0.09499, over 4809.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2479, pruned_loss=0.05321, over 920307.66 frames. ], batch size: 51, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:50:03,542 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98030.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:50:03,575 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98030.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:50:14,380 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98046.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:50:18,390 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98051.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:50:26,530 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.729e+02 2.021e+02 2.442e+02 4.739e+02, threshold=4.043e+02, percent-clipped=3.0 +2023-04-27 10:50:48,619 INFO [finetune.py:976] (3/7) Epoch 18, batch 700, loss[loss=0.1905, simple_loss=0.2615, pruned_loss=0.05974, over 4829.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2482, pruned_loss=0.05286, over 929177.67 frames. ], batch size: 47, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:50:58,809 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:51:25,133 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5541, 3.2880, 2.6580, 3.0430, 2.1880, 2.8689, 2.7742, 2.1501], + device='cuda:3'), covar=tensor([0.1981, 0.1147, 0.0728, 0.1037, 0.3420, 0.1091, 0.1838, 0.2601], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0307, 0.0220, 0.0282, 0.0315, 0.0261, 0.0251, 0.0267], + device='cuda:3'), out_proj_covar=tensor([1.1536e-04, 1.2191e-04, 8.7647e-05, 1.1178e-04, 1.2787e-04, 1.0353e-04, + 1.0175e-04, 1.0598e-04], device='cuda:3') +2023-04-27 10:51:33,384 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98106.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:51:52,856 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-27 10:51:53,738 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5763, 1.6905, 1.8846, 1.9553, 1.7765, 1.8638, 2.0443, 2.0405], + device='cuda:3'), covar=tensor([0.4342, 0.6078, 0.4999, 0.5080, 0.6192, 0.8029, 0.5601, 0.5232], + device='cuda:3'), in_proj_covar=tensor([0.0333, 0.0375, 0.0320, 0.0334, 0.0346, 0.0397, 0.0356, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 10:51:54,809 INFO [finetune.py:976] (3/7) Epoch 18, batch 750, loss[loss=0.1781, simple_loss=0.2473, pruned_loss=0.05448, over 4867.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2499, pruned_loss=0.05323, over 937055.27 frames. ], batch size: 31, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:52:28,547 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6328, 1.4674, 0.7037, 1.3116, 1.4575, 1.4851, 1.3498, 1.4090], + device='cuda:3'), covar=tensor([0.0498, 0.0391, 0.0383, 0.0580, 0.0290, 0.0527, 0.0554, 0.0570], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 10:52:31,992 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.562e+02 1.753e+02 1.961e+02 3.537e+02, threshold=3.506e+02, percent-clipped=0.0 +2023-04-27 10:52:42,172 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98167.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:52:44,507 INFO [finetune.py:976] (3/7) Epoch 18, batch 800, loss[loss=0.1784, simple_loss=0.2424, pruned_loss=0.05723, over 4921.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2494, pruned_loss=0.05306, over 940659.39 frames. ], batch size: 38, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:52:50,715 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3897, 3.1910, 1.0350, 1.8125, 1.7385, 2.3329, 1.8976, 1.1130], + device='cuda:3'), covar=tensor([0.1392, 0.0956, 0.1835, 0.1189, 0.1141, 0.0961, 0.1403, 0.1927], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0242, 0.0136, 0.0120, 0.0133, 0.0152, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 10:53:10,422 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98211.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:53:17,963 INFO [finetune.py:976] (3/7) Epoch 18, batch 850, loss[loss=0.167, simple_loss=0.2407, pruned_loss=0.0467, over 4822.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.247, pruned_loss=0.05256, over 942297.69 frames. ], batch size: 39, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:53:32,030 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98244.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:53:38,451 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.514e+02 1.837e+02 2.117e+02 3.312e+02, threshold=3.674e+02, percent-clipped=0.0 +2023-04-27 10:53:42,072 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98259.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:53:51,234 INFO [finetune.py:976] (3/7) Epoch 18, batch 900, loss[loss=0.1531, simple_loss=0.2221, pruned_loss=0.04204, over 4830.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2438, pruned_loss=0.05103, over 945817.31 frames. ], batch size: 38, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:54:13,160 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5541, 2.9693, 0.9073, 1.7016, 2.0241, 1.6285, 4.1546, 2.1062], + device='cuda:3'), covar=tensor([0.0593, 0.0795, 0.0872, 0.1197, 0.0531, 0.0891, 0.0211, 0.0578], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0065, 0.0048, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 10:54:24,149 INFO [finetune.py:976] (3/7) Epoch 18, batch 950, loss[loss=0.1984, simple_loss=0.2661, pruned_loss=0.0654, over 4935.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2428, pruned_loss=0.05102, over 949955.21 frames. ], batch size: 38, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:54:30,291 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9434, 1.3207, 4.6381, 4.3795, 4.0949, 4.3004, 4.1350, 4.0702], + device='cuda:3'), covar=tensor([0.6695, 0.5804, 0.1099, 0.1655, 0.1083, 0.1339, 0.2329, 0.1683], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0305, 0.0401, 0.0405, 0.0347, 0.0403, 0.0311, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 10:54:30,311 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98330.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:54:44,941 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.562e+02 1.785e+02 2.020e+02 3.278e+02, threshold=3.570e+02, percent-clipped=0.0 +2023-04-27 10:54:57,898 INFO [finetune.py:976] (3/7) Epoch 18, batch 1000, loss[loss=0.2518, simple_loss=0.3241, pruned_loss=0.08973, over 4810.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2437, pruned_loss=0.05137, over 950759.67 frames. ], batch size: 40, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:55:01,606 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3335, 2.8665, 0.8217, 1.6203, 1.9746, 1.4863, 4.0475, 2.0315], + device='cuda:3'), covar=tensor([0.0684, 0.0808, 0.0916, 0.1274, 0.0577, 0.0982, 0.0290, 0.0633], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 10:55:02,792 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98378.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:55:22,429 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98410.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:55:27,641 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8762, 1.5374, 1.3615, 1.6635, 2.0725, 1.6470, 1.4805, 1.3402], + device='cuda:3'), covar=tensor([0.1653, 0.1603, 0.2210, 0.1449, 0.0952, 0.1814, 0.2216, 0.2415], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0313, 0.0352, 0.0288, 0.0329, 0.0309, 0.0300, 0.0367], + device='cuda:3'), out_proj_covar=tensor([6.3664e-05, 6.5189e-05, 7.4815e-05, 5.8447e-05, 6.8374e-05, 6.5154e-05, + 6.3125e-05, 7.8148e-05], device='cuda:3') +2023-04-27 10:55:29,959 INFO [finetune.py:976] (3/7) Epoch 18, batch 1050, loss[loss=0.1954, simple_loss=0.2749, pruned_loss=0.05793, over 4925.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.247, pruned_loss=0.05222, over 952808.43 frames. ], batch size: 38, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:55:52,030 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.114e+02 1.610e+02 1.888e+02 2.252e+02 4.818e+02, threshold=3.776e+02, percent-clipped=3.0 +2023-04-27 10:55:56,957 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:56:08,981 INFO [finetune.py:976] (3/7) Epoch 18, batch 1100, loss[loss=0.1562, simple_loss=0.2227, pruned_loss=0.04488, over 4757.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2488, pruned_loss=0.05252, over 952432.09 frames. ], batch size: 27, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:56:09,095 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:57:13,273 INFO [finetune.py:976] (3/7) Epoch 18, batch 1150, loss[loss=0.2189, simple_loss=0.2857, pruned_loss=0.07602, over 4100.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.249, pruned_loss=0.053, over 952209.85 frames. ], batch size: 66, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:57:13,477 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.65 vs. limit=5.0 +2023-04-27 10:57:45,479 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98543.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:57:46,073 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98544.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:57:57,924 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.919e+01 1.656e+02 1.874e+02 2.209e+02 6.337e+02, threshold=3.749e+02, percent-clipped=3.0 +2023-04-27 10:58:14,178 INFO [finetune.py:976] (3/7) Epoch 18, batch 1200, loss[loss=0.1783, simple_loss=0.2397, pruned_loss=0.05839, over 4894.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2492, pruned_loss=0.05361, over 953870.91 frames. ], batch size: 36, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:58:23,065 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1485, 2.7624, 2.1599, 2.1580, 1.5803, 1.5369, 2.2869, 1.4926], + device='cuda:3'), covar=tensor([0.1555, 0.1482, 0.1381, 0.1681, 0.2214, 0.1824, 0.0924, 0.1933], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0167, 0.0204, 0.0200, 0.0184, 0.0156, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 10:58:28,911 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98592.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:58:36,427 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98604.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:58:46,674 INFO [finetune.py:976] (3/7) Epoch 18, batch 1250, loss[loss=0.1605, simple_loss=0.2134, pruned_loss=0.05375, over 4000.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.246, pruned_loss=0.05253, over 953264.98 frames. ], batch size: 17, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:58:52,703 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 10:59:10,194 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.558e+02 1.895e+02 2.216e+02 4.222e+02, threshold=3.789e+02, percent-clipped=2.0 +2023-04-27 10:59:20,473 INFO [finetune.py:976] (3/7) Epoch 18, batch 1300, loss[loss=0.1713, simple_loss=0.2378, pruned_loss=0.05244, over 4760.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2431, pruned_loss=0.0514, over 955009.54 frames. ], batch size: 28, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:59:32,136 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7472, 1.0051, 1.6704, 2.1418, 1.7665, 1.7191, 1.7299, 1.7158], + device='cuda:3'), covar=tensor([0.4949, 0.6897, 0.6345, 0.6406, 0.6271, 0.8139, 0.7970, 0.8317], + device='cuda:3'), in_proj_covar=tensor([0.0420, 0.0406, 0.0495, 0.0501, 0.0449, 0.0476, 0.0482, 0.0486], + device='cuda:3'), out_proj_covar=tensor([1.0110e-04, 9.9927e-05, 1.1140e-04, 1.1974e-04, 1.0770e-04, 1.1434e-04, + 1.1441e-04, 1.1471e-04], device='cuda:3') +2023-04-27 10:59:41,647 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2966, 1.4945, 1.3521, 1.7818, 1.6366, 1.9458, 1.3732, 3.4660], + device='cuda:3'), covar=tensor([0.0560, 0.0787, 0.0801, 0.1155, 0.0616, 0.0509, 0.0744, 0.0157], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 10:59:46,752 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 10:59:53,698 INFO [finetune.py:976] (3/7) Epoch 18, batch 1350, loss[loss=0.2072, simple_loss=0.2879, pruned_loss=0.06327, over 4872.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2437, pruned_loss=0.05226, over 954496.41 frames. ], batch size: 44, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:00:17,037 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.614e+02 2.005e+02 2.367e+02 4.015e+02, threshold=4.011e+02, percent-clipped=1.0 +2023-04-27 11:00:18,415 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 11:00:22,052 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98762.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:00:24,444 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:00:27,468 INFO [finetune.py:976] (3/7) Epoch 18, batch 1400, loss[loss=0.2198, simple_loss=0.2951, pruned_loss=0.07227, over 4934.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.248, pruned_loss=0.0541, over 955090.65 frames. ], batch size: 42, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:00:32,709 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-27 11:00:42,236 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0132, 1.4853, 1.3919, 1.9016, 2.1363, 1.7316, 1.6927, 1.5297], + device='cuda:3'), covar=tensor([0.1965, 0.1942, 0.1795, 0.1390, 0.1276, 0.2791, 0.2319, 0.2237], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0314, 0.0353, 0.0288, 0.0330, 0.0310, 0.0301, 0.0367], + device='cuda:3'), out_proj_covar=tensor([6.3916e-05, 6.5439e-05, 7.4988e-05, 5.8499e-05, 6.8523e-05, 6.5381e-05, + 6.3407e-05, 7.8332e-05], device='cuda:3') +2023-04-27 11:00:49,527 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9580, 2.5070, 2.0070, 2.1468, 1.6208, 1.5856, 2.0745, 1.5102], + device='cuda:3'), covar=tensor([0.1356, 0.1330, 0.1203, 0.1395, 0.1905, 0.1671, 0.0846, 0.1789], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0213, 0.0168, 0.0204, 0.0199, 0.0184, 0.0156, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 11:00:54,325 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98810.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:01:00,975 INFO [finetune.py:976] (3/7) Epoch 18, batch 1450, loss[loss=0.16, simple_loss=0.2313, pruned_loss=0.04436, over 4741.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2487, pruned_loss=0.05341, over 953607.71 frames. ], batch size: 59, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:01:24,436 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.637e+02 1.928e+02 2.404e+02 4.442e+02, threshold=3.855e+02, percent-clipped=1.0 +2023-04-27 11:01:34,709 INFO [finetune.py:976] (3/7) Epoch 18, batch 1500, loss[loss=0.1583, simple_loss=0.2374, pruned_loss=0.03959, over 4819.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2507, pruned_loss=0.05421, over 954697.45 frames. ], batch size: 39, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:01:52,805 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98890.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:02:00,791 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 11:02:02,788 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98895.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:02:11,047 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98899.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:02:35,632 INFO [finetune.py:976] (3/7) Epoch 18, batch 1550, loss[loss=0.1657, simple_loss=0.233, pruned_loss=0.04922, over 4790.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2508, pruned_loss=0.05412, over 955455.90 frames. ], batch size: 25, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:03:17,460 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98951.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:03:19,153 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.300e+01 1.636e+02 1.920e+02 2.366e+02 5.799e+02, threshold=3.840e+02, percent-clipped=2.0 +2023-04-27 11:03:25,945 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98956.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:03:41,674 INFO [finetune.py:976] (3/7) Epoch 18, batch 1600, loss[loss=0.1496, simple_loss=0.2212, pruned_loss=0.03899, over 4803.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2485, pruned_loss=0.05325, over 955325.37 frames. ], batch size: 45, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:03:41,817 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98971.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:04:20,540 INFO [finetune.py:976] (3/7) Epoch 18, batch 1650, loss[loss=0.1499, simple_loss=0.2227, pruned_loss=0.03857, over 4690.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.246, pruned_loss=0.05264, over 955843.59 frames. ], batch size: 23, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:04:21,878 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8910, 1.8451, 2.3155, 2.3612, 1.7683, 1.5506, 1.8491, 1.0255], + device='cuda:3'), covar=tensor([0.0639, 0.0771, 0.0437, 0.0898, 0.0828, 0.1221, 0.0817, 0.0880], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0069, 0.0067, 0.0067, 0.0075, 0.0095, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 11:04:27,358 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:04:40,837 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99051.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:04:42,972 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.609e+02 1.883e+02 2.336e+02 5.190e+02, threshold=3.766e+02, percent-clipped=4.0 +2023-04-27 11:04:50,886 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99066.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:04:53,813 INFO [finetune.py:976] (3/7) Epoch 18, batch 1700, loss[loss=0.2307, simple_loss=0.2891, pruned_loss=0.08613, over 4205.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2454, pruned_loss=0.05282, over 955210.85 frames. ], batch size: 65, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:05:21,713 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99112.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:05:22,864 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99114.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:05:27,558 INFO [finetune.py:976] (3/7) Epoch 18, batch 1750, loss[loss=0.1401, simple_loss=0.2087, pruned_loss=0.03572, over 4777.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2479, pruned_loss=0.0542, over 955325.19 frames. ], batch size: 28, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:05:50,009 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.654e+02 1.955e+02 2.443e+02 4.969e+02, threshold=3.909e+02, percent-clipped=5.0 +2023-04-27 11:06:01,218 INFO [finetune.py:976] (3/7) Epoch 18, batch 1800, loss[loss=0.1884, simple_loss=0.2592, pruned_loss=0.05883, over 4932.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2511, pruned_loss=0.05544, over 955095.42 frames. ], batch size: 33, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:06:04,384 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 11:06:18,773 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99199.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:22,378 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1521, 1.7220, 2.0449, 2.5253, 2.0491, 1.6308, 1.4674, 1.8311], + device='cuda:3'), covar=tensor([0.3010, 0.2919, 0.1554, 0.2039, 0.2446, 0.2462, 0.3911, 0.1944], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0246, 0.0227, 0.0315, 0.0218, 0.0231, 0.0228, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 11:06:34,181 INFO [finetune.py:976] (3/7) Epoch 18, batch 1850, loss[loss=0.1604, simple_loss=0.2452, pruned_loss=0.03778, over 4821.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2527, pruned_loss=0.05642, over 954966.49 frames. ], batch size: 39, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:06:39,647 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:49,927 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99246.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:50,534 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99247.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:53,450 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99251.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:55,189 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5323, 2.6853, 2.2312, 2.3837, 2.7836, 2.2411, 3.5604, 1.9392], + device='cuda:3'), covar=tensor([0.3928, 0.2416, 0.4735, 0.3080, 0.1767, 0.2663, 0.1327, 0.4171], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0348, 0.0429, 0.0355, 0.0383, 0.0381, 0.0372, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:06:55,634 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.735e+02 2.090e+02 2.546e+02 5.570e+02, threshold=4.180e+02, percent-clipped=4.0 +2023-04-27 11:07:06,163 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99268.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:07:07,878 INFO [finetune.py:976] (3/7) Epoch 18, batch 1900, loss[loss=0.1562, simple_loss=0.2267, pruned_loss=0.04284, over 4754.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2513, pruned_loss=0.05469, over 955647.08 frames. ], batch size: 28, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:07:20,138 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99290.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:08:03,838 INFO [finetune.py:976] (3/7) Epoch 18, batch 1950, loss[loss=0.1948, simple_loss=0.2605, pruned_loss=0.06459, over 4811.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2493, pruned_loss=0.05368, over 955430.07 frames. ], batch size: 40, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:08:05,844 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99324.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:08:07,588 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99327.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:08:09,363 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:08:30,192 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.408e+01 1.522e+02 1.826e+02 2.209e+02 4.570e+02, threshold=3.652e+02, percent-clipped=1.0 +2023-04-27 11:08:52,739 INFO [finetune.py:976] (3/7) Epoch 18, batch 2000, loss[loss=0.1411, simple_loss=0.2146, pruned_loss=0.03379, over 4906.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2456, pruned_loss=0.05229, over 952197.59 frames. ], batch size: 43, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:09:11,137 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9063, 1.4743, 4.8033, 4.5040, 4.2208, 4.4813, 4.2392, 4.2151], + device='cuda:3'), covar=tensor([0.6703, 0.5840, 0.1071, 0.1770, 0.1061, 0.1447, 0.1799, 0.1523], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0308, 0.0404, 0.0407, 0.0351, 0.0404, 0.0311, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:09:13,019 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:09:24,408 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3637, 1.4258, 1.7108, 1.7301, 1.4364, 1.2375, 1.5150, 1.0240], + device='cuda:3'), covar=tensor([0.0672, 0.0454, 0.0413, 0.0389, 0.0827, 0.0958, 0.0560, 0.0561], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 11:09:25,616 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5782, 1.6590, 1.4780, 1.1627, 1.2245, 1.2036, 1.4650, 1.1512], + device='cuda:3'), covar=tensor([0.1610, 0.1318, 0.1438, 0.1628, 0.2219, 0.1814, 0.1061, 0.1966], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0213, 0.0168, 0.0205, 0.0201, 0.0185, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 11:09:44,021 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99407.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:09:49,625 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8382, 3.8292, 2.7287, 4.4872, 3.9095, 3.9627, 1.5587, 3.7728], + device='cuda:3'), covar=tensor([0.1854, 0.1168, 0.3048, 0.1627, 0.3823, 0.1851, 0.6614, 0.2488], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0214, 0.0249, 0.0305, 0.0298, 0.0249, 0.0272, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:09:54,356 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99420.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:09:54,850 INFO [finetune.py:976] (3/7) Epoch 18, batch 2050, loss[loss=0.1552, simple_loss=0.2302, pruned_loss=0.04015, over 4901.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.243, pruned_loss=0.05202, over 950721.06 frames. ], batch size: 35, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:10:06,377 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4461, 1.0699, 1.1754, 1.2045, 1.6558, 1.3212, 1.1011, 1.1268], + device='cuda:3'), covar=tensor([0.1545, 0.1204, 0.1825, 0.1261, 0.0648, 0.1255, 0.1811, 0.2136], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0313, 0.0351, 0.0288, 0.0329, 0.0309, 0.0301, 0.0367], + device='cuda:3'), out_proj_covar=tensor([6.3985e-05, 6.5160e-05, 7.4673e-05, 5.8459e-05, 6.8268e-05, 6.5070e-05, + 6.3447e-05, 7.8320e-05], device='cuda:3') +2023-04-27 11:10:15,959 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.635e+02 1.952e+02 2.334e+02 5.427e+02, threshold=3.904e+02, percent-clipped=2.0 +2023-04-27 11:10:26,032 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99468.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:10:28,727 INFO [finetune.py:976] (3/7) Epoch 18, batch 2100, loss[loss=0.1626, simple_loss=0.2261, pruned_loss=0.04958, over 4765.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2433, pruned_loss=0.05283, over 949018.19 frames. ], batch size: 27, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:10:34,268 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.38 vs. limit=5.0 +2023-04-27 11:10:35,442 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99481.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:02,725 INFO [finetune.py:976] (3/7) Epoch 18, batch 2150, loss[loss=0.1649, simple_loss=0.244, pruned_loss=0.04292, over 4155.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2475, pruned_loss=0.05394, over 949472.27 frames. ], batch size: 65, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:11:08,197 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99529.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:11:18,914 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99546.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:20,758 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1810, 1.4598, 1.3133, 1.6984, 1.5931, 1.8897, 1.3727, 3.4335], + device='cuda:3'), covar=tensor([0.0600, 0.0811, 0.0806, 0.1186, 0.0650, 0.0526, 0.0764, 0.0142], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 11:11:21,996 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99551.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:23,739 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.609e+02 1.810e+02 2.358e+02 5.458e+02, threshold=3.621e+02, percent-clipped=4.0 +2023-04-27 11:11:32,856 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99568.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:35,036 INFO [finetune.py:976] (3/7) Epoch 18, batch 2200, loss[loss=0.1642, simple_loss=0.2316, pruned_loss=0.0484, over 4788.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2501, pruned_loss=0.05447, over 952565.32 frames. ], batch size: 26, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:11:45,026 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99585.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:47,520 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99589.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:51,021 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99594.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:54,027 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99599.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:08,620 INFO [finetune.py:976] (3/7) Epoch 18, batch 2250, loss[loss=0.2056, simple_loss=0.2649, pruned_loss=0.07316, over 4806.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2509, pruned_loss=0.05465, over 952460.43 frames. ], batch size: 45, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:12:11,000 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99624.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:11,115 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 11:12:13,833 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99627.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:15,109 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99629.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:28,860 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:31,121 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.508e+01 1.589e+02 1.867e+02 2.262e+02 4.430e+02, threshold=3.734e+02, percent-clipped=1.0 +2023-04-27 11:12:41,913 INFO [finetune.py:976] (3/7) Epoch 18, batch 2300, loss[loss=0.1972, simple_loss=0.274, pruned_loss=0.06026, over 4769.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2506, pruned_loss=0.05406, over 953913.66 frames. ], batch size: 51, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:12:45,388 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99675.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:49,371 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:49,397 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:13:01,086 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-04-27 11:13:12,570 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:13:26,439 INFO [finetune.py:976] (3/7) Epoch 18, batch 2350, loss[loss=0.2119, simple_loss=0.2812, pruned_loss=0.07128, over 4812.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2482, pruned_loss=0.053, over 953986.03 frames. ], batch size: 40, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:13:34,809 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 11:13:56,986 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99741.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:14:16,628 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.602e+02 1.875e+02 2.250e+02 5.277e+02, threshold=3.750e+02, percent-clipped=1.0 +2023-04-27 11:14:17,849 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99755.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:14:18,499 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3920, 2.1939, 1.8520, 1.9259, 2.3438, 1.8409, 2.5955, 1.6556], + device='cuda:3'), covar=tensor([0.3582, 0.1829, 0.4589, 0.2925, 0.1504, 0.2469, 0.1515, 0.4563], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0346, 0.0425, 0.0353, 0.0380, 0.0378, 0.0368, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:14:26,618 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 11:14:33,758 INFO [finetune.py:976] (3/7) Epoch 18, batch 2400, loss[loss=0.1737, simple_loss=0.2466, pruned_loss=0.05035, over 4931.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2447, pruned_loss=0.0519, over 954046.14 frames. ], batch size: 33, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:14:36,883 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99776.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:15:35,837 INFO [finetune.py:976] (3/7) Epoch 18, batch 2450, loss[loss=0.1626, simple_loss=0.2343, pruned_loss=0.04547, over 4872.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2422, pruned_loss=0.05129, over 955756.57 frames. ], batch size: 34, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:15:37,742 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99824.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:16:08,396 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.453e+01 1.597e+02 1.870e+02 2.238e+02 4.750e+02, threshold=3.741e+02, percent-clipped=1.0 +2023-04-27 11:16:19,143 INFO [finetune.py:976] (3/7) Epoch 18, batch 2500, loss[loss=0.1886, simple_loss=0.255, pruned_loss=0.06109, over 4755.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2429, pruned_loss=0.05147, over 955408.28 frames. ], batch size: 54, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:16:28,672 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99885.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:16:36,191 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3708, 3.3938, 2.5385, 4.0437, 3.5519, 3.3903, 1.9215, 3.4204], + device='cuda:3'), covar=tensor([0.1913, 0.1475, 0.3720, 0.1945, 0.3063, 0.1969, 0.5258, 0.2618], + device='cuda:3'), in_proj_covar=tensor([0.0248, 0.0217, 0.0253, 0.0309, 0.0301, 0.0252, 0.0276, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:16:50,583 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3098, 2.7699, 1.0169, 1.5613, 2.2047, 1.4624, 3.7831, 2.0899], + device='cuda:3'), covar=tensor([0.0644, 0.0623, 0.0786, 0.1257, 0.0470, 0.0933, 0.0274, 0.0563], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0049, 0.0052, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 11:16:52,953 INFO [finetune.py:976] (3/7) Epoch 18, batch 2550, loss[loss=0.2039, simple_loss=0.2689, pruned_loss=0.06943, over 4105.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.247, pruned_loss=0.0532, over 954189.73 frames. ], batch size: 65, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:16:54,879 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:16:54,897 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:00,811 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:00,850 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4440, 2.9114, 0.8114, 1.6159, 2.3085, 1.4696, 4.0906, 1.9730], + device='cuda:3'), covar=tensor([0.0643, 0.0773, 0.0888, 0.1289, 0.0502, 0.1023, 0.0198, 0.0585], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0049, 0.0052, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 11:17:10,140 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:16,095 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.590e+02 1.888e+02 2.384e+02 3.870e+02, threshold=3.776e+02, percent-clipped=2.0 +2023-04-27 11:17:26,891 INFO [finetune.py:976] (3/7) Epoch 18, batch 2600, loss[loss=0.2188, simple_loss=0.2923, pruned_loss=0.07263, over 4830.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2479, pruned_loss=0.05272, over 956170.79 frames. ], batch size: 39, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:17:27,555 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:28,953 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 11:17:32,445 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:36,136 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7131, 2.2226, 1.8930, 1.7203, 1.2348, 1.3530, 1.9153, 1.2244], + device='cuda:3'), covar=tensor([0.1768, 0.1409, 0.1261, 0.1528, 0.2306, 0.1922, 0.0914, 0.2047], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0212, 0.0168, 0.0205, 0.0200, 0.0184, 0.0156, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 11:17:52,675 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100007.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:18:01,563 INFO [finetune.py:976] (3/7) Epoch 18, batch 2650, loss[loss=0.179, simple_loss=0.2588, pruned_loss=0.04961, over 4849.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2471, pruned_loss=0.05205, over 952696.39 frames. ], batch size: 44, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:18:05,843 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100028.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:18:10,683 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:18:24,402 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.666e+02 1.842e+02 2.203e+02 3.197e+02, threshold=3.685e+02, percent-clipped=0.0 +2023-04-27 11:18:33,085 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100068.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:18:35,203 INFO [finetune.py:976] (3/7) Epoch 18, batch 2700, loss[loss=0.1874, simple_loss=0.2547, pruned_loss=0.06008, over 4816.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2463, pruned_loss=0.05117, over 953199.31 frames. ], batch size: 40, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:18:38,368 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100076.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:18:53,815 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2232, 2.4580, 2.2988, 2.4639, 2.2428, 2.3743, 2.4124, 2.3130], + device='cuda:3'), covar=tensor([0.4036, 0.5834, 0.5102, 0.4703, 0.5956, 0.7274, 0.6498, 0.6121], + device='cuda:3'), in_proj_covar=tensor([0.0331, 0.0371, 0.0318, 0.0332, 0.0342, 0.0392, 0.0353, 0.0324], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:19:37,242 INFO [finetune.py:976] (3/7) Epoch 18, batch 2750, loss[loss=0.1373, simple_loss=0.2115, pruned_loss=0.03157, over 4803.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2441, pruned_loss=0.05088, over 953604.08 frames. ], batch size: 25, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:19:39,077 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:19:39,124 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100124.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:20:11,922 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.977e+01 1.603e+02 1.951e+02 2.448e+02 3.827e+02, threshold=3.902e+02, percent-clipped=2.0 +2023-04-27 11:20:28,164 INFO [finetune.py:976] (3/7) Epoch 18, batch 2800, loss[loss=0.1668, simple_loss=0.2266, pruned_loss=0.05348, over 4915.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2416, pruned_loss=0.05039, over 953071.83 frames. ], batch size: 37, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:20:34,238 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100172.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:20:40,242 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1261, 1.8488, 2.3229, 2.4907, 2.1836, 2.0726, 2.1612, 2.2188], + device='cuda:3'), covar=tensor([0.5497, 0.7326, 0.8370, 0.6942, 0.6651, 0.9768, 0.9937, 0.9713], + device='cuda:3'), in_proj_covar=tensor([0.0423, 0.0408, 0.0498, 0.0502, 0.0452, 0.0477, 0.0484, 0.0489], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:20:47,553 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 11:21:07,287 INFO [finetune.py:976] (3/7) Epoch 18, batch 2850, loss[loss=0.2487, simple_loss=0.3012, pruned_loss=0.09814, over 4844.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.242, pruned_loss=0.05122, over 954571.84 frames. ], batch size: 44, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:21:15,059 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100224.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:21:29,447 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100238.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:21:39,955 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100245.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:21:52,185 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.601e+02 1.939e+02 2.362e+02 3.802e+02, threshold=3.878e+02, percent-clipped=0.0 +2023-04-27 11:22:14,789 INFO [finetune.py:976] (3/7) Epoch 18, batch 2900, loss[loss=0.2027, simple_loss=0.2636, pruned_loss=0.0709, over 4915.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2462, pruned_loss=0.05259, over 953323.71 frames. ], batch size: 36, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:22:15,466 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:22:28,745 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:22:32,448 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:22:37,925 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 11:22:48,679 INFO [finetune.py:976] (3/7) Epoch 18, batch 2950, loss[loss=0.1627, simple_loss=0.2378, pruned_loss=0.0438, over 4832.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.249, pruned_loss=0.05306, over 953644.42 frames. ], batch size: 47, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:22:54,839 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6310, 3.9824, 0.7344, 2.2706, 2.2350, 2.5230, 2.3674, 1.0377], + device='cuda:3'), covar=tensor([0.1432, 0.0923, 0.2218, 0.1197, 0.1044, 0.1128, 0.1409, 0.2097], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0242, 0.0136, 0.0120, 0.0132, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 11:22:57,897 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:23:01,414 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-04-27 11:23:09,851 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.535e+02 1.848e+02 2.418e+02 4.913e+02, threshold=3.697e+02, percent-clipped=1.0 +2023-04-27 11:23:15,824 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100363.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:23:21,745 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 11:23:22,041 INFO [finetune.py:976] (3/7) Epoch 18, batch 3000, loss[loss=0.1976, simple_loss=0.2691, pruned_loss=0.06304, over 4908.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2484, pruned_loss=0.05279, over 950698.68 frames. ], batch size: 36, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:23:22,041 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 11:23:32,629 INFO [finetune.py:1010] (3/7) Epoch 18, validation: loss=0.1524, simple_loss=0.2231, pruned_loss=0.04086, over 2265189.00 frames. +2023-04-27 11:23:32,630 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 11:23:41,189 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:23:52,518 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100402.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:23:56,867 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 11:24:04,565 INFO [finetune.py:976] (3/7) Epoch 18, batch 3050, loss[loss=0.1668, simple_loss=0.2172, pruned_loss=0.05823, over 4101.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2488, pruned_loss=0.05254, over 949248.43 frames. ], batch size: 17, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:24:43,542 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.525e+02 1.858e+02 2.098e+02 3.467e+02, threshold=3.716e+02, percent-clipped=0.0 +2023-04-27 11:24:55,032 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100463.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:24:59,804 INFO [finetune.py:976] (3/7) Epoch 18, batch 3100, loss[loss=0.1705, simple_loss=0.2309, pruned_loss=0.05505, over 4245.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2478, pruned_loss=0.05233, over 948285.59 frames. ], batch size: 18, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:25:39,568 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 11:26:02,762 INFO [finetune.py:976] (3/7) Epoch 18, batch 3150, loss[loss=0.1572, simple_loss=0.2155, pruned_loss=0.04949, over 4803.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2461, pruned_loss=0.05263, over 950124.40 frames. ], batch size: 25, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:26:23,110 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.0289, 2.2530, 2.2012, 2.3270, 2.0971, 2.2674, 2.3785, 2.2501], + device='cuda:3'), covar=tensor([0.4740, 0.6533, 0.5331, 0.4584, 0.6366, 0.7511, 0.6189, 0.5674], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0374, 0.0320, 0.0333, 0.0343, 0.0393, 0.0355, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:26:37,705 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.912e+01 1.520e+02 1.805e+02 2.268e+02 8.838e+02, threshold=3.610e+02, percent-clipped=4.0 +2023-04-27 11:26:59,500 INFO [finetune.py:976] (3/7) Epoch 18, batch 3200, loss[loss=0.1539, simple_loss=0.2276, pruned_loss=0.04006, over 4823.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2415, pruned_loss=0.05078, over 950754.67 frames. ], batch size: 38, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:27:33,044 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100594.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:27:40,545 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 11:27:55,352 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.66 vs. limit=5.0 +2023-04-27 11:28:02,071 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6833, 3.5797, 2.7929, 4.3041, 3.6604, 3.7071, 1.6189, 3.6109], + device='cuda:3'), covar=tensor([0.1712, 0.1296, 0.3320, 0.1427, 0.3959, 0.1656, 0.5908, 0.2562], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0215, 0.0250, 0.0305, 0.0297, 0.0248, 0.0271, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:28:06,109 INFO [finetune.py:976] (3/7) Epoch 18, batch 3250, loss[loss=0.2112, simple_loss=0.2612, pruned_loss=0.0806, over 4766.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2431, pruned_loss=0.05145, over 952830.90 frames. ], batch size: 28, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:28:14,719 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7611, 1.6942, 1.6546, 1.2684, 1.7987, 1.5267, 2.2670, 1.4120], + device='cuda:3'), covar=tensor([0.3258, 0.1721, 0.4402, 0.2868, 0.1640, 0.2225, 0.1414, 0.4713], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0347, 0.0432, 0.0354, 0.0384, 0.0382, 0.0371, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:28:30,300 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.424e+01 1.529e+02 1.873e+02 2.219e+02 4.824e+02, threshold=3.746e+02, percent-clipped=4.0 +2023-04-27 11:28:35,348 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100663.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:28:40,641 INFO [finetune.py:976] (3/7) Epoch 18, batch 3300, loss[loss=0.1839, simple_loss=0.2522, pruned_loss=0.0578, over 4877.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2461, pruned_loss=0.05205, over 953263.70 frames. ], batch size: 31, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:29:07,778 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100711.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:29:13,775 INFO [finetune.py:976] (3/7) Epoch 18, batch 3350, loss[loss=0.1848, simple_loss=0.2556, pruned_loss=0.05699, over 4801.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2491, pruned_loss=0.05361, over 952992.49 frames. ], batch size: 45, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:29:19,916 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100730.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:29:33,304 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 11:29:37,231 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.219e+01 1.703e+02 2.119e+02 2.646e+02 1.102e+03, threshold=4.237e+02, percent-clipped=5.0 +2023-04-27 11:29:39,140 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100758.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:29:47,004 INFO [finetune.py:976] (3/7) Epoch 18, batch 3400, loss[loss=0.1666, simple_loss=0.2455, pruned_loss=0.04385, over 4806.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2511, pruned_loss=0.0544, over 953629.65 frames. ], batch size: 29, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:30:00,852 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100791.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:30:20,269 INFO [finetune.py:976] (3/7) Epoch 18, batch 3450, loss[loss=0.166, simple_loss=0.2317, pruned_loss=0.05019, over 4763.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2503, pruned_loss=0.05377, over 954594.88 frames. ], batch size: 27, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:30:22,209 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9221, 2.3663, 1.9571, 2.2742, 1.7980, 2.0819, 2.1473, 1.7151], + device='cuda:3'), covar=tensor([0.1681, 0.0960, 0.0707, 0.0906, 0.2460, 0.0851, 0.1405, 0.1988], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0306, 0.0220, 0.0282, 0.0313, 0.0262, 0.0251, 0.0268], + device='cuda:3'), out_proj_covar=tensor([1.1563e-04, 1.2156e-04, 8.7283e-05, 1.1189e-04, 1.2711e-04, 1.0381e-04, + 1.0140e-04, 1.0649e-04], device='cuda:3') +2023-04-27 11:30:45,556 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 11:30:54,165 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.615e+02 1.948e+02 2.378e+02 4.172e+02, threshold=3.896e+02, percent-clipped=0.0 +2023-04-27 11:31:08,670 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-04-27 11:31:09,718 INFO [finetune.py:976] (3/7) Epoch 18, batch 3500, loss[loss=0.1774, simple_loss=0.2463, pruned_loss=0.0542, over 4895.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2478, pruned_loss=0.0529, over 955989.51 frames. ], batch size: 32, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:31:22,391 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6802, 1.6181, 1.9244, 1.9625, 1.5318, 1.4226, 1.6063, 1.0614], + device='cuda:3'), covar=tensor([0.0536, 0.0563, 0.0391, 0.0545, 0.0696, 0.1097, 0.0558, 0.0595], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0068, 0.0067, 0.0067, 0.0074, 0.0095, 0.0073, 0.0066], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 11:31:29,655 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 11:31:30,618 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100894.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:31:47,774 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0160, 2.7550, 1.9735, 2.2907, 1.6891, 1.5806, 1.9904, 1.5757], + device='cuda:3'), covar=tensor([0.1545, 0.1373, 0.1519, 0.1577, 0.2223, 0.2035, 0.1049, 0.1970], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0215, 0.0170, 0.0206, 0.0201, 0.0186, 0.0157, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 11:31:49,440 INFO [finetune.py:976] (3/7) Epoch 18, batch 3550, loss[loss=0.1476, simple_loss=0.2147, pruned_loss=0.04029, over 4745.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2438, pruned_loss=0.05123, over 956008.97 frames. ], batch size: 27, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:31:59,814 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2039, 1.7195, 2.0598, 2.6779, 1.9690, 1.6549, 1.4609, 1.8655], + device='cuda:3'), covar=tensor([0.3204, 0.3258, 0.1774, 0.2140, 0.3033, 0.2869, 0.4296, 0.2215], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0245, 0.0224, 0.0312, 0.0217, 0.0230, 0.0227, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 11:32:02,729 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100942.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:32:23,484 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.524e+02 1.758e+02 2.158e+02 5.191e+02, threshold=3.516e+02, percent-clipped=1.0 +2023-04-27 11:32:39,052 INFO [finetune.py:976] (3/7) Epoch 18, batch 3600, loss[loss=0.1783, simple_loss=0.2516, pruned_loss=0.05251, over 4886.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2409, pruned_loss=0.05047, over 954161.80 frames. ], batch size: 35, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:33:40,675 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5686, 1.2544, 4.0745, 3.7700, 3.5594, 3.7999, 3.7195, 3.5759], + device='cuda:3'), covar=tensor([0.7137, 0.5951, 0.1074, 0.1906, 0.1138, 0.1388, 0.2255, 0.1431], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0307, 0.0405, 0.0409, 0.0351, 0.0405, 0.0313, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:33:50,172 INFO [finetune.py:976] (3/7) Epoch 18, batch 3650, loss[loss=0.1066, simple_loss=0.1769, pruned_loss=0.01813, over 4678.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2437, pruned_loss=0.0514, over 953085.21 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:34:31,992 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.646e+02 1.981e+02 2.447e+02 4.488e+02, threshold=3.961e+02, percent-clipped=2.0 +2023-04-27 11:34:34,871 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101058.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:34:48,426 INFO [finetune.py:976] (3/7) Epoch 18, batch 3700, loss[loss=0.2044, simple_loss=0.2721, pruned_loss=0.0684, over 4868.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2467, pruned_loss=0.05253, over 952483.49 frames. ], batch size: 31, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:34:57,641 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101086.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 11:35:10,798 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101106.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:35:11,930 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101107.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:35:22,248 INFO [finetune.py:976] (3/7) Epoch 18, batch 3750, loss[loss=0.1588, simple_loss=0.218, pruned_loss=0.04983, over 4746.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2479, pruned_loss=0.05336, over 953035.67 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:35:43,442 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.572e+02 1.895e+02 2.251e+02 5.219e+02, threshold=3.790e+02, percent-clipped=2.0 +2023-04-27 11:35:54,454 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101168.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:35:56,187 INFO [finetune.py:976] (3/7) Epoch 18, batch 3800, loss[loss=0.1797, simple_loss=0.248, pruned_loss=0.05575, over 4847.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2493, pruned_loss=0.05411, over 952122.64 frames. ], batch size: 44, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:36:25,188 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 11:36:30,058 INFO [finetune.py:976] (3/7) Epoch 18, batch 3850, loss[loss=0.1751, simple_loss=0.2446, pruned_loss=0.05282, over 4882.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2482, pruned_loss=0.0532, over 952725.40 frames. ], batch size: 32, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:36:46,579 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3222, 1.7552, 2.1336, 2.6571, 2.1379, 1.6908, 1.2790, 1.9808], + device='cuda:3'), covar=tensor([0.3404, 0.3277, 0.1760, 0.2101, 0.2682, 0.2827, 0.4161, 0.1946], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0244, 0.0224, 0.0311, 0.0216, 0.0229, 0.0226, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 11:36:50,615 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.513e+02 1.818e+02 2.217e+02 6.339e+02, threshold=3.636e+02, percent-clipped=4.0 +2023-04-27 11:37:01,752 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4914, 1.7214, 1.9063, 1.9962, 1.9275, 1.9386, 1.9959, 1.9827], + device='cuda:3'), covar=tensor([0.4303, 0.5155, 0.4723, 0.4705, 0.5623, 0.7287, 0.5247, 0.4552], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0376, 0.0322, 0.0335, 0.0346, 0.0396, 0.0358, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:37:02,693 INFO [finetune.py:976] (3/7) Epoch 18, batch 3900, loss[loss=0.1651, simple_loss=0.241, pruned_loss=0.04458, over 4737.00 frames. ], tot_loss[loss=0.176, simple_loss=0.246, pruned_loss=0.05298, over 952396.71 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:37:23,704 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7177, 1.0251, 1.6943, 2.2338, 1.8282, 1.6551, 1.7002, 1.6808], + device='cuda:3'), covar=tensor([0.4525, 0.6232, 0.6394, 0.5756, 0.5606, 0.7648, 0.7589, 0.8247], + device='cuda:3'), in_proj_covar=tensor([0.0421, 0.0406, 0.0497, 0.0501, 0.0451, 0.0476, 0.0484, 0.0487], + device='cuda:3'), out_proj_covar=tensor([1.0145e-04, 9.9723e-05, 1.1175e-04, 1.1954e-04, 1.0806e-04, 1.1424e-04, + 1.1469e-04, 1.1486e-04], device='cuda:3') +2023-04-27 11:37:35,479 INFO [finetune.py:976] (3/7) Epoch 18, batch 3950, loss[loss=0.1508, simple_loss=0.2208, pruned_loss=0.04042, over 4823.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2428, pruned_loss=0.05214, over 951071.34 frames. ], batch size: 38, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:37:44,554 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7053, 1.8212, 0.8496, 1.3934, 1.8811, 1.6009, 1.4337, 1.5945], + device='cuda:3'), covar=tensor([0.0496, 0.0363, 0.0367, 0.0559, 0.0257, 0.0530, 0.0541, 0.0534], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 11:38:09,080 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.783e+01 1.521e+02 1.787e+02 2.150e+02 4.001e+02, threshold=3.574e+02, percent-clipped=1.0 +2023-04-27 11:38:30,022 INFO [finetune.py:976] (3/7) Epoch 18, batch 4000, loss[loss=0.2213, simple_loss=0.29, pruned_loss=0.07636, over 4928.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2424, pruned_loss=0.05213, over 951322.16 frames. ], batch size: 38, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:38:40,633 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5281, 1.7991, 1.8136, 1.9140, 1.7718, 1.8318, 1.9260, 1.8620], + device='cuda:3'), covar=tensor([0.4339, 0.5469, 0.4631, 0.4465, 0.5462, 0.7531, 0.5585, 0.5011], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0376, 0.0323, 0.0335, 0.0347, 0.0396, 0.0359, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:38:52,315 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101386.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:39:34,184 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5287, 1.5729, 0.8734, 1.2363, 1.8181, 1.4461, 1.3000, 1.3758], + device='cuda:3'), covar=tensor([0.0492, 0.0374, 0.0346, 0.0550, 0.0281, 0.0490, 0.0483, 0.0568], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 11:39:35,329 INFO [finetune.py:976] (3/7) Epoch 18, batch 4050, loss[loss=0.1668, simple_loss=0.2435, pruned_loss=0.04506, over 4236.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2469, pruned_loss=0.05384, over 950123.17 frames. ], batch size: 65, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:39:46,407 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2372, 1.8283, 2.1120, 2.6164, 2.1880, 1.7837, 1.8003, 2.0476], + device='cuda:3'), covar=tensor([0.2466, 0.2584, 0.1384, 0.2041, 0.2303, 0.2223, 0.3548, 0.1917], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0244, 0.0224, 0.0312, 0.0217, 0.0229, 0.0226, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 11:39:55,122 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101434.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:40:01,878 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0330, 2.3685, 0.9956, 1.3292, 1.7538, 1.2456, 3.0198, 1.4982], + device='cuda:3'), covar=tensor([0.0680, 0.0627, 0.0773, 0.1382, 0.0529, 0.1069, 0.0289, 0.0734], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0048, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 11:40:07,832 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.179e+02 1.675e+02 1.988e+02 2.422e+02 4.320e+02, threshold=3.976e+02, percent-clipped=3.0 +2023-04-27 11:40:12,815 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101463.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:40:18,125 INFO [finetune.py:976] (3/7) Epoch 18, batch 4100, loss[loss=0.1808, simple_loss=0.2542, pruned_loss=0.05373, over 4826.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2499, pruned_loss=0.0545, over 951497.00 frames. ], batch size: 33, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:40:31,811 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 11:40:46,532 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5461, 3.4991, 2.6642, 4.0926, 3.5613, 3.4739, 1.7824, 3.5293], + device='cuda:3'), covar=tensor([0.2133, 0.1329, 0.3463, 0.2209, 0.3210, 0.2235, 0.5808, 0.2944], + device='cuda:3'), in_proj_covar=tensor([0.0249, 0.0218, 0.0254, 0.0308, 0.0303, 0.0252, 0.0277, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:40:51,379 INFO [finetune.py:976] (3/7) Epoch 18, batch 4150, loss[loss=0.1677, simple_loss=0.2424, pruned_loss=0.04651, over 4775.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2516, pruned_loss=0.05487, over 952981.98 frames. ], batch size: 28, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:41:11,722 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-04-27 11:41:13,287 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3158, 1.6294, 1.5526, 1.9702, 1.7768, 1.9745, 1.5747, 4.0939], + device='cuda:3'), covar=tensor([0.0605, 0.0909, 0.0867, 0.1243, 0.0682, 0.0601, 0.0841, 0.0147], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 11:41:14,401 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.437e+01 1.627e+02 2.025e+02 2.401e+02 3.721e+02, threshold=4.051e+02, percent-clipped=0.0 +2023-04-27 11:41:24,218 INFO [finetune.py:976] (3/7) Epoch 18, batch 4200, loss[loss=0.1723, simple_loss=0.2454, pruned_loss=0.04956, over 4851.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2506, pruned_loss=0.05351, over 953463.45 frames. ], batch size: 44, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:41:58,016 INFO [finetune.py:976] (3/7) Epoch 18, batch 4250, loss[loss=0.131, simple_loss=0.2073, pruned_loss=0.02742, over 4867.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2479, pruned_loss=0.0526, over 951902.73 frames. ], batch size: 31, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:42:06,360 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4547, 1.7313, 1.6626, 1.8092, 1.7643, 1.8551, 1.7814, 1.7416], + device='cuda:3'), covar=tensor([0.3850, 0.5333, 0.4944, 0.4261, 0.5494, 0.7213, 0.5898, 0.5285], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0374, 0.0321, 0.0334, 0.0345, 0.0395, 0.0358, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:42:21,984 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.254e+01 1.472e+02 1.774e+02 2.178e+02 3.033e+02, threshold=3.548e+02, percent-clipped=0.0 +2023-04-27 11:42:31,626 INFO [finetune.py:976] (3/7) Epoch 18, batch 4300, loss[loss=0.1353, simple_loss=0.2071, pruned_loss=0.03177, over 4796.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2458, pruned_loss=0.05228, over 951989.75 frames. ], batch size: 25, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:42:40,101 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101684.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:43:04,508 INFO [finetune.py:976] (3/7) Epoch 18, batch 4350, loss[loss=0.147, simple_loss=0.2187, pruned_loss=0.03766, over 4748.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.242, pruned_loss=0.05094, over 952464.41 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:43:09,591 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 11:43:20,699 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101745.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:43:32,959 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.731e+01 1.464e+02 1.839e+02 2.155e+02 7.335e+02, threshold=3.679e+02, percent-clipped=2.0 +2023-04-27 11:43:43,337 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101763.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:43:48,147 INFO [finetune.py:976] (3/7) Epoch 18, batch 4400, loss[loss=0.2079, simple_loss=0.2697, pruned_loss=0.07303, over 4751.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2436, pruned_loss=0.05187, over 953194.73 frames. ], batch size: 27, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:44:22,041 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3370, 3.4280, 0.8727, 1.7159, 1.6602, 2.3729, 1.9872, 1.0874], + device='cuda:3'), covar=tensor([0.1793, 0.1625, 0.2481, 0.1805, 0.1432, 0.1304, 0.1706, 0.2139], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0246, 0.0138, 0.0122, 0.0134, 0.0154, 0.0119, 0.0121], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 11:44:37,288 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101811.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:44:49,150 INFO [finetune.py:976] (3/7) Epoch 18, batch 4450, loss[loss=0.1679, simple_loss=0.2545, pruned_loss=0.04069, over 4819.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2472, pruned_loss=0.05282, over 954388.57 frames. ], batch size: 51, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:44:58,796 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1973, 1.1995, 3.8442, 3.5764, 3.4385, 3.7088, 3.7214, 3.3713], + device='cuda:3'), covar=tensor([0.7031, 0.5738, 0.1187, 0.1805, 0.1111, 0.1880, 0.1398, 0.1510], + device='cuda:3'), in_proj_covar=tensor([0.0304, 0.0303, 0.0400, 0.0403, 0.0346, 0.0402, 0.0309, 0.0363], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:45:11,574 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.02 vs. limit=5.0 +2023-04-27 11:45:32,216 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.703e+02 1.979e+02 2.478e+02 5.839e+02, threshold=3.957e+02, percent-clipped=3.0 +2023-04-27 11:45:32,392 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7809, 1.0491, 1.7052, 2.2050, 1.8311, 1.6552, 1.6790, 1.6881], + device='cuda:3'), covar=tensor([0.4700, 0.6589, 0.6435, 0.5880, 0.5840, 0.8184, 0.8299, 0.8286], + device='cuda:3'), in_proj_covar=tensor([0.0423, 0.0407, 0.0499, 0.0503, 0.0452, 0.0477, 0.0484, 0.0489], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 11:45:42,461 INFO [finetune.py:976] (3/7) Epoch 18, batch 4500, loss[loss=0.1918, simple_loss=0.2751, pruned_loss=0.05426, over 4815.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2485, pruned_loss=0.05251, over 954734.30 frames. ], batch size: 45, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:45:55,435 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5680, 0.9567, 1.5810, 2.0708, 1.6656, 1.4985, 1.5456, 1.5282], + device='cuda:3'), covar=tensor([0.4486, 0.6582, 0.6147, 0.6046, 0.5699, 0.7672, 0.7985, 0.8452], + device='cuda:3'), in_proj_covar=tensor([0.0423, 0.0406, 0.0498, 0.0502, 0.0451, 0.0476, 0.0484, 0.0488], + device='cuda:3'), out_proj_covar=tensor([1.0183e-04, 9.9790e-05, 1.1200e-04, 1.1974e-04, 1.0808e-04, 1.1424e-04, + 1.1466e-04, 1.1526e-04], device='cuda:3') +2023-04-27 11:46:15,938 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101920.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:46:16,439 INFO [finetune.py:976] (3/7) Epoch 18, batch 4550, loss[loss=0.1735, simple_loss=0.2393, pruned_loss=0.05388, over 4907.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2505, pruned_loss=0.05337, over 954393.70 frames. ], batch size: 37, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:46:25,797 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7441, 1.4049, 1.3159, 1.5021, 1.9129, 1.4769, 1.2444, 1.3078], + device='cuda:3'), covar=tensor([0.1547, 0.1424, 0.1981, 0.1194, 0.0865, 0.1746, 0.2064, 0.2070], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0314, 0.0353, 0.0291, 0.0331, 0.0309, 0.0303, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.3948e-05, 6.5312e-05, 7.4986e-05, 5.9060e-05, 6.8816e-05, 6.5004e-05, + 6.3761e-05, 7.9282e-05], device='cuda:3') +2023-04-27 11:46:35,987 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 11:46:38,530 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.683e+02 1.915e+02 2.320e+02 4.613e+02, threshold=3.831e+02, percent-clipped=2.0 +2023-04-27 11:46:49,873 INFO [finetune.py:976] (3/7) Epoch 18, batch 4600, loss[loss=0.1653, simple_loss=0.2412, pruned_loss=0.04467, over 4755.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2504, pruned_loss=0.05347, over 954473.69 frames. ], batch size: 26, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:46:56,093 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101981.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:46:59,720 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:47:24,662 INFO [finetune.py:976] (3/7) Epoch 18, batch 4650, loss[loss=0.1307, simple_loss=0.2095, pruned_loss=0.0259, over 4794.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2476, pruned_loss=0.05248, over 955785.09 frames. ], batch size: 26, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:47:36,317 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:47:41,236 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102048.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:47:45,411 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.010e+02 1.596e+02 1.902e+02 2.192e+02 3.719e+02, threshold=3.804e+02, percent-clipped=0.0 +2023-04-27 11:47:58,102 INFO [finetune.py:976] (3/7) Epoch 18, batch 4700, loss[loss=0.1496, simple_loss=0.2209, pruned_loss=0.03915, over 4755.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2429, pruned_loss=0.05048, over 956717.35 frames. ], batch size: 26, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:48:15,317 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.91 vs. limit=5.0 +2023-04-27 11:48:29,309 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 11:48:30,322 INFO [finetune.py:976] (3/7) Epoch 18, batch 4750, loss[loss=0.2164, simple_loss=0.2634, pruned_loss=0.08472, over 4026.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2413, pruned_loss=0.05016, over 954415.51 frames. ], batch size: 66, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:48:47,201 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8964, 2.4720, 1.9547, 1.8074, 1.4552, 1.4447, 2.1044, 1.3818], + device='cuda:3'), covar=tensor([0.1703, 0.1385, 0.1426, 0.1670, 0.2142, 0.1896, 0.0915, 0.1990], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0168, 0.0205, 0.0201, 0.0185, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 11:48:50,290 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8111, 1.2769, 1.8673, 2.2653, 1.8585, 1.8179, 1.8577, 1.7521], + device='cuda:3'), covar=tensor([0.4628, 0.6538, 0.6301, 0.5741, 0.5819, 0.7607, 0.7301, 0.7928], + device='cuda:3'), in_proj_covar=tensor([0.0422, 0.0406, 0.0499, 0.0502, 0.0452, 0.0476, 0.0484, 0.0489], + device='cuda:3'), out_proj_covar=tensor([1.0176e-04, 9.9898e-05, 1.1214e-04, 1.1960e-04, 1.0809e-04, 1.1427e-04, + 1.1466e-04, 1.1529e-04], device='cuda:3') +2023-04-27 11:48:51,941 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.496e+02 1.838e+02 2.058e+02 4.267e+02, threshold=3.676e+02, percent-clipped=2.0 +2023-04-27 11:49:09,300 INFO [finetune.py:976] (3/7) Epoch 18, batch 4800, loss[loss=0.2234, simple_loss=0.3026, pruned_loss=0.07208, over 4818.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2443, pruned_loss=0.05126, over 954090.73 frames. ], batch size: 38, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:50:15,445 INFO [finetune.py:976] (3/7) Epoch 18, batch 4850, loss[loss=0.1937, simple_loss=0.2685, pruned_loss=0.05946, over 4832.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2477, pruned_loss=0.052, over 955075.39 frames. ], batch size: 30, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:50:58,312 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.240e+02 1.715e+02 2.079e+02 2.399e+02 4.707e+02, threshold=4.157e+02, percent-clipped=1.0 +2023-04-27 11:51:19,609 INFO [finetune.py:976] (3/7) Epoch 18, batch 4900, loss[loss=0.1816, simple_loss=0.2639, pruned_loss=0.04965, over 4860.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2491, pruned_loss=0.05285, over 952971.73 frames. ], batch size: 31, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:51:23,242 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102276.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:52:17,696 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8122, 2.8294, 2.1733, 3.2484, 2.8033, 2.8052, 1.2840, 2.7492], + device='cuda:3'), covar=tensor([0.2287, 0.1740, 0.3664, 0.3293, 0.2774, 0.2265, 0.5242, 0.3052], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0216, 0.0251, 0.0304, 0.0299, 0.0250, 0.0273, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 11:52:25,539 INFO [finetune.py:976] (3/7) Epoch 18, batch 4950, loss[loss=0.1438, simple_loss=0.2168, pruned_loss=0.03542, over 4756.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2502, pruned_loss=0.05295, over 954091.88 frames. ], batch size: 28, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:52:50,469 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102340.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:52:57,585 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102343.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:53:04,800 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.629e+02 1.929e+02 2.326e+02 4.533e+02, threshold=3.857e+02, percent-clipped=1.0 +2023-04-27 11:53:09,775 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-04-27 11:53:14,474 INFO [finetune.py:976] (3/7) Epoch 18, batch 5000, loss[loss=0.1502, simple_loss=0.2205, pruned_loss=0.04, over 4793.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2483, pruned_loss=0.05278, over 954470.61 frames. ], batch size: 29, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:53:28,308 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102388.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:53:48,401 INFO [finetune.py:976] (3/7) Epoch 18, batch 5050, loss[loss=0.1313, simple_loss=0.2086, pruned_loss=0.02697, over 4829.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2456, pruned_loss=0.05234, over 953677.10 frames. ], batch size: 39, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:54:23,902 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.725e+01 1.630e+02 1.921e+02 2.242e+02 3.810e+02, threshold=3.842e+02, percent-clipped=0.0 +2023-04-27 11:54:45,116 INFO [finetune.py:976] (3/7) Epoch 18, batch 5100, loss[loss=0.1605, simple_loss=0.2316, pruned_loss=0.04471, over 4832.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2424, pruned_loss=0.05124, over 955041.03 frames. ], batch size: 47, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:55:18,948 INFO [finetune.py:976] (3/7) Epoch 18, batch 5150, loss[loss=0.221, simple_loss=0.2892, pruned_loss=0.07645, over 4926.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2439, pruned_loss=0.05189, over 955968.70 frames. ], batch size: 42, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:55:52,827 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.715e+02 2.103e+02 2.437e+02 3.988e+02, threshold=4.205e+02, percent-clipped=0.0 +2023-04-27 11:56:08,073 INFO [finetune.py:976] (3/7) Epoch 18, batch 5200, loss[loss=0.1652, simple_loss=0.2517, pruned_loss=0.03934, over 4820.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2472, pruned_loss=0.05295, over 954246.13 frames. ], batch size: 39, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:56:11,160 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102576.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:56:42,077 INFO [finetune.py:976] (3/7) Epoch 18, batch 5250, loss[loss=0.174, simple_loss=0.2387, pruned_loss=0.05467, over 4801.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2487, pruned_loss=0.05345, over 952366.03 frames. ], batch size: 25, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:56:43,955 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102624.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:56:45,761 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4674, 1.2400, 0.6355, 1.1671, 1.4750, 1.3627, 1.2296, 1.2964], + device='cuda:3'), covar=tensor([0.0500, 0.0405, 0.0406, 0.0567, 0.0298, 0.0525, 0.0511, 0.0598], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 11:56:57,447 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:57:11,279 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.529e+02 1.797e+02 2.306e+02 3.318e+02, threshold=3.594e+02, percent-clipped=0.0 +2023-04-27 11:57:26,676 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-27 11:57:26,878 INFO [finetune.py:976] (3/7) Epoch 18, batch 5300, loss[loss=0.1835, simple_loss=0.255, pruned_loss=0.05602, over 4807.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2497, pruned_loss=0.0533, over 951768.68 frames. ], batch size: 39, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:57:55,087 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102691.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:58:31,735 INFO [finetune.py:976] (3/7) Epoch 18, batch 5350, loss[loss=0.1489, simple_loss=0.2269, pruned_loss=0.03547, over 4811.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2492, pruned_loss=0.05255, over 952884.07 frames. ], batch size: 41, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:59:22,079 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.383e+01 1.595e+02 1.837e+02 2.151e+02 4.228e+02, threshold=3.674e+02, percent-clipped=3.0 +2023-04-27 11:59:23,498 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-27 11:59:37,386 INFO [finetune.py:976] (3/7) Epoch 18, batch 5400, loss[loss=0.1563, simple_loss=0.2169, pruned_loss=0.04778, over 4717.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2458, pruned_loss=0.05152, over 953144.59 frames. ], batch size: 23, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:00:49,107 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9330, 1.5054, 2.0168, 2.3601, 2.0194, 1.8650, 1.9172, 1.9106], + device='cuda:3'), covar=tensor([0.4539, 0.5793, 0.5881, 0.5761, 0.6034, 0.7796, 0.7926, 0.6965], + device='cuda:3'), in_proj_covar=tensor([0.0423, 0.0407, 0.0498, 0.0502, 0.0450, 0.0477, 0.0483, 0.0489], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:00:49,560 INFO [finetune.py:976] (3/7) Epoch 18, batch 5450, loss[loss=0.1866, simple_loss=0.2456, pruned_loss=0.06376, over 3977.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2422, pruned_loss=0.05072, over 952972.49 frames. ], batch size: 17, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:00:59,787 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5289, 1.5605, 1.8590, 1.8713, 1.3972, 1.3369, 1.5780, 0.9757], + device='cuda:3'), covar=tensor([0.0628, 0.0744, 0.0410, 0.0606, 0.0845, 0.1091, 0.0670, 0.0653], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0075, 0.0095, 0.0073, 0.0066], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:01:23,556 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102851.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:01:33,555 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.567e+02 1.920e+02 2.316e+02 4.548e+02, threshold=3.840e+02, percent-clipped=3.0 +2023-04-27 12:01:44,221 INFO [finetune.py:976] (3/7) Epoch 18, batch 5500, loss[loss=0.1626, simple_loss=0.234, pruned_loss=0.04556, over 4895.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2398, pruned_loss=0.04983, over 954137.25 frames. ], batch size: 35, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:01:59,973 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102887.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:02:17,481 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102912.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:02:23,856 INFO [finetune.py:976] (3/7) Epoch 18, batch 5550, loss[loss=0.1611, simple_loss=0.2269, pruned_loss=0.04762, over 4729.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2408, pruned_loss=0.05002, over 954996.79 frames. ], batch size: 23, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:02:40,492 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102948.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:02:42,845 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.9176, 4.8752, 3.1687, 5.5926, 4.8343, 4.9187, 2.3627, 4.8464], + device='cuda:3'), covar=tensor([0.1405, 0.0735, 0.2913, 0.0838, 0.4275, 0.1441, 0.5297, 0.2003], + device='cuda:3'), in_proj_covar=tensor([0.0241, 0.0213, 0.0247, 0.0303, 0.0297, 0.0246, 0.0268, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 12:02:45,149 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.541e+02 1.915e+02 2.352e+02 3.794e+02, threshold=3.830e+02, percent-clipped=0.0 +2023-04-27 12:02:54,417 INFO [finetune.py:976] (3/7) Epoch 18, batch 5600, loss[loss=0.194, simple_loss=0.2642, pruned_loss=0.06189, over 4861.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2436, pruned_loss=0.05108, over 951740.98 frames. ], batch size: 31, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:03:02,676 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 12:03:05,458 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4638, 1.6551, 1.8279, 1.9693, 1.8934, 1.9601, 1.9312, 1.9166], + device='cuda:3'), covar=tensor([0.4439, 0.5832, 0.4884, 0.4875, 0.5443, 0.7312, 0.5738, 0.5081], + device='cuda:3'), in_proj_covar=tensor([0.0332, 0.0373, 0.0321, 0.0331, 0.0343, 0.0394, 0.0356, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 12:03:16,572 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2859, 1.4673, 1.3434, 1.4471, 1.2386, 1.2331, 1.2847, 1.0470], + device='cuda:3'), covar=tensor([0.1647, 0.1234, 0.0881, 0.1184, 0.3590, 0.1234, 0.1705, 0.2169], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0305, 0.0218, 0.0281, 0.0312, 0.0260, 0.0251, 0.0267], + device='cuda:3'), out_proj_covar=tensor([1.1563e-04, 1.2135e-04, 8.6656e-05, 1.1175e-04, 1.2667e-04, 1.0323e-04, + 1.0126e-04, 1.0571e-04], device='cuda:3') +2023-04-27 12:03:24,714 INFO [finetune.py:976] (3/7) Epoch 18, batch 5650, loss[loss=0.1848, simple_loss=0.2653, pruned_loss=0.05212, over 4845.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2475, pruned_loss=0.05171, over 951598.58 frames. ], batch size: 44, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:03:39,295 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2664, 1.4375, 1.3649, 1.4527, 1.2739, 1.1988, 1.3563, 1.0218], + device='cuda:3'), covar=tensor([0.1290, 0.1041, 0.0763, 0.0923, 0.2664, 0.0994, 0.1343, 0.1813], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0305, 0.0219, 0.0282, 0.0312, 0.0260, 0.0251, 0.0267], + device='cuda:3'), out_proj_covar=tensor([1.1587e-04, 1.2145e-04, 8.6826e-05, 1.1196e-04, 1.2689e-04, 1.0328e-04, + 1.0149e-04, 1.0604e-04], device='cuda:3') +2023-04-27 12:03:46,255 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.540e+02 1.863e+02 2.247e+02 6.465e+02, threshold=3.725e+02, percent-clipped=2.0 +2023-04-27 12:03:55,213 INFO [finetune.py:976] (3/7) Epoch 18, batch 5700, loss[loss=0.13, simple_loss=0.2005, pruned_loss=0.02976, over 3956.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.244, pruned_loss=0.05171, over 931617.52 frames. ], batch size: 17, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:04:24,056 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103098.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:04:24,533 INFO [finetune.py:976] (3/7) Epoch 19, batch 0, loss[loss=0.2184, simple_loss=0.2851, pruned_loss=0.07588, over 4898.00 frames. ], tot_loss[loss=0.2184, simple_loss=0.2851, pruned_loss=0.07588, over 4898.00 frames. ], batch size: 37, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:04:24,533 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 12:04:35,098 INFO [finetune.py:1010] (3/7) Epoch 19, validation: loss=0.1545, simple_loss=0.2248, pruned_loss=0.04209, over 2265189.00 frames. +2023-04-27 12:04:35,099 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 12:04:56,251 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103132.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:05:02,843 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 12:05:07,369 INFO [finetune.py:976] (3/7) Epoch 19, batch 50, loss[loss=0.1634, simple_loss=0.2385, pruned_loss=0.04412, over 4879.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2525, pruned_loss=0.0534, over 217614.26 frames. ], batch size: 43, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:05:13,237 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 12:05:13,472 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.545e+02 1.855e+02 2.320e+02 4.324e+02, threshold=3.710e+02, percent-clipped=2.0 +2023-04-27 12:05:15,437 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103159.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:05:57,685 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103193.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:06:01,236 INFO [finetune.py:976] (3/7) Epoch 19, batch 100, loss[loss=0.1793, simple_loss=0.2506, pruned_loss=0.05399, over 4860.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2468, pruned_loss=0.053, over 382635.24 frames. ], batch size: 31, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:06:13,379 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103207.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:07:01,895 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103243.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:07:06,071 INFO [finetune.py:976] (3/7) Epoch 19, batch 150, loss[loss=0.1549, simple_loss=0.2271, pruned_loss=0.04133, over 4819.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2414, pruned_loss=0.05162, over 509968.43 frames. ], batch size: 40, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:07:16,415 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.567e+02 1.850e+02 2.251e+02 4.056e+02, threshold=3.701e+02, percent-clipped=1.0 +2023-04-27 12:07:27,812 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4155, 3.3996, 0.9543, 1.8993, 1.8865, 2.4332, 2.1134, 0.9978], + device='cuda:3'), covar=tensor([0.1654, 0.1225, 0.2187, 0.1430, 0.1206, 0.1229, 0.1457, 0.2254], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0243, 0.0137, 0.0120, 0.0132, 0.0153, 0.0118, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:07:35,083 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3144, 1.5493, 1.5593, 1.9195, 1.7476, 1.9141, 1.5253, 3.6193], + device='cuda:3'), covar=tensor([0.0614, 0.0811, 0.0806, 0.1149, 0.0639, 0.0502, 0.0740, 0.0137], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 12:07:45,051 INFO [finetune.py:976] (3/7) Epoch 19, batch 200, loss[loss=0.1769, simple_loss=0.2439, pruned_loss=0.0549, over 4822.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2413, pruned_loss=0.05251, over 609505.19 frames. ], batch size: 41, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:07:55,052 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6809, 1.8197, 0.7324, 1.3122, 1.9231, 1.5888, 1.4593, 1.5226], + device='cuda:3'), covar=tensor([0.0490, 0.0344, 0.0352, 0.0550, 0.0257, 0.0504, 0.0477, 0.0569], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0050, 0.0050], + device='cuda:3') +2023-04-27 12:07:56,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5891, 1.3851, 4.4214, 4.1665, 3.7681, 4.1866, 4.0992, 3.9020], + device='cuda:3'), covar=tensor([0.7071, 0.6359, 0.1203, 0.1864, 0.1220, 0.1949, 0.1405, 0.1576], + device='cuda:3'), in_proj_covar=tensor([0.0305, 0.0306, 0.0404, 0.0406, 0.0347, 0.0405, 0.0313, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:08:13,280 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 12:08:33,911 INFO [finetune.py:976] (3/7) Epoch 19, batch 250, loss[loss=0.1254, simple_loss=0.204, pruned_loss=0.02341, over 4767.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2452, pruned_loss=0.05257, over 687441.29 frames. ], batch size: 28, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:08:44,147 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.956e+01 1.630e+02 1.978e+02 2.388e+02 4.596e+02, threshold=3.957e+02, percent-clipped=1.0 +2023-04-27 12:08:56,543 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103366.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 12:09:09,294 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 12:09:22,448 INFO [finetune.py:976] (3/7) Epoch 19, batch 300, loss[loss=0.1425, simple_loss=0.2244, pruned_loss=0.03034, over 4926.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2468, pruned_loss=0.0522, over 748316.33 frames. ], batch size: 33, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:09:42,577 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103427.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 12:09:55,813 INFO [finetune.py:976] (3/7) Epoch 19, batch 350, loss[loss=0.1494, simple_loss=0.2248, pruned_loss=0.037, over 4923.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2488, pruned_loss=0.05245, over 794524.13 frames. ], batch size: 33, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:09:59,435 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103454.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:10:00,584 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.010e+02 1.607e+02 1.938e+02 2.366e+02 5.284e+02, threshold=3.875e+02, percent-clipped=4.0 +2023-04-27 12:10:17,148 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 12:10:20,046 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6775, 2.2372, 2.5746, 3.2490, 2.6083, 2.0205, 1.9814, 2.5106], + device='cuda:3'), covar=tensor([0.3056, 0.3056, 0.1559, 0.2177, 0.2548, 0.2450, 0.3705, 0.1990], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0245, 0.0226, 0.0313, 0.0218, 0.0231, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 12:10:22,422 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103488.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:10:29,035 INFO [finetune.py:976] (3/7) Epoch 19, batch 400, loss[loss=0.1615, simple_loss=0.2299, pruned_loss=0.04652, over 4757.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2496, pruned_loss=0.05256, over 830812.57 frames. ], batch size: 27, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:10:30,681 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 12:10:35,038 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103507.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:10:47,359 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.78 vs. limit=5.0 +2023-04-27 12:10:57,127 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8215, 2.3238, 1.8642, 1.6750, 1.4019, 1.3822, 1.8607, 1.3349], + device='cuda:3'), covar=tensor([0.1645, 0.1309, 0.1449, 0.1677, 0.2250, 0.1924, 0.0986, 0.1951], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0213, 0.0169, 0.0205, 0.0201, 0.0185, 0.0157, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 12:10:58,915 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103543.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:11:02,474 INFO [finetune.py:976] (3/7) Epoch 19, batch 450, loss[loss=0.1664, simple_loss=0.2264, pruned_loss=0.05321, over 3730.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2481, pruned_loss=0.05253, over 858009.07 frames. ], batch size: 16, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:11:06,185 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:11:06,718 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.516e+02 1.781e+02 2.073e+02 5.569e+02, threshold=3.563e+02, percent-clipped=1.0 +2023-04-27 12:11:34,260 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2934, 3.2082, 2.5602, 3.8209, 3.2586, 3.2881, 1.4874, 3.2730], + device='cuda:3'), covar=tensor([0.1867, 0.1367, 0.3379, 0.2209, 0.2751, 0.1803, 0.5299, 0.2394], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0213, 0.0247, 0.0302, 0.0297, 0.0245, 0.0269, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 12:11:42,149 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103591.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:11:46,990 INFO [finetune.py:976] (3/7) Epoch 19, batch 500, loss[loss=0.1669, simple_loss=0.2281, pruned_loss=0.05281, over 4942.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2461, pruned_loss=0.05234, over 880184.13 frames. ], batch size: 33, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:12:16,135 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9509, 2.1866, 2.1134, 2.3066, 2.0916, 2.2095, 2.2530, 2.1529], + device='cuda:3'), covar=tensor([0.3847, 0.6290, 0.5036, 0.4889, 0.5940, 0.7372, 0.6505, 0.6054], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0374, 0.0323, 0.0334, 0.0347, 0.0396, 0.0359, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 12:12:30,576 INFO [finetune.py:976] (3/7) Epoch 19, batch 550, loss[loss=0.1602, simple_loss=0.2275, pruned_loss=0.04649, over 4861.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2438, pruned_loss=0.05179, over 898630.20 frames. ], batch size: 47, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:12:30,671 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6581, 1.3302, 4.5540, 4.2591, 3.9025, 4.3556, 4.1987, 4.0265], + device='cuda:3'), covar=tensor([0.6886, 0.6111, 0.0897, 0.1574, 0.1003, 0.1326, 0.1231, 0.1425], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0307, 0.0406, 0.0408, 0.0350, 0.0408, 0.0314, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:12:34,853 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.551e+02 1.854e+02 2.248e+02 4.068e+02, threshold=3.707e+02, percent-clipped=1.0 +2023-04-27 12:12:38,336 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6680, 2.6824, 2.1944, 3.1016, 2.6503, 2.6799, 1.3411, 2.6991], + device='cuda:3'), covar=tensor([0.2375, 0.1829, 0.4064, 0.3322, 0.3864, 0.2216, 0.4976, 0.2920], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0213, 0.0248, 0.0303, 0.0297, 0.0245, 0.0269, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 12:13:01,825 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0627, 2.7164, 0.9351, 1.5888, 1.8648, 1.2682, 3.5120, 1.9344], + device='cuda:3'), covar=tensor([0.0684, 0.0615, 0.0787, 0.1108, 0.0544, 0.0938, 0.0317, 0.0549], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 12:13:02,466 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0814, 1.9091, 2.4728, 2.6827, 1.8571, 1.7956, 2.1154, 1.0569], + device='cuda:3'), covar=tensor([0.0567, 0.0767, 0.0390, 0.0702, 0.0807, 0.1082, 0.0742, 0.0817], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0067, 0.0067, 0.0066, 0.0074, 0.0094, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:13:04,203 INFO [finetune.py:976] (3/7) Epoch 19, batch 600, loss[loss=0.159, simple_loss=0.2245, pruned_loss=0.04669, over 4708.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2433, pruned_loss=0.05188, over 909172.80 frames. ], batch size: 23, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:13:05,506 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4003, 1.8679, 1.7756, 2.1745, 2.0778, 2.0354, 1.8350, 4.4977], + device='cuda:3'), covar=tensor([0.0537, 0.0757, 0.0748, 0.1125, 0.0566, 0.0528, 0.0687, 0.0118], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 12:13:08,600 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 12:13:19,610 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103722.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 12:13:43,215 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103740.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:13:53,494 INFO [finetune.py:976] (3/7) Epoch 19, batch 650, loss[loss=0.1862, simple_loss=0.2484, pruned_loss=0.06203, over 4886.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2467, pruned_loss=0.05312, over 919642.64 frames. ], batch size: 32, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:14:01,813 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103754.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:02,872 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.593e+02 1.866e+02 2.212e+02 4.115e+02, threshold=3.733e+02, percent-clipped=2.0 +2023-04-27 12:14:14,701 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:30,690 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103788.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:37,245 INFO [finetune.py:976] (3/7) Epoch 19, batch 700, loss[loss=0.1534, simple_loss=0.2334, pruned_loss=0.0367, over 4726.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2487, pruned_loss=0.05383, over 927554.13 frames. ], batch size: 59, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:14:38,099 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1306, 1.7102, 1.9873, 2.4313, 2.4287, 2.0010, 1.8063, 2.2710], + device='cuda:3'), covar=tensor([0.0699, 0.1068, 0.0609, 0.0448, 0.0484, 0.0753, 0.0610, 0.0459], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0196, 0.0177, 0.0167, 0.0173, 0.0177, 0.0147, 0.0173], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:14:38,184 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 12:14:38,726 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103801.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:39,253 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103802.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:55,305 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103826.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:15:02,440 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103836.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:15:10,878 INFO [finetune.py:976] (3/7) Epoch 19, batch 750, loss[loss=0.219, simple_loss=0.2842, pruned_loss=0.07689, over 4813.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2484, pruned_loss=0.05362, over 932141.55 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:15:15,074 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.609e+02 1.947e+02 2.389e+02 3.942e+02, threshold=3.894e+02, percent-clipped=2.0 +2023-04-27 12:15:28,388 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103876.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:15:44,007 INFO [finetune.py:976] (3/7) Epoch 19, batch 800, loss[loss=0.2056, simple_loss=0.2671, pruned_loss=0.07206, over 4826.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.248, pruned_loss=0.05314, over 937409.77 frames. ], batch size: 30, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:16:09,171 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103937.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 12:16:17,416 INFO [finetune.py:976] (3/7) Epoch 19, batch 850, loss[loss=0.1316, simple_loss=0.1979, pruned_loss=0.03262, over 4780.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2445, pruned_loss=0.05159, over 943910.52 frames. ], batch size: 28, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:16:21,648 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.943e+01 1.500e+02 1.730e+02 2.145e+02 3.862e+02, threshold=3.461e+02, percent-clipped=0.0 +2023-04-27 12:16:59,790 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103990.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:17:02,208 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1677, 2.5511, 1.0341, 1.4791, 1.7956, 1.4555, 3.3157, 1.9156], + device='cuda:3'), covar=tensor([0.0625, 0.0573, 0.0820, 0.1255, 0.0541, 0.0938, 0.0230, 0.0602], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 12:17:05,177 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103997.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:17:10,830 INFO [finetune.py:976] (3/7) Epoch 19, batch 900, loss[loss=0.1972, simple_loss=0.2705, pruned_loss=0.06192, over 4891.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2424, pruned_loss=0.05099, over 946054.54 frames. ], batch size: 35, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:17:21,005 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 12:17:43,378 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104022.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 12:17:57,079 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0280, 2.4862, 1.0027, 1.3934, 1.7668, 1.3412, 3.2538, 1.7673], + device='cuda:3'), covar=tensor([0.0687, 0.0589, 0.0760, 0.1313, 0.0542, 0.0993, 0.0331, 0.0641], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 12:18:18,255 INFO [finetune.py:976] (3/7) Epoch 19, batch 950, loss[loss=0.1373, simple_loss=0.2239, pruned_loss=0.02535, over 4899.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2409, pruned_loss=0.0505, over 948238.87 frames. ], batch size: 43, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:18:18,354 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=104049.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:18:19,599 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104051.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:18:21,377 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5170, 3.3855, 1.0248, 1.9626, 1.8430, 2.5770, 1.9341, 1.1033], + device='cuda:3'), covar=tensor([0.1360, 0.1006, 0.1917, 0.1182, 0.1104, 0.0853, 0.1572, 0.1853], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0242, 0.0137, 0.0120, 0.0132, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:18:27,365 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1967, 1.5641, 1.4082, 1.6806, 1.5207, 1.8984, 1.3925, 3.3613], + device='cuda:3'), covar=tensor([0.0607, 0.0726, 0.0756, 0.1222, 0.0651, 0.0516, 0.0752, 0.0146], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 12:18:27,839 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.269e+02 1.627e+02 1.900e+02 2.232e+02 4.587e+02, threshold=3.799e+02, percent-clipped=1.0 +2023-04-27 12:18:29,209 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104058.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:18:41,643 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104070.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 12:19:14,740 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104096.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:19:21,310 INFO [finetune.py:976] (3/7) Epoch 19, batch 1000, loss[loss=0.2164, simple_loss=0.3012, pruned_loss=0.06586, over 4852.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2443, pruned_loss=0.05171, over 951649.11 frames. ], batch size: 44, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:19:33,862 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:19:41,922 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7687, 1.8746, 1.8930, 1.5157, 1.8269, 1.5259, 2.3480, 1.4755], + device='cuda:3'), covar=tensor([0.3389, 0.1652, 0.3838, 0.2560, 0.1548, 0.2497, 0.1349, 0.4298], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0350, 0.0429, 0.0355, 0.0382, 0.0381, 0.0374, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:19:51,327 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104121.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:20:16,418 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 12:20:25,268 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-04-27 12:20:25,623 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 12:20:27,135 INFO [finetune.py:976] (3/7) Epoch 19, batch 1050, loss[loss=0.1715, simple_loss=0.2486, pruned_loss=0.04719, over 4906.00 frames. ], tot_loss[loss=0.175, simple_loss=0.246, pruned_loss=0.05196, over 953062.55 frames. ], batch size: 46, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:20:37,628 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.418e+01 1.570e+02 1.795e+02 2.236e+02 4.469e+02, threshold=3.589e+02, percent-clipped=1.0 +2023-04-27 12:21:05,398 INFO [finetune.py:976] (3/7) Epoch 19, batch 1100, loss[loss=0.1727, simple_loss=0.2465, pruned_loss=0.04946, over 4883.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2476, pruned_loss=0.05229, over 953802.80 frames. ], batch size: 32, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:21:14,545 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.25 vs. limit=5.0 +2023-04-27 12:21:19,721 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9314, 2.6233, 2.8093, 3.3712, 3.1849, 2.8004, 2.2574, 2.9470], + device='cuda:3'), covar=tensor([0.0749, 0.0847, 0.0562, 0.0437, 0.0521, 0.0675, 0.0726, 0.0498], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0198, 0.0179, 0.0169, 0.0174, 0.0179, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:21:27,607 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104232.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 12:21:39,374 INFO [finetune.py:976] (3/7) Epoch 19, batch 1150, loss[loss=0.1703, simple_loss=0.2422, pruned_loss=0.04916, over 4916.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2489, pruned_loss=0.05294, over 954916.61 frames. ], batch size: 38, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:21:44,616 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.137e+01 1.582e+02 1.911e+02 2.328e+02 3.877e+02, threshold=3.822e+02, percent-clipped=1.0 +2023-04-27 12:21:51,728 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5327, 1.3595, 1.7922, 1.7443, 1.3823, 1.2718, 1.4420, 0.9637], + device='cuda:3'), covar=tensor([0.0580, 0.0694, 0.0400, 0.0530, 0.0918, 0.1275, 0.0578, 0.0630], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0075, 0.0095, 0.0074, 0.0066], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:22:12,702 INFO [finetune.py:976] (3/7) Epoch 19, batch 1200, loss[loss=0.1469, simple_loss=0.2171, pruned_loss=0.03831, over 4763.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2481, pruned_loss=0.05253, over 955885.01 frames. ], batch size: 28, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:22:27,974 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1953, 2.8810, 2.4994, 2.7094, 2.0637, 2.5028, 2.6951, 1.8385], + device='cuda:3'), covar=tensor([0.2238, 0.1171, 0.0673, 0.1069, 0.2944, 0.1039, 0.1844, 0.2779], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0302, 0.0216, 0.0278, 0.0310, 0.0257, 0.0250, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1516e-04, 1.1985e-04, 8.5663e-05, 1.1031e-04, 1.2601e-04, 1.0208e-04, + 1.0084e-04, 1.0496e-04], device='cuda:3') +2023-04-27 12:22:44,721 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104346.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:22:46,459 INFO [finetune.py:976] (3/7) Epoch 19, batch 1250, loss[loss=0.1679, simple_loss=0.2323, pruned_loss=0.05171, over 4869.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2452, pruned_loss=0.05114, over 957133.64 frames. ], batch size: 34, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:22:49,474 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104353.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:22:50,131 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7385, 1.5169, 1.3568, 1.6206, 1.9492, 1.6155, 1.4072, 1.3232], + device='cuda:3'), covar=tensor([0.1413, 0.1290, 0.1515, 0.1287, 0.0846, 0.1415, 0.1923, 0.1830], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0311, 0.0348, 0.0289, 0.0328, 0.0307, 0.0300, 0.0366], + device='cuda:3'), out_proj_covar=tensor([6.2994e-05, 6.4563e-05, 7.3854e-05, 5.8700e-05, 6.8185e-05, 6.4498e-05, + 6.2936e-05, 7.8026e-05], device='cuda:3') +2023-04-27 12:22:51,233 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.445e+01 1.483e+02 1.801e+02 2.223e+02 4.756e+02, threshold=3.603e+02, percent-clipped=1.0 +2023-04-27 12:23:45,753 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:23:47,993 INFO [finetune.py:976] (3/7) Epoch 19, batch 1300, loss[loss=0.1811, simple_loss=0.2493, pruned_loss=0.05643, over 4773.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2417, pruned_loss=0.05, over 955717.66 frames. ], batch size: 28, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:23:56,154 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104405.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:24:18,860 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104421.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:24:44,515 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104444.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:24:53,499 INFO [finetune.py:976] (3/7) Epoch 19, batch 1350, loss[loss=0.1826, simple_loss=0.2649, pruned_loss=0.05017, over 4865.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2406, pruned_loss=0.0493, over 956068.73 frames. ], batch size: 44, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:25:03,935 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.562e+02 1.851e+02 2.195e+02 3.087e+02, threshold=3.702e+02, percent-clipped=0.0 +2023-04-27 12:25:23,558 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104469.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:25:58,403 INFO [finetune.py:976] (3/7) Epoch 19, batch 1400, loss[loss=0.2082, simple_loss=0.2946, pruned_loss=0.06086, over 4848.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2445, pruned_loss=0.05131, over 954652.80 frames. ], batch size: 49, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:26:30,550 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104532.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 12:26:41,377 INFO [finetune.py:976] (3/7) Epoch 19, batch 1450, loss[loss=0.2001, simple_loss=0.2707, pruned_loss=0.06477, over 4823.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2455, pruned_loss=0.05126, over 954735.68 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:26:46,127 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.618e+02 1.883e+02 2.290e+02 5.063e+02, threshold=3.766e+02, percent-clipped=2.0 +2023-04-27 12:27:03,310 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104580.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:27:14,737 INFO [finetune.py:976] (3/7) Epoch 19, batch 1500, loss[loss=0.2045, simple_loss=0.2784, pruned_loss=0.0653, over 4844.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2466, pruned_loss=0.05172, over 955944.16 frames. ], batch size: 44, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:27:39,379 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-04-27 12:27:46,879 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104646.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:27:48,602 INFO [finetune.py:976] (3/7) Epoch 19, batch 1550, loss[loss=0.1936, simple_loss=0.2604, pruned_loss=0.0634, over 4813.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2483, pruned_loss=0.0523, over 955918.71 frames. ], batch size: 40, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:27:51,151 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104653.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:27:53,360 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.681e+02 1.907e+02 2.272e+02 3.950e+02, threshold=3.814e+02, percent-clipped=3.0 +2023-04-27 12:28:00,899 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-27 12:28:33,194 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 12:28:35,128 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104694.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:28:43,386 INFO [finetune.py:976] (3/7) Epoch 19, batch 1600, loss[loss=0.2084, simple_loss=0.2698, pruned_loss=0.07353, over 4894.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2462, pruned_loss=0.0511, over 954291.10 frames. ], batch size: 32, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:28:44,681 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104701.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:28:53,299 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104705.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:29:13,051 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-04-27 12:29:16,414 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8050, 2.4272, 1.7003, 1.8719, 1.3761, 1.3883, 1.7601, 1.2362], + device='cuda:3'), covar=tensor([0.1793, 0.1263, 0.1567, 0.1630, 0.2401, 0.2189, 0.1035, 0.2082], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0213, 0.0168, 0.0205, 0.0200, 0.0185, 0.0157, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 12:29:27,150 INFO [finetune.py:976] (3/7) Epoch 19, batch 1650, loss[loss=0.124, simple_loss=0.2013, pruned_loss=0.02335, over 4830.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2436, pruned_loss=0.05042, over 955655.36 frames. ], batch size: 40, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:29:29,682 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104753.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:29:31,432 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.063e+01 1.479e+02 1.732e+02 2.090e+02 3.270e+02, threshold=3.465e+02, percent-clipped=0.0 +2023-04-27 12:30:01,072 INFO [finetune.py:976] (3/7) Epoch 19, batch 1700, loss[loss=0.1616, simple_loss=0.2414, pruned_loss=0.0409, over 4932.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2412, pruned_loss=0.04972, over 957220.44 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:30:05,251 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-04-27 12:30:48,748 INFO [finetune.py:976] (3/7) Epoch 19, batch 1750, loss[loss=0.1739, simple_loss=0.2374, pruned_loss=0.05519, over 4893.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2433, pruned_loss=0.05101, over 958418.41 frames. ], batch size: 32, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:30:53,010 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.623e+02 1.966e+02 2.378e+02 4.217e+02, threshold=3.932e+02, percent-clipped=2.0 +2023-04-27 12:31:30,727 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 12:31:37,541 INFO [finetune.py:976] (3/7) Epoch 19, batch 1800, loss[loss=0.1612, simple_loss=0.2213, pruned_loss=0.05059, over 4429.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2447, pruned_loss=0.0509, over 958446.63 frames. ], batch size: 19, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:31:42,619 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.49 vs. limit=5.0 +2023-04-27 12:31:52,864 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-27 12:31:56,751 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5151, 1.7132, 1.4363, 1.0802, 1.1512, 1.1128, 1.4049, 1.0785], + device='cuda:3'), covar=tensor([0.1728, 0.1284, 0.1522, 0.1752, 0.2415, 0.1998, 0.1073, 0.2101], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0215, 0.0170, 0.0208, 0.0203, 0.0187, 0.0158, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 12:32:04,480 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4043, 2.2492, 1.7717, 1.9145, 2.1102, 1.7917, 2.4331, 1.4036], + device='cuda:3'), covar=tensor([0.2922, 0.1349, 0.3745, 0.2271, 0.1540, 0.2126, 0.1474, 0.4340], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0351, 0.0432, 0.0356, 0.0385, 0.0383, 0.0376, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:32:06,736 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2270, 1.1699, 3.7870, 3.5532, 3.3439, 3.6353, 3.6093, 3.3751], + device='cuda:3'), covar=tensor([0.6943, 0.5559, 0.1186, 0.1723, 0.1174, 0.1539, 0.1701, 0.1496], + device='cuda:3'), in_proj_covar=tensor([0.0306, 0.0304, 0.0404, 0.0406, 0.0349, 0.0407, 0.0313, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:32:10,885 INFO [finetune.py:976] (3/7) Epoch 19, batch 1850, loss[loss=0.166, simple_loss=0.2387, pruned_loss=0.04668, over 4910.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2471, pruned_loss=0.05235, over 959207.32 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:32:15,623 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.634e+02 1.902e+02 2.257e+02 6.149e+02, threshold=3.804e+02, percent-clipped=1.0 +2023-04-27 12:32:44,648 INFO [finetune.py:976] (3/7) Epoch 19, batch 1900, loss[loss=0.1922, simple_loss=0.2509, pruned_loss=0.06681, over 4733.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2465, pruned_loss=0.05195, over 958049.17 frames. ], batch size: 23, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:33:18,502 INFO [finetune.py:976] (3/7) Epoch 19, batch 1950, loss[loss=0.1496, simple_loss=0.2254, pruned_loss=0.03687, over 4761.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2456, pruned_loss=0.05179, over 956495.08 frames. ], batch size: 28, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:33:22,765 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.554e+02 1.843e+02 2.166e+02 4.298e+02, threshold=3.686e+02, percent-clipped=1.0 +2023-04-27 12:33:33,310 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1209, 2.6263, 1.2902, 1.4808, 2.0872, 1.4171, 3.1580, 1.8462], + device='cuda:3'), covar=tensor([0.0594, 0.0652, 0.0763, 0.0990, 0.0424, 0.0803, 0.0189, 0.0530], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 12:33:56,298 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9076, 1.3086, 1.6498, 1.7057, 1.6234, 1.3908, 0.7284, 1.2334], + device='cuda:3'), covar=tensor([0.3393, 0.3689, 0.1900, 0.2347, 0.2780, 0.2627, 0.4292, 0.2246], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0246, 0.0226, 0.0314, 0.0218, 0.0231, 0.0227, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 12:34:05,150 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-04-27 12:34:18,470 INFO [finetune.py:976] (3/7) Epoch 19, batch 2000, loss[loss=0.1684, simple_loss=0.2386, pruned_loss=0.04915, over 4889.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2429, pruned_loss=0.05108, over 956182.36 frames. ], batch size: 35, lr: 3.29e-03, grad_scale: 64.0 +2023-04-27 12:35:14,515 INFO [finetune.py:976] (3/7) Epoch 19, batch 2050, loss[loss=0.1625, simple_loss=0.2276, pruned_loss=0.0487, over 4827.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2409, pruned_loss=0.05063, over 956676.22 frames. ], batch size: 30, lr: 3.29e-03, grad_scale: 64.0 +2023-04-27 12:35:18,783 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.655e+01 1.540e+02 1.837e+02 2.232e+02 4.472e+02, threshold=3.673e+02, percent-clipped=5.0 +2023-04-27 12:35:52,485 INFO [finetune.py:976] (3/7) Epoch 19, batch 2100, loss[loss=0.2578, simple_loss=0.3128, pruned_loss=0.1014, over 4354.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2425, pruned_loss=0.05184, over 955142.18 frames. ], batch size: 65, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:36:37,773 INFO [finetune.py:976] (3/7) Epoch 19, batch 2150, loss[loss=0.2528, simple_loss=0.3185, pruned_loss=0.0936, over 4911.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2473, pruned_loss=0.05345, over 955285.85 frames. ], batch size: 42, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:36:48,984 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.234e+01 1.652e+02 1.972e+02 2.377e+02 5.881e+02, threshold=3.945e+02, percent-clipped=2.0 +2023-04-27 12:37:36,679 INFO [finetune.py:976] (3/7) Epoch 19, batch 2200, loss[loss=0.1608, simple_loss=0.2335, pruned_loss=0.04404, over 4791.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2499, pruned_loss=0.05413, over 955604.11 frames. ], batch size: 26, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:38:41,671 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0628, 2.5584, 2.1006, 2.3869, 1.7133, 2.1461, 2.0938, 1.7516], + device='cuda:3'), covar=tensor([0.1871, 0.1253, 0.0779, 0.1108, 0.3110, 0.1111, 0.2150, 0.2688], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0302, 0.0216, 0.0279, 0.0310, 0.0258, 0.0250, 0.0264], + device='cuda:3'), out_proj_covar=tensor([1.1492e-04, 1.2011e-04, 8.5649e-05, 1.1091e-04, 1.2594e-04, 1.0228e-04, + 1.0102e-04, 1.0470e-04], device='cuda:3') +2023-04-27 12:38:49,116 INFO [finetune.py:976] (3/7) Epoch 19, batch 2250, loss[loss=0.1936, simple_loss=0.2687, pruned_loss=0.05927, over 4886.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2514, pruned_loss=0.05506, over 955931.76 frames. ], batch size: 43, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:38:59,482 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.635e+02 1.938e+02 2.374e+02 3.739e+02, threshold=3.876e+02, percent-clipped=0.0 +2023-04-27 12:38:59,821 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.17 vs. limit=5.0 +2023-04-27 12:39:32,283 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2008, 1.5872, 5.2859, 5.0320, 4.6461, 5.1856, 4.7044, 4.7311], + device='cuda:3'), covar=tensor([0.5911, 0.5398, 0.0863, 0.1274, 0.0820, 0.1100, 0.1071, 0.1201], + device='cuda:3'), in_proj_covar=tensor([0.0302, 0.0300, 0.0400, 0.0401, 0.0345, 0.0401, 0.0310, 0.0362], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:39:34,796 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105388.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:39:53,434 INFO [finetune.py:976] (3/7) Epoch 19, batch 2300, loss[loss=0.1216, simple_loss=0.1995, pruned_loss=0.02182, over 4762.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2505, pruned_loss=0.05388, over 955905.51 frames. ], batch size: 26, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:39:56,166 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3919, 1.8975, 2.3099, 2.9704, 2.2972, 1.8457, 1.8533, 2.1559], + device='cuda:3'), covar=tensor([0.3357, 0.3275, 0.1628, 0.2290, 0.2727, 0.2674, 0.3850, 0.1998], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0246, 0.0226, 0.0314, 0.0217, 0.0231, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 12:40:58,676 INFO [finetune.py:976] (3/7) Epoch 19, batch 2350, loss[loss=0.1598, simple_loss=0.2225, pruned_loss=0.04861, over 4756.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2478, pruned_loss=0.05259, over 954438.69 frames. ], batch size: 27, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:40:59,312 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105449.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 12:41:09,881 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.591e+02 1.880e+02 2.234e+02 4.098e+02, threshold=3.760e+02, percent-clipped=1.0 +2023-04-27 12:41:22,092 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5919, 1.6886, 1.4643, 1.0949, 1.2548, 1.1951, 1.4054, 1.1866], + device='cuda:3'), covar=tensor([0.1584, 0.1300, 0.1484, 0.1792, 0.2273, 0.1910, 0.1065, 0.1970], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0213, 0.0170, 0.0206, 0.0200, 0.0186, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 12:41:52,223 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105486.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:42:06,554 INFO [finetune.py:976] (3/7) Epoch 19, batch 2400, loss[loss=0.1868, simple_loss=0.252, pruned_loss=0.06076, over 4890.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2447, pruned_loss=0.05164, over 955860.46 frames. ], batch size: 32, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:42:09,008 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105502.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:42:39,430 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105547.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:42:40,534 INFO [finetune.py:976] (3/7) Epoch 19, batch 2450, loss[loss=0.1336, simple_loss=0.1995, pruned_loss=0.03391, over 4763.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2415, pruned_loss=0.05088, over 956147.57 frames. ], batch size: 26, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:42:45,853 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.599e+02 1.963e+02 2.373e+02 4.323e+02, threshold=3.926e+02, percent-clipped=3.0 +2023-04-27 12:42:50,143 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105563.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:43:07,252 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 12:43:14,478 INFO [finetune.py:976] (3/7) Epoch 19, batch 2500, loss[loss=0.1843, simple_loss=0.2616, pruned_loss=0.05347, over 4802.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2445, pruned_loss=0.05245, over 954062.08 frames. ], batch size: 29, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:43:37,959 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.55 vs. limit=5.0 +2023-04-27 12:43:47,976 INFO [finetune.py:976] (3/7) Epoch 19, batch 2550, loss[loss=0.1676, simple_loss=0.2467, pruned_loss=0.04426, over 4803.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2466, pruned_loss=0.0527, over 956005.03 frames. ], batch size: 40, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:43:53,309 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.536e+02 1.856e+02 2.246e+02 3.753e+02, threshold=3.711e+02, percent-clipped=0.0 +2023-04-27 12:44:26,606 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3539, 1.2818, 1.5642, 1.5534, 1.2772, 1.1589, 1.2651, 0.6989], + device='cuda:3'), covar=tensor([0.0527, 0.0575, 0.0413, 0.0623, 0.0801, 0.1135, 0.0596, 0.0683], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0067, 0.0066, 0.0066, 0.0074, 0.0095, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:44:27,227 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7254, 2.0761, 1.8139, 2.0117, 1.5608, 1.6717, 1.7348, 1.3282], + device='cuda:3'), covar=tensor([0.1672, 0.1137, 0.0797, 0.1077, 0.3318, 0.1154, 0.1752, 0.2444], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0304, 0.0217, 0.0280, 0.0312, 0.0259, 0.0251, 0.0266], + device='cuda:3'), out_proj_covar=tensor([1.1529e-04, 1.2071e-04, 8.5926e-05, 1.1109e-04, 1.2666e-04, 1.0263e-04, + 1.0132e-04, 1.0540e-04], device='cuda:3') +2023-04-27 12:44:27,713 INFO [finetune.py:976] (3/7) Epoch 19, batch 2600, loss[loss=0.1828, simple_loss=0.2591, pruned_loss=0.05321, over 4864.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.249, pruned_loss=0.05316, over 957663.74 frames. ], batch size: 31, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:44:58,149 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105744.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 12:45:01,138 INFO [finetune.py:976] (3/7) Epoch 19, batch 2650, loss[loss=0.1584, simple_loss=0.2337, pruned_loss=0.04158, over 4752.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2488, pruned_loss=0.05291, over 956521.92 frames. ], batch size: 54, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:45:06,397 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.206e+02 1.583e+02 1.907e+02 2.231e+02 5.599e+02, threshold=3.814e+02, percent-clipped=2.0 +2023-04-27 12:45:30,671 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105792.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:45:34,814 INFO [finetune.py:976] (3/7) Epoch 19, batch 2700, loss[loss=0.1418, simple_loss=0.2164, pruned_loss=0.03366, over 4824.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2473, pruned_loss=0.05209, over 956975.74 frames. ], batch size: 30, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:45:40,227 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105799.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:46:36,273 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105842.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:46:40,483 INFO [finetune.py:976] (3/7) Epoch 19, batch 2750, loss[loss=0.1774, simple_loss=0.2424, pruned_loss=0.05618, over 4744.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2453, pruned_loss=0.0514, over 956461.85 frames. ], batch size: 54, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:46:43,047 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105853.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:46:45,896 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.998e+01 1.495e+02 1.843e+02 2.455e+02 4.470e+02, threshold=3.686e+02, percent-clipped=1.0 +2023-04-27 12:46:46,585 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6656, 3.5775, 2.7031, 4.2344, 3.6533, 3.6694, 1.6338, 3.5591], + device='cuda:3'), covar=tensor([0.1699, 0.1274, 0.3205, 0.1911, 0.2297, 0.1910, 0.5835, 0.2649], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0214, 0.0246, 0.0306, 0.0298, 0.0246, 0.0270, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 12:46:46,592 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105858.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:46:47,897 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105860.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:47:35,995 INFO [finetune.py:976] (3/7) Epoch 19, batch 2800, loss[loss=0.1392, simple_loss=0.2056, pruned_loss=0.03644, over 4723.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2418, pruned_loss=0.05, over 957600.08 frames. ], batch size: 23, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:48:43,333 INFO [finetune.py:976] (3/7) Epoch 19, batch 2850, loss[loss=0.1867, simple_loss=0.2486, pruned_loss=0.06235, over 4820.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2405, pruned_loss=0.04983, over 955642.27 frames. ], batch size: 39, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:48:53,755 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.431e+02 1.804e+02 2.150e+02 4.441e+02, threshold=3.609e+02, percent-clipped=4.0 +2023-04-27 12:49:49,886 INFO [finetune.py:976] (3/7) Epoch 19, batch 2900, loss[loss=0.2053, simple_loss=0.2829, pruned_loss=0.0638, over 4805.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2439, pruned_loss=0.05099, over 954629.10 frames. ], batch size: 45, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:50:01,075 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8274, 1.4919, 1.4830, 1.6600, 2.0470, 1.6855, 1.3019, 1.3613], + device='cuda:3'), covar=tensor([0.1417, 0.1421, 0.2035, 0.1421, 0.0832, 0.1532, 0.2254, 0.2481], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0314, 0.0353, 0.0293, 0.0330, 0.0312, 0.0303, 0.0373], + device='cuda:3'), out_proj_covar=tensor([6.4030e-05, 6.5195e-05, 7.4905e-05, 5.9409e-05, 6.8458e-05, 6.5506e-05, + 6.3558e-05, 7.9402e-05], device='cuda:3') +2023-04-27 12:50:40,771 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106044.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:50:43,676 INFO [finetune.py:976] (3/7) Epoch 19, batch 2950, loss[loss=0.1604, simple_loss=0.2272, pruned_loss=0.04679, over 4736.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2467, pruned_loss=0.0516, over 954711.99 frames. ], batch size: 23, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:50:48,560 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.617e+02 1.806e+02 2.204e+02 5.278e+02, threshold=3.611e+02, percent-clipped=1.0 +2023-04-27 12:51:01,786 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106077.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:51:12,301 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106092.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:51:17,074 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 12:51:17,511 INFO [finetune.py:976] (3/7) Epoch 19, batch 3000, loss[loss=0.1853, simple_loss=0.2535, pruned_loss=0.0586, over 4927.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2493, pruned_loss=0.05314, over 953788.22 frames. ], batch size: 33, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:51:17,511 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 12:51:22,254 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1369, 2.5989, 1.0283, 1.3658, 1.9125, 1.3764, 3.0719, 1.7075], + device='cuda:3'), covar=tensor([0.0673, 0.0525, 0.0732, 0.1303, 0.0456, 0.0922, 0.0252, 0.0595], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0051, 0.0052, 0.0074, 0.0052], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 12:51:33,615 INFO [finetune.py:1010] (3/7) Epoch 19, validation: loss=0.1523, simple_loss=0.2226, pruned_loss=0.04099, over 2265189.00 frames. +2023-04-27 12:51:33,615 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 12:51:45,352 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-04-27 12:51:53,270 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8933, 2.6049, 2.0282, 1.9120, 1.4400, 1.4128, 2.0719, 1.3951], + device='cuda:3'), covar=tensor([0.1649, 0.1333, 0.1314, 0.1648, 0.2181, 0.1906, 0.0941, 0.2001], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0211, 0.0168, 0.0203, 0.0198, 0.0183, 0.0155, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 12:52:18,481 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106138.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:52:25,386 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4006, 3.8771, 0.8085, 1.8822, 1.8433, 2.6294, 2.2669, 0.8103], + device='cuda:3'), covar=tensor([0.1736, 0.1658, 0.2389, 0.1712, 0.1315, 0.1309, 0.1583, 0.2389], + device='cuda:3'), in_proj_covar=tensor([0.0114, 0.0238, 0.0135, 0.0118, 0.0129, 0.0150, 0.0114, 0.0117], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:52:26,610 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106142.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:52:36,646 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106148.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:52:37,150 INFO [finetune.py:976] (3/7) Epoch 19, batch 3050, loss[loss=0.2104, simple_loss=0.2846, pruned_loss=0.0681, over 4908.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2512, pruned_loss=0.05343, over 953848.28 frames. ], batch size: 37, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:52:39,146 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8238, 1.4460, 1.9789, 2.3523, 1.9003, 1.8250, 1.9220, 1.8608], + device='cuda:3'), covar=tensor([0.4390, 0.6043, 0.5814, 0.5186, 0.5269, 0.6916, 0.7337, 0.7974], + device='cuda:3'), in_proj_covar=tensor([0.0425, 0.0407, 0.0501, 0.0502, 0.0453, 0.0480, 0.0486, 0.0491], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 12:52:46,801 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106155.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:52:48,481 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.467e+02 1.745e+02 2.097e+02 4.231e+02, threshold=3.490e+02, percent-clipped=3.0 +2023-04-27 12:52:49,216 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:53:31,157 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106190.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:53:42,131 INFO [finetune.py:976] (3/7) Epoch 19, batch 3100, loss[loss=0.1551, simple_loss=0.2241, pruned_loss=0.04306, over 4860.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2491, pruned_loss=0.05291, over 954100.80 frames. ], batch size: 34, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:53:51,784 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:54:35,396 INFO [finetune.py:976] (3/7) Epoch 19, batch 3150, loss[loss=0.184, simple_loss=0.245, pruned_loss=0.0615, over 4899.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2456, pruned_loss=0.05202, over 956338.14 frames. ], batch size: 35, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:54:40,720 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.609e+02 1.934e+02 2.307e+02 5.875e+02, threshold=3.868e+02, percent-clipped=2.0 +2023-04-27 12:54:49,451 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106268.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:55:23,648 INFO [finetune.py:976] (3/7) Epoch 19, batch 3200, loss[loss=0.1662, simple_loss=0.2315, pruned_loss=0.0504, over 4868.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2422, pruned_loss=0.05108, over 956979.34 frames. ], batch size: 31, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:55:28,475 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3790, 3.5001, 0.8118, 1.7641, 1.7678, 2.3825, 1.9786, 1.0613], + device='cuda:3'), covar=tensor([0.1623, 0.1300, 0.2290, 0.1533, 0.1211, 0.1180, 0.1710, 0.1988], + device='cuda:3'), in_proj_covar=tensor([0.0115, 0.0240, 0.0136, 0.0119, 0.0130, 0.0151, 0.0115, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 12:55:50,400 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:55:54,732 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:55:59,326 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 12:56:03,600 INFO [finetune.py:976] (3/7) Epoch 19, batch 3250, loss[loss=0.1653, simple_loss=0.2477, pruned_loss=0.0414, over 4738.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2418, pruned_loss=0.05047, over 956547.90 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:56:08,525 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.638e+02 1.854e+02 2.216e+02 4.860e+02, threshold=3.708e+02, percent-clipped=2.0 +2023-04-27 12:56:35,798 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106397.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:56:37,426 INFO [finetune.py:976] (3/7) Epoch 19, batch 3300, loss[loss=0.2029, simple_loss=0.3017, pruned_loss=0.05206, over 4801.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2454, pruned_loss=0.05142, over 955670.84 frames. ], batch size: 45, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:56:44,915 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7305, 2.0856, 0.9848, 1.3997, 1.9274, 1.6265, 1.4937, 1.6058], + device='cuda:3'), covar=tensor([0.0509, 0.0338, 0.0345, 0.0551, 0.0276, 0.0492, 0.0490, 0.0556], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 12:56:58,595 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0867, 2.5727, 1.0773, 1.3802, 2.2447, 1.3537, 3.5445, 1.7981], + device='cuda:3'), covar=tensor([0.0726, 0.0598, 0.0787, 0.1437, 0.0487, 0.1056, 0.0284, 0.0695], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 12:57:00,451 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106433.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:09,678 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106448.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:10,217 INFO [finetune.py:976] (3/7) Epoch 19, batch 3350, loss[loss=0.1609, simple_loss=0.2263, pruned_loss=0.0477, over 4716.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2462, pruned_loss=0.05144, over 953643.48 frames. ], batch size: 23, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:57:14,362 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106454.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:14,945 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106455.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:16,069 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.680e+02 1.880e+02 2.311e+02 4.481e+02, threshold=3.759e+02, percent-clipped=2.0 +2023-04-27 12:57:42,141 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106496.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:43,905 INFO [finetune.py:976] (3/7) Epoch 19, batch 3400, loss[loss=0.2226, simple_loss=0.2957, pruned_loss=0.07474, over 4912.00 frames. ], tot_loss[loss=0.176, simple_loss=0.248, pruned_loss=0.05195, over 952314.18 frames. ], batch size: 38, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:57:46,913 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106503.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:55,251 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106515.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:58:17,777 INFO [finetune.py:976] (3/7) Epoch 19, batch 3450, loss[loss=0.1867, simple_loss=0.2499, pruned_loss=0.06172, over 4816.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2479, pruned_loss=0.05192, over 952681.89 frames. ], batch size: 25, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:58:20,348 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7615, 2.4392, 1.7538, 1.8691, 1.2965, 1.3068, 1.7389, 1.2323], + device='cuda:3'), covar=tensor([0.1905, 0.1348, 0.1583, 0.1698, 0.2504, 0.2222, 0.1113, 0.2153], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0168, 0.0204, 0.0199, 0.0185, 0.0156, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 12:58:23,138 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.592e+02 1.859e+02 2.190e+02 3.296e+02, threshold=3.717e+02, percent-clipped=0.0 +2023-04-27 12:58:30,243 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7991, 3.7269, 2.7965, 4.4266, 3.7295, 3.7927, 1.5604, 3.7922], + device='cuda:3'), covar=tensor([0.1838, 0.1513, 0.3403, 0.1597, 0.3229, 0.1857, 0.5842, 0.2456], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0216, 0.0248, 0.0308, 0.0298, 0.0247, 0.0271, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 12:58:47,503 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4835, 3.0481, 1.0045, 1.6305, 2.4020, 1.6391, 4.1260, 2.1782], + device='cuda:3'), covar=tensor([0.0599, 0.0817, 0.0855, 0.1223, 0.0471, 0.0898, 0.0184, 0.0600], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 12:58:57,665 INFO [finetune.py:976] (3/7) Epoch 19, batch 3500, loss[loss=0.1576, simple_loss=0.2281, pruned_loss=0.04362, over 4797.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2463, pruned_loss=0.05215, over 954678.80 frames. ], batch size: 29, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:59:14,545 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106624.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:59:36,355 INFO [finetune.py:976] (3/7) Epoch 19, batch 3550, loss[loss=0.1685, simple_loss=0.2453, pruned_loss=0.04583, over 4809.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.244, pruned_loss=0.05178, over 954670.54 frames. ], batch size: 45, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:59:41,156 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.386e+01 1.482e+02 1.819e+02 2.170e+02 5.506e+02, threshold=3.638e+02, percent-clipped=1.0 +2023-04-27 12:59:46,102 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-27 13:00:05,460 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:00:14,738 INFO [finetune.py:976] (3/7) Epoch 19, batch 3600, loss[loss=0.2502, simple_loss=0.2916, pruned_loss=0.1044, over 4127.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2417, pruned_loss=0.05074, over 954706.92 frames. ], batch size: 65, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:00:46,380 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9395, 2.3836, 0.7240, 1.2356, 1.2838, 1.7483, 1.5106, 0.8355], + device='cuda:3'), covar=tensor([0.2141, 0.2126, 0.2445, 0.2116, 0.1589, 0.1448, 0.1961, 0.2034], + device='cuda:3'), in_proj_covar=tensor([0.0115, 0.0240, 0.0136, 0.0118, 0.0131, 0.0150, 0.0115, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:00:55,324 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-27 13:00:58,221 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106733.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:01:08,805 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 13:01:14,589 INFO [finetune.py:976] (3/7) Epoch 19, batch 3650, loss[loss=0.1949, simple_loss=0.2755, pruned_loss=0.05716, over 4865.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2446, pruned_loss=0.05186, over 954306.58 frames. ], batch size: 44, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:01:16,511 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2675, 2.9892, 0.8152, 1.6991, 1.6380, 2.2096, 1.7927, 0.9015], + device='cuda:3'), covar=tensor([0.1409, 0.0931, 0.1937, 0.1237, 0.1128, 0.0912, 0.1501, 0.1971], + device='cuda:3'), in_proj_covar=tensor([0.0115, 0.0240, 0.0136, 0.0119, 0.0131, 0.0150, 0.0115, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:01:19,423 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.799e+01 1.618e+02 2.033e+02 2.562e+02 4.341e+02, threshold=4.066e+02, percent-clipped=3.0 +2023-04-27 13:01:26,469 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-27 13:01:34,639 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5946, 2.0886, 0.9522, 1.3636, 2.0577, 1.5714, 1.4829, 1.5524], + device='cuda:3'), covar=tensor([0.0535, 0.0340, 0.0341, 0.0595, 0.0270, 0.0512, 0.0526, 0.0588], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 13:01:35,803 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106781.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:01:47,625 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5415, 2.3461, 2.5714, 3.1042, 2.9899, 2.6069, 2.0061, 2.7750], + device='cuda:3'), covar=tensor([0.0888, 0.0919, 0.0597, 0.0517, 0.0531, 0.0778, 0.0718, 0.0520], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0200, 0.0181, 0.0171, 0.0174, 0.0181, 0.0150, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:01:48,135 INFO [finetune.py:976] (3/7) Epoch 19, batch 3700, loss[loss=0.155, simple_loss=0.2282, pruned_loss=0.04089, over 4936.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2469, pruned_loss=0.0518, over 952565.61 frames. ], batch size: 33, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:01:55,079 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106810.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:01:56,207 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-27 13:02:22,252 INFO [finetune.py:976] (3/7) Epoch 19, batch 3750, loss[loss=0.1779, simple_loss=0.2581, pruned_loss=0.04881, over 4812.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2479, pruned_loss=0.05251, over 952466.08 frames. ], batch size: 39, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:02:23,318 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 13:02:27,098 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.601e+01 1.567e+02 1.830e+02 2.266e+02 4.856e+02, threshold=3.659e+02, percent-clipped=1.0 +2023-04-27 13:02:28,113 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-27 13:02:52,585 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8079, 1.7538, 1.6155, 1.4084, 1.8109, 1.5681, 2.3064, 1.3465], + device='cuda:3'), covar=tensor([0.3279, 0.1534, 0.4900, 0.2624, 0.1441, 0.2099, 0.1310, 0.4641], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0346, 0.0429, 0.0355, 0.0383, 0.0378, 0.0373, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:02:55,380 INFO [finetune.py:976] (3/7) Epoch 19, batch 3800, loss[loss=0.1893, simple_loss=0.2571, pruned_loss=0.06077, over 4812.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.248, pruned_loss=0.05265, over 950872.45 frames. ], batch size: 25, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:03:10,669 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106924.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:03:21,866 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1675, 1.7412, 2.0218, 2.5683, 2.0745, 1.6650, 1.5803, 1.8835], + device='cuda:3'), covar=tensor([0.2775, 0.2913, 0.1494, 0.1802, 0.2335, 0.2402, 0.4083, 0.2074], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0245, 0.0226, 0.0315, 0.0218, 0.0231, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 13:03:27,656 INFO [finetune.py:976] (3/7) Epoch 19, batch 3850, loss[loss=0.1525, simple_loss=0.2212, pruned_loss=0.0419, over 4895.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2467, pruned_loss=0.05224, over 951341.50 frames. ], batch size: 35, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:03:33,087 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.551e+02 1.789e+02 2.154e+02 4.176e+02, threshold=3.578e+02, percent-clipped=2.0 +2023-04-27 13:03:42,311 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:03:56,029 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106992.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:04:00,619 INFO [finetune.py:976] (3/7) Epoch 19, batch 3900, loss[loss=0.2078, simple_loss=0.2765, pruned_loss=0.06954, over 4925.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2441, pruned_loss=0.05175, over 950621.82 frames. ], batch size: 38, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:04:33,284 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:04:35,020 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2779, 1.4535, 1.3488, 1.8333, 1.7881, 1.8960, 1.4699, 3.5562], + device='cuda:3'), covar=tensor([0.0691, 0.0959, 0.1024, 0.1265, 0.0714, 0.0505, 0.0854, 0.0180], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 13:04:41,522 INFO [finetune.py:976] (3/7) Epoch 19, batch 3950, loss[loss=0.1598, simple_loss=0.2221, pruned_loss=0.0488, over 4804.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2425, pruned_loss=0.05141, over 950199.66 frames. ], batch size: 26, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:04:54,081 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.559e+01 1.455e+02 1.767e+02 2.125e+02 3.704e+02, threshold=3.534e+02, percent-clipped=1.0 +2023-04-27 13:05:24,906 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.58 vs. limit=5.0 +2023-04-27 13:05:26,564 INFO [finetune.py:976] (3/7) Epoch 19, batch 4000, loss[loss=0.1607, simple_loss=0.2371, pruned_loss=0.04214, over 4895.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2426, pruned_loss=0.05173, over 949283.06 frames. ], batch size: 32, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:05:35,386 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:06:15,750 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107140.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:06:28,918 INFO [finetune.py:976] (3/7) Epoch 19, batch 4050, loss[loss=0.1417, simple_loss=0.2035, pruned_loss=0.03994, over 4741.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2455, pruned_loss=0.05282, over 949545.90 frames. ], batch size: 23, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:06:30,811 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107152.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:06:35,807 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.479e+01 1.702e+02 1.942e+02 2.364e+02 4.795e+02, threshold=3.885e+02, percent-clipped=4.0 +2023-04-27 13:06:36,474 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:07:01,384 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1230, 1.5547, 1.4421, 2.0181, 2.1872, 1.8294, 1.8071, 1.4699], + device='cuda:3'), covar=tensor([0.1736, 0.1591, 0.1870, 0.1695, 0.1083, 0.1819, 0.1877, 0.2198], + device='cuda:3'), in_proj_covar=tensor([0.0304, 0.0309, 0.0346, 0.0287, 0.0325, 0.0305, 0.0296, 0.0367], + device='cuda:3'), out_proj_covar=tensor([6.2690e-05, 6.4060e-05, 7.3228e-05, 5.8090e-05, 6.7332e-05, 6.4057e-05, + 6.2005e-05, 7.8104e-05], device='cuda:3') +2023-04-27 13:07:15,574 INFO [finetune.py:976] (3/7) Epoch 19, batch 4100, loss[loss=0.1482, simple_loss=0.2244, pruned_loss=0.036, over 4859.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2498, pruned_loss=0.05378, over 951107.93 frames. ], batch size: 44, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:07:17,375 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107201.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:07:26,164 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107213.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:07:27,838 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3722, 2.1551, 2.5098, 2.7775, 2.8206, 2.2938, 1.7787, 2.5702], + device='cuda:3'), covar=tensor([0.0938, 0.1048, 0.0621, 0.0623, 0.0645, 0.0922, 0.0836, 0.0546], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0201, 0.0183, 0.0173, 0.0176, 0.0182, 0.0151, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:07:48,305 INFO [finetune.py:976] (3/7) Epoch 19, batch 4150, loss[loss=0.2197, simple_loss=0.2778, pruned_loss=0.08079, over 4846.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2511, pruned_loss=0.05435, over 950440.09 frames. ], batch size: 31, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:07:51,212 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0430, 2.6736, 1.0911, 1.5136, 2.0491, 1.2740, 3.4722, 1.9436], + device='cuda:3'), covar=tensor([0.0680, 0.0612, 0.0715, 0.1123, 0.0468, 0.0983, 0.0197, 0.0580], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 13:07:54,579 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.951e+01 1.595e+02 1.845e+02 2.201e+02 3.820e+02, threshold=3.690e+02, percent-clipped=0.0 +2023-04-27 13:08:10,212 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-27 13:08:11,753 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1744, 1.6335, 2.0037, 2.3877, 1.9730, 1.5993, 1.3106, 1.8534], + device='cuda:3'), covar=tensor([0.2813, 0.2894, 0.1431, 0.1938, 0.2427, 0.2479, 0.4097, 0.1919], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0246, 0.0227, 0.0316, 0.0219, 0.0231, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 13:08:15,943 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6153, 1.8636, 0.7801, 1.3625, 1.7081, 1.5160, 1.4119, 1.5299], + device='cuda:3'), covar=tensor([0.0491, 0.0344, 0.0341, 0.0538, 0.0267, 0.0502, 0.0502, 0.0563], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0049, 0.0051], + device='cuda:3') +2023-04-27 13:08:21,561 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7431, 1.5837, 1.9972, 2.1349, 1.5813, 1.3942, 1.7127, 1.1082], + device='cuda:3'), covar=tensor([0.0431, 0.0631, 0.0384, 0.0488, 0.0630, 0.1058, 0.0598, 0.0636], + device='cuda:3'), in_proj_covar=tensor([0.0067, 0.0067, 0.0066, 0.0066, 0.0074, 0.0095, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:08:22,020 INFO [finetune.py:976] (3/7) Epoch 19, batch 4200, loss[loss=0.1673, simple_loss=0.2397, pruned_loss=0.04743, over 4763.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2494, pruned_loss=0.05292, over 948939.04 frames. ], batch size: 28, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:08:55,957 INFO [finetune.py:976] (3/7) Epoch 19, batch 4250, loss[loss=0.1707, simple_loss=0.2454, pruned_loss=0.04798, over 4738.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2472, pruned_loss=0.05217, over 950365.54 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:09:01,364 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.574e+01 1.612e+02 1.786e+02 2.196e+02 3.792e+02, threshold=3.571e+02, percent-clipped=1.0 +2023-04-27 13:09:10,348 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=2.03 vs. limit=2.0 +2023-04-27 13:09:29,660 INFO [finetune.py:976] (3/7) Epoch 19, batch 4300, loss[loss=0.1148, simple_loss=0.1938, pruned_loss=0.0179, over 4765.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2447, pruned_loss=0.05165, over 949319.01 frames. ], batch size: 28, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:10:02,816 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6468, 1.3116, 4.5318, 4.2725, 3.9701, 4.3665, 4.2473, 3.9255], + device='cuda:3'), covar=tensor([0.6408, 0.5757, 0.0814, 0.1305, 0.0901, 0.1563, 0.1053, 0.1300], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0303, 0.0406, 0.0404, 0.0348, 0.0405, 0.0312, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:10:04,025 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5485, 1.2348, 4.4293, 4.1369, 3.8928, 4.2358, 4.1546, 3.8620], + device='cuda:3'), covar=tensor([0.6792, 0.6058, 0.0934, 0.1649, 0.0975, 0.1386, 0.1210, 0.1566], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0303, 0.0406, 0.0404, 0.0348, 0.0405, 0.0312, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:10:04,050 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2811, 1.5912, 1.5387, 1.7799, 1.8191, 1.9700, 1.4388, 3.6599], + device='cuda:3'), covar=tensor([0.0603, 0.0840, 0.0751, 0.1213, 0.0608, 0.0504, 0.0757, 0.0111], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 13:10:05,747 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1017, 2.5872, 2.1619, 2.4688, 1.8418, 2.1450, 2.1940, 1.6091], + device='cuda:3'), covar=tensor([0.1620, 0.0939, 0.0624, 0.1030, 0.2655, 0.0928, 0.1624, 0.2538], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0299, 0.0213, 0.0274, 0.0307, 0.0256, 0.0247, 0.0261], + device='cuda:3'), out_proj_covar=tensor([1.1346e-04, 1.1860e-04, 8.4570e-05, 1.0883e-04, 1.2472e-04, 1.0124e-04, + 9.9821e-05, 1.0347e-04], device='cuda:3') +2023-04-27 13:10:31,546 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-04-27 13:10:31,845 INFO [finetune.py:976] (3/7) Epoch 19, batch 4350, loss[loss=0.146, simple_loss=0.2257, pruned_loss=0.03319, over 4689.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2417, pruned_loss=0.05049, over 950534.20 frames. ], batch size: 23, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:10:37,327 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.596e+02 1.862e+02 2.295e+02 4.192e+02, threshold=3.723e+02, percent-clipped=1.0 +2023-04-27 13:11:03,014 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107496.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 13:11:04,782 INFO [finetune.py:976] (3/7) Epoch 19, batch 4400, loss[loss=0.181, simple_loss=0.2347, pruned_loss=0.0636, over 4778.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2422, pruned_loss=0.05104, over 951915.09 frames. ], batch size: 25, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:11:10,414 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107508.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:11:11,653 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107510.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:11:49,530 INFO [finetune.py:976] (3/7) Epoch 19, batch 4450, loss[loss=0.1554, simple_loss=0.2188, pruned_loss=0.04596, over 4750.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2451, pruned_loss=0.05191, over 952254.98 frames. ], batch size: 23, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:11:59,077 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-04-27 13:12:00,086 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.632e+02 1.928e+02 2.326e+02 3.887e+02, threshold=3.856e+02, percent-clipped=1.0 +2023-04-27 13:12:12,157 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-27 13:12:19,126 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107571.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:12:43,864 INFO [finetune.py:976] (3/7) Epoch 19, batch 4500, loss[loss=0.1788, simple_loss=0.2551, pruned_loss=0.05127, over 4926.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2462, pruned_loss=0.05203, over 954638.16 frames. ], batch size: 33, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:13:17,593 INFO [finetune.py:976] (3/7) Epoch 19, batch 4550, loss[loss=0.199, simple_loss=0.2675, pruned_loss=0.06525, over 4910.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.247, pruned_loss=0.0519, over 955765.55 frames. ], batch size: 37, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:13:20,740 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0143, 1.7650, 2.0029, 2.3057, 2.3541, 1.9555, 1.4979, 2.2298], + device='cuda:3'), covar=tensor([0.0871, 0.1188, 0.0748, 0.0657, 0.0595, 0.0922, 0.0805, 0.0513], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0203, 0.0183, 0.0173, 0.0177, 0.0182, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:13:23,030 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.628e+02 1.814e+02 2.190e+02 5.311e+02, threshold=3.629e+02, percent-clipped=2.0 +2023-04-27 13:13:23,726 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:13:42,356 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0404, 4.3336, 1.0887, 2.2997, 2.5072, 2.7121, 2.4625, 1.0078], + device='cuda:3'), covar=tensor([0.1303, 0.0967, 0.1999, 0.1190, 0.0938, 0.1113, 0.1390, 0.2113], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0242, 0.0138, 0.0120, 0.0133, 0.0152, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:13:46,979 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6111, 1.8477, 0.7194, 1.3429, 1.8750, 1.4949, 1.3844, 1.5027], + device='cuda:3'), covar=tensor([0.0487, 0.0338, 0.0343, 0.0532, 0.0262, 0.0476, 0.0492, 0.0539], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:3') +2023-04-27 13:13:51,122 INFO [finetune.py:976] (3/7) Epoch 19, batch 4600, loss[loss=0.1792, simple_loss=0.2422, pruned_loss=0.05815, over 4818.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2466, pruned_loss=0.05176, over 956699.66 frames. ], batch size: 33, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:13:51,469 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 13:14:04,029 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:14:16,218 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107737.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:14:24,297 INFO [finetune.py:976] (3/7) Epoch 19, batch 4650, loss[loss=0.1595, simple_loss=0.2327, pruned_loss=0.04308, over 4820.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2448, pruned_loss=0.05149, over 958355.54 frames. ], batch size: 40, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:14:29,768 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.588e+02 1.876e+02 2.187e+02 4.173e+02, threshold=3.752e+02, percent-clipped=2.0 +2023-04-27 13:15:06,651 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107796.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:15:07,913 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107798.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 13:15:08,401 INFO [finetune.py:976] (3/7) Epoch 19, batch 4700, loss[loss=0.1803, simple_loss=0.257, pruned_loss=0.05181, over 4818.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2427, pruned_loss=0.05138, over 958872.78 frames. ], batch size: 39, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:15:19,599 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107808.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:15:48,344 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107844.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:15:52,874 INFO [finetune.py:976] (3/7) Epoch 19, batch 4750, loss[loss=0.1641, simple_loss=0.2417, pruned_loss=0.04327, over 4917.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2413, pruned_loss=0.05094, over 956470.01 frames. ], batch size: 37, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:15:57,097 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107856.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:15:58,278 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.851e+01 1.548e+02 1.806e+02 2.304e+02 4.398e+02, threshold=3.612e+02, percent-clipped=1.0 +2023-04-27 13:16:09,257 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107866.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:16:53,223 INFO [finetune.py:976] (3/7) Epoch 19, batch 4800, loss[loss=0.1867, simple_loss=0.2598, pruned_loss=0.05685, over 4936.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2446, pruned_loss=0.05255, over 956144.10 frames. ], batch size: 38, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:17:31,834 INFO [finetune.py:976] (3/7) Epoch 19, batch 4850, loss[loss=0.1978, simple_loss=0.259, pruned_loss=0.06834, over 4323.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2467, pruned_loss=0.05245, over 954894.43 frames. ], batch size: 66, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:17:37,756 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.607e+02 1.917e+02 2.218e+02 6.571e+02, threshold=3.834e+02, percent-clipped=2.0 +2023-04-27 13:17:39,040 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.5940, 4.6335, 2.9651, 5.2200, 4.6268, 4.5268, 1.6068, 4.5574], + device='cuda:3'), covar=tensor([0.1531, 0.0805, 0.3416, 0.0840, 0.2453, 0.1386, 0.5895, 0.2032], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0217, 0.0251, 0.0308, 0.0300, 0.0250, 0.0275, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:17:50,411 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-27 13:18:04,088 INFO [finetune.py:976] (3/7) Epoch 19, batch 4900, loss[loss=0.2175, simple_loss=0.2877, pruned_loss=0.07366, over 4811.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2484, pruned_loss=0.05321, over 957836.78 frames. ], batch size: 38, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:18:15,607 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 13:18:16,704 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108015.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:18:17,472 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-27 13:18:38,322 INFO [finetune.py:976] (3/7) Epoch 19, batch 4950, loss[loss=0.182, simple_loss=0.2441, pruned_loss=0.05991, over 4804.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2482, pruned_loss=0.05246, over 957824.17 frames. ], batch size: 25, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:18:45,331 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.581e+02 1.882e+02 2.406e+02 4.521e+02, threshold=3.765e+02, percent-clipped=4.0 +2023-04-27 13:19:07,760 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108093.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:19:11,337 INFO [finetune.py:976] (3/7) Epoch 19, batch 5000, loss[loss=0.2033, simple_loss=0.2676, pruned_loss=0.06949, over 4900.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2461, pruned_loss=0.0519, over 956276.29 frames. ], batch size: 37, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:19:22,392 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8090, 1.4156, 1.9243, 2.2876, 1.8926, 1.8017, 1.8306, 1.8281], + device='cuda:3'), covar=tensor([0.4646, 0.6798, 0.6741, 0.5381, 0.5951, 0.7778, 0.7884, 0.9438], + device='cuda:3'), in_proj_covar=tensor([0.0427, 0.0410, 0.0503, 0.0505, 0.0455, 0.0483, 0.0488, 0.0496], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:19:39,444 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0035, 1.6412, 2.1642, 2.4774, 2.1131, 1.9225, 2.0007, 1.9905], + device='cuda:3'), covar=tensor([0.4723, 0.7191, 0.7320, 0.5883, 0.5994, 0.8778, 0.9133, 1.0152], + device='cuda:3'), in_proj_covar=tensor([0.0429, 0.0411, 0.0504, 0.0506, 0.0456, 0.0484, 0.0489, 0.0497], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:19:44,757 INFO [finetune.py:976] (3/7) Epoch 19, batch 5050, loss[loss=0.2045, simple_loss=0.2637, pruned_loss=0.07268, over 4908.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2437, pruned_loss=0.05121, over 957663.11 frames. ], batch size: 43, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:19:51,169 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.142e+01 1.502e+02 1.806e+02 2.114e+02 4.959e+02, threshold=3.611e+02, percent-clipped=2.0 +2023-04-27 13:19:57,094 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:20:29,640 INFO [finetune.py:976] (3/7) Epoch 19, batch 5100, loss[loss=0.1813, simple_loss=0.2422, pruned_loss=0.0602, over 4874.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2389, pruned_loss=0.04927, over 956776.35 frames. ], batch size: 31, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:20:50,603 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108214.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:21:02,996 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.53 vs. limit=5.0 +2023-04-27 13:21:34,996 INFO [finetune.py:976] (3/7) Epoch 19, batch 5150, loss[loss=0.1251, simple_loss=0.1922, pruned_loss=0.02905, over 3959.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2397, pruned_loss=0.0499, over 955637.12 frames. ], batch size: 17, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:21:44,949 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.855e+01 1.529e+02 1.881e+02 2.347e+02 3.720e+02, threshold=3.762e+02, percent-clipped=1.0 +2023-04-27 13:22:29,162 INFO [finetune.py:976] (3/7) Epoch 19, batch 5200, loss[loss=0.2284, simple_loss=0.2974, pruned_loss=0.07976, over 4894.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2446, pruned_loss=0.05136, over 957028.74 frames. ], batch size: 35, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:22:41,862 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5180, 3.6123, 2.5523, 4.1201, 3.6028, 3.5354, 1.6034, 3.5138], + device='cuda:3'), covar=tensor([0.1875, 0.1217, 0.3408, 0.1826, 0.3881, 0.1716, 0.5563, 0.2492], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0214, 0.0248, 0.0304, 0.0296, 0.0247, 0.0271, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:22:51,812 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108315.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:23:04,353 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-27 13:23:05,671 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108325.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:23:36,574 INFO [finetune.py:976] (3/7) Epoch 19, batch 5250, loss[loss=0.1818, simple_loss=0.2482, pruned_loss=0.05772, over 4064.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2471, pruned_loss=0.05205, over 955880.21 frames. ], batch size: 65, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:23:44,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.684e+02 2.011e+02 2.395e+02 3.683e+02, threshold=4.022e+02, percent-clipped=0.0 +2023-04-27 13:23:53,522 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108363.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:19,500 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108386.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:23,685 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108393.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:27,187 INFO [finetune.py:976] (3/7) Epoch 19, batch 5300, loss[loss=0.1861, simple_loss=0.2575, pruned_loss=0.05736, over 4892.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2494, pruned_loss=0.05297, over 957247.65 frames. ], batch size: 36, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:24:29,894 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9305, 1.8448, 1.7729, 1.5311, 1.9915, 1.5482, 2.5004, 1.5412], + device='cuda:3'), covar=tensor([0.3832, 0.1915, 0.4741, 0.2965, 0.1863, 0.2454, 0.1420, 0.4281], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0345, 0.0426, 0.0353, 0.0382, 0.0377, 0.0372, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:24:50,800 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:50,832 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9592, 2.2717, 2.0564, 2.2627, 1.9948, 2.1122, 2.1678, 2.1284], + device='cuda:3'), covar=tensor([0.3908, 0.6514, 0.5123, 0.4525, 0.6159, 0.7716, 0.6416, 0.6003], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0372, 0.0321, 0.0334, 0.0345, 0.0394, 0.0357, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:24:56,168 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108441.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:25:01,039 INFO [finetune.py:976] (3/7) Epoch 19, batch 5350, loss[loss=0.1369, simple_loss=0.2076, pruned_loss=0.0331, over 4717.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2485, pruned_loss=0.05177, over 956005.70 frames. ], batch size: 23, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:25:06,503 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.659e+02 1.988e+02 2.329e+02 5.055e+02, threshold=3.977e+02, percent-clipped=1.0 +2023-04-27 13:25:31,251 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108493.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:25:34,727 INFO [finetune.py:976] (3/7) Epoch 19, batch 5400, loss[loss=0.1393, simple_loss=0.2075, pruned_loss=0.03557, over 4867.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2453, pruned_loss=0.05067, over 956678.53 frames. ], batch size: 31, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:26:21,264 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6329, 2.1209, 2.4958, 2.9589, 2.4207, 1.9587, 1.8208, 2.3776], + device='cuda:3'), covar=tensor([0.3292, 0.3185, 0.1743, 0.2622, 0.3013, 0.2796, 0.4000, 0.2164], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0246, 0.0227, 0.0317, 0.0219, 0.0233, 0.0228, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 13:26:31,590 INFO [finetune.py:976] (3/7) Epoch 19, batch 5450, loss[loss=0.1596, simple_loss=0.2298, pruned_loss=0.04471, over 4866.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2424, pruned_loss=0.0503, over 954922.95 frames. ], batch size: 31, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:26:42,377 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.852e+01 1.533e+02 1.903e+02 2.214e+02 4.592e+02, threshold=3.806e+02, percent-clipped=2.0 +2023-04-27 13:26:42,849 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-27 13:27:10,224 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:27:25,457 INFO [finetune.py:976] (3/7) Epoch 19, batch 5500, loss[loss=0.1255, simple_loss=0.1984, pruned_loss=0.02631, over 4783.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.239, pruned_loss=0.04933, over 954513.73 frames. ], batch size: 26, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:27:35,793 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-04-27 13:27:51,046 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108638.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:27:58,828 INFO [finetune.py:976] (3/7) Epoch 19, batch 5550, loss[loss=0.1956, simple_loss=0.2616, pruned_loss=0.06476, over 4832.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2411, pruned_loss=0.04996, over 951386.95 frames. ], batch size: 30, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:28:10,001 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.131e+01 1.592e+02 1.937e+02 2.305e+02 3.019e+02, threshold=3.873e+02, percent-clipped=0.0 +2023-04-27 13:28:41,417 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108681.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:29:03,284 INFO [finetune.py:976] (3/7) Epoch 19, batch 5600, loss[loss=0.1596, simple_loss=0.2468, pruned_loss=0.03623, over 4805.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2447, pruned_loss=0.05079, over 953046.32 frames. ], batch size: 51, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:29:12,626 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0296, 1.6574, 1.9558, 2.3991, 2.3952, 1.9712, 1.6746, 2.1024], + device='cuda:3'), covar=tensor([0.0821, 0.1214, 0.0737, 0.0574, 0.0630, 0.0895, 0.0764, 0.0594], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0203, 0.0184, 0.0173, 0.0180, 0.0183, 0.0153, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:29:27,117 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 13:30:00,556 INFO [finetune.py:976] (3/7) Epoch 19, batch 5650, loss[loss=0.2004, simple_loss=0.2729, pruned_loss=0.06394, over 4758.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2461, pruned_loss=0.05066, over 953419.00 frames. ], batch size: 27, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:30:17,474 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.527e+02 1.868e+02 2.111e+02 3.831e+02, threshold=3.735e+02, percent-clipped=0.0 +2023-04-27 13:30:52,596 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108788.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:30:54,436 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108791.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:31:04,845 INFO [finetune.py:976] (3/7) Epoch 19, batch 5700, loss[loss=0.1915, simple_loss=0.2453, pruned_loss=0.06886, over 3545.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2428, pruned_loss=0.05078, over 934590.70 frames. ], batch size: 15, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:31:40,721 INFO [finetune.py:976] (3/7) Epoch 20, batch 0, loss[loss=0.1115, simple_loss=0.1812, pruned_loss=0.02087, over 4714.00 frames. ], tot_loss[loss=0.1115, simple_loss=0.1812, pruned_loss=0.02087, over 4714.00 frames. ], batch size: 23, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:31:40,722 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 13:31:48,587 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0875, 2.6153, 1.1108, 1.4429, 1.7787, 1.3467, 2.9828, 1.7546], + device='cuda:3'), covar=tensor([0.0628, 0.0597, 0.0686, 0.1104, 0.0475, 0.0850, 0.0254, 0.0537], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 13:31:49,965 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2973, 1.5174, 1.8145, 1.9663, 1.8967, 2.0305, 1.8238, 1.8828], + device='cuda:3'), covar=tensor([0.4648, 0.5745, 0.4773, 0.4803, 0.6018, 0.7207, 0.6009, 0.5147], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0373, 0.0321, 0.0334, 0.0345, 0.0394, 0.0357, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:31:57,318 INFO [finetune.py:1010] (3/7) Epoch 20, validation: loss=0.1536, simple_loss=0.2249, pruned_loss=0.04109, over 2265189.00 frames. +2023-04-27 13:31:57,318 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 13:32:13,986 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:32:17,460 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.429e+02 1.789e+02 2.182e+02 4.169e+02, threshold=3.578e+02, percent-clipped=1.0 +2023-04-27 13:32:21,703 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2778, 1.6593, 1.5301, 1.8488, 1.7214, 2.2260, 1.4949, 3.7222], + device='cuda:3'), covar=tensor([0.0637, 0.0816, 0.0845, 0.1229, 0.0680, 0.0409, 0.0761, 0.0151], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 13:32:30,439 INFO [finetune.py:976] (3/7) Epoch 20, batch 50, loss[loss=0.1991, simple_loss=0.2493, pruned_loss=0.07449, over 4889.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2497, pruned_loss=0.05459, over 217497.35 frames. ], batch size: 43, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:32:45,185 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8359, 2.8362, 2.2410, 3.2742, 2.8810, 2.8253, 1.1556, 2.7574], + device='cuda:3'), covar=tensor([0.2448, 0.1592, 0.3328, 0.3097, 0.2904, 0.2297, 0.5923, 0.2963], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0213, 0.0247, 0.0303, 0.0293, 0.0246, 0.0269, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:33:03,608 INFO [finetune.py:976] (3/7) Epoch 20, batch 100, loss[loss=0.1761, simple_loss=0.2481, pruned_loss=0.05206, over 4912.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2427, pruned_loss=0.05077, over 382304.01 frames. ], batch size: 46, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:33:08,749 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:33:22,263 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7347, 2.0388, 1.7267, 2.0597, 1.5926, 1.7353, 1.7265, 1.3761], + device='cuda:3'), covar=tensor([0.1605, 0.1152, 0.0835, 0.1032, 0.3282, 0.1101, 0.1855, 0.2231], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0303, 0.0218, 0.0279, 0.0314, 0.0260, 0.0251, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1557e-04, 1.2006e-04, 8.6276e-05, 1.1079e-04, 1.2740e-04, 1.0276e-04, + 1.0141e-04, 1.0494e-04], device='cuda:3') +2023-04-27 13:33:23,934 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.275e+01 1.453e+02 1.763e+02 2.128e+02 3.737e+02, threshold=3.527e+02, percent-clipped=2.0 +2023-04-27 13:33:36,946 INFO [finetune.py:976] (3/7) Epoch 20, batch 150, loss[loss=0.1369, simple_loss=0.2089, pruned_loss=0.0325, over 4904.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2376, pruned_loss=0.04967, over 507986.99 frames. ], batch size: 43, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:33:39,882 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:33:40,995 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5227, 1.5046, 1.8264, 1.9173, 1.3949, 1.3065, 1.5242, 0.9584], + device='cuda:3'), covar=tensor([0.0615, 0.0714, 0.0396, 0.0538, 0.0783, 0.1185, 0.0714, 0.0694], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0074, 0.0096, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:33:59,804 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6961, 3.1589, 1.0343, 1.8723, 2.4741, 2.0413, 4.4389, 2.6845], + device='cuda:3'), covar=tensor([0.0594, 0.0741, 0.0841, 0.1271, 0.0516, 0.0809, 0.0299, 0.0503], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 13:34:09,747 INFO [finetune.py:976] (3/7) Epoch 20, batch 200, loss[loss=0.1698, simple_loss=0.24, pruned_loss=0.04984, over 4860.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2358, pruned_loss=0.04929, over 606922.26 frames. ], batch size: 44, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:34:09,853 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.1683, 3.4876, 1.4780, 2.3002, 2.8324, 2.5533, 4.5921, 3.1088], + device='cuda:3'), covar=tensor([0.0479, 0.0608, 0.0702, 0.1062, 0.0442, 0.0670, 0.0228, 0.0395], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 13:34:11,043 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109029.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:34:13,508 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9058, 2.2706, 0.8666, 1.2501, 1.4912, 1.2280, 2.4893, 1.4558], + device='cuda:3'), covar=tensor([0.0707, 0.0614, 0.0655, 0.1257, 0.0494, 0.0970, 0.0357, 0.0644], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 13:34:16,937 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109037.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:34:29,598 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.602e+02 1.908e+02 2.213e+02 3.365e+02, threshold=3.817e+02, percent-clipped=0.0 +2023-04-27 13:34:42,778 INFO [finetune.py:976] (3/7) Epoch 20, batch 250, loss[loss=0.157, simple_loss=0.2236, pruned_loss=0.04514, over 4764.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2398, pruned_loss=0.05019, over 684987.99 frames. ], batch size: 27, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:35:02,393 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:35:13,160 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109098.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:35:31,288 INFO [finetune.py:976] (3/7) Epoch 20, batch 300, loss[loss=0.1955, simple_loss=0.2792, pruned_loss=0.05585, over 4812.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2445, pruned_loss=0.05143, over 746560.90 frames. ], batch size: 40, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:35:49,172 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:35:56,827 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109147.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:36:03,480 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.663e+02 2.046e+02 2.623e+02 4.738e+02, threshold=4.093e+02, percent-clipped=3.0 +2023-04-27 13:36:18,047 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6382, 2.1662, 1.6315, 1.5239, 1.2736, 1.2855, 1.6785, 1.2161], + device='cuda:3'), covar=tensor([0.1564, 0.1235, 0.1437, 0.1637, 0.2245, 0.1923, 0.0932, 0.1950], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0169, 0.0204, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 13:36:25,684 INFO [finetune.py:976] (3/7) Epoch 20, batch 350, loss[loss=0.1831, simple_loss=0.2616, pruned_loss=0.05229, over 4816.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2469, pruned_loss=0.05237, over 793152.24 frames. ], batch size: 39, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:36:47,085 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109194.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:37:21,409 INFO [finetune.py:976] (3/7) Epoch 20, batch 400, loss[loss=0.1586, simple_loss=0.2347, pruned_loss=0.04122, over 4773.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2494, pruned_loss=0.05332, over 831327.02 frames. ], batch size: 26, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:37:31,621 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109233.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:03,286 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1285, 2.5744, 2.2194, 2.0809, 1.7103, 1.7299, 2.2658, 1.6633], + device='cuda:3'), covar=tensor([0.1578, 0.1244, 0.1361, 0.1538, 0.2197, 0.1794, 0.0886, 0.1892], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0211, 0.0168, 0.0202, 0.0199, 0.0183, 0.0154, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 13:38:03,877 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109255.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:05,568 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.658e+02 1.948e+02 2.358e+02 4.073e+02, threshold=3.896e+02, percent-clipped=0.0 +2023-04-27 13:38:22,932 INFO [finetune.py:976] (3/7) Epoch 20, batch 450, loss[loss=0.1573, simple_loss=0.2345, pruned_loss=0.04003, over 4815.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2474, pruned_loss=0.05203, over 858911.91 frames. ], batch size: 38, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:38:25,341 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3673, 3.4457, 0.9917, 1.8325, 1.8389, 2.4360, 1.9397, 1.0046], + device='cuda:3'), covar=tensor([0.1375, 0.1173, 0.1871, 0.1161, 0.1084, 0.0930, 0.1439, 0.1893], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0240, 0.0137, 0.0118, 0.0131, 0.0151, 0.0116, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:38:25,937 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109281.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:34,962 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:50,557 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109318.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:54,509 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-27 13:38:55,901 INFO [finetune.py:976] (3/7) Epoch 20, batch 500, loss[loss=0.2282, simple_loss=0.2896, pruned_loss=0.08339, over 4766.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2464, pruned_loss=0.05253, over 880181.59 frames. ], batch size: 26, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:39:16,056 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109355.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:39:17,761 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.900e+01 1.534e+02 1.939e+02 2.475e+02 3.939e+02, threshold=3.877e+02, percent-clipped=2.0 +2023-04-27 13:39:24,024 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1767, 1.8134, 1.5936, 2.1295, 2.3391, 1.9427, 1.8561, 1.6663], + device='cuda:3'), covar=tensor([0.1752, 0.1651, 0.1659, 0.1403, 0.1137, 0.1771, 0.2003, 0.2155], + device='cuda:3'), in_proj_covar=tensor([0.0303, 0.0306, 0.0344, 0.0284, 0.0320, 0.0302, 0.0295, 0.0364], + device='cuda:3'), out_proj_covar=tensor([6.2256e-05, 6.3329e-05, 7.2701e-05, 5.7506e-05, 6.6126e-05, 6.3451e-05, + 6.1823e-05, 7.7397e-05], device='cuda:3') +2023-04-27 13:39:29,353 INFO [finetune.py:976] (3/7) Epoch 20, batch 550, loss[loss=0.1401, simple_loss=0.2053, pruned_loss=0.03751, over 4742.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2423, pruned_loss=0.05123, over 895327.02 frames. ], batch size: 23, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:39:30,092 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5932, 1.4570, 1.7793, 1.8508, 1.4558, 1.3976, 1.5446, 1.0498], + device='cuda:3'), covar=tensor([0.0529, 0.0583, 0.0406, 0.0518, 0.0764, 0.1172, 0.0482, 0.0584], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0074, 0.0096, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:39:30,711 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109379.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:39:40,651 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109393.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:40:01,838 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9534, 1.1738, 1.6014, 1.7194, 1.6478, 1.7039, 1.6216, 1.6025], + device='cuda:3'), covar=tensor([0.3693, 0.4747, 0.4007, 0.3935, 0.5004, 0.6537, 0.4415, 0.4097], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0373, 0.0322, 0.0335, 0.0345, 0.0394, 0.0358, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:40:02,926 INFO [finetune.py:976] (3/7) Epoch 20, batch 600, loss[loss=0.1645, simple_loss=0.2368, pruned_loss=0.04611, over 4865.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2428, pruned_loss=0.0514, over 909334.28 frames. ], batch size: 34, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:40:10,663 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6436, 1.2715, 1.8306, 2.1775, 1.7973, 1.6214, 1.7138, 1.6912], + device='cuda:3'), covar=tensor([0.4419, 0.6221, 0.5809, 0.5406, 0.5558, 0.7273, 0.7317, 0.8063], + device='cuda:3'), in_proj_covar=tensor([0.0430, 0.0411, 0.0506, 0.0506, 0.0457, 0.0485, 0.0492, 0.0499], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:40:16,604 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109447.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:40:24,266 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.599e+02 2.024e+02 2.481e+02 5.741e+02, threshold=4.048e+02, percent-clipped=1.0 +2023-04-27 13:40:33,467 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-27 13:40:36,279 INFO [finetune.py:976] (3/7) Epoch 20, batch 650, loss[loss=0.1712, simple_loss=0.2451, pruned_loss=0.04867, over 4875.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2447, pruned_loss=0.05147, over 919723.86 frames. ], batch size: 34, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:40:48,774 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:41:10,002 INFO [finetune.py:976] (3/7) Epoch 20, batch 700, loss[loss=0.1665, simple_loss=0.2439, pruned_loss=0.04453, over 4807.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2471, pruned_loss=0.05215, over 927279.27 frames. ], batch size: 40, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:41:10,760 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9590, 1.5887, 1.5549, 1.7080, 2.1568, 1.7510, 1.4545, 1.4257], + device='cuda:3'), covar=tensor([0.1471, 0.1390, 0.1961, 0.1295, 0.0872, 0.1672, 0.2314, 0.2504], + device='cuda:3'), in_proj_covar=tensor([0.0304, 0.0307, 0.0346, 0.0285, 0.0321, 0.0304, 0.0296, 0.0366], + device='cuda:3'), out_proj_covar=tensor([6.2609e-05, 6.3584e-05, 7.3140e-05, 5.7737e-05, 6.6438e-05, 6.3810e-05, + 6.2081e-05, 7.7909e-05], device='cuda:3') +2023-04-27 13:41:14,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8924, 2.7240, 2.0168, 1.9834, 1.3606, 1.4405, 2.0863, 1.3839], + device='cuda:3'), covar=tensor([0.1700, 0.1268, 0.1465, 0.1626, 0.2421, 0.1943, 0.1005, 0.1990], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0211, 0.0168, 0.0203, 0.0199, 0.0184, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 13:41:25,596 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109550.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:41:27,181 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-27 13:41:30,317 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.590e+02 1.898e+02 2.337e+02 3.520e+02, threshold=3.796e+02, percent-clipped=0.0 +2023-04-27 13:41:43,341 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9054, 1.1509, 1.5246, 1.6259, 1.5789, 1.6681, 1.5591, 1.5897], + device='cuda:3'), covar=tensor([0.3893, 0.4814, 0.3967, 0.4112, 0.5105, 0.6863, 0.4783, 0.4358], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0375, 0.0323, 0.0337, 0.0347, 0.0396, 0.0360, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:41:43,810 INFO [finetune.py:976] (3/7) Epoch 20, batch 750, loss[loss=0.2336, simple_loss=0.2972, pruned_loss=0.085, over 4808.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2482, pruned_loss=0.05258, over 931643.44 frames. ], batch size: 40, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:41:45,095 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109579.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:42:49,901 INFO [finetune.py:976] (3/7) Epoch 20, batch 800, loss[loss=0.1507, simple_loss=0.2169, pruned_loss=0.0422, over 4865.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2473, pruned_loss=0.05157, over 937832.06 frames. ], batch size: 34, lr: 3.25e-03, grad_scale: 64.0 +2023-04-27 13:43:07,634 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109640.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:43:19,832 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:43:22,865 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1437, 1.6442, 1.4642, 1.7983, 1.7306, 2.0033, 1.5181, 3.4032], + device='cuda:3'), covar=tensor([0.0580, 0.0686, 0.0688, 0.1017, 0.0555, 0.0541, 0.0636, 0.0159], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 13:43:27,912 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.469e+02 1.785e+02 2.276e+02 4.254e+02, threshold=3.571e+02, percent-clipped=2.0 +2023-04-27 13:43:28,671 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:43:50,512 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109674.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:43:52,235 INFO [finetune.py:976] (3/7) Epoch 20, batch 850, loss[loss=0.1943, simple_loss=0.2614, pruned_loss=0.06363, over 4863.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2451, pruned_loss=0.05104, over 942565.09 frames. ], batch size: 34, lr: 3.25e-03, grad_scale: 64.0 +2023-04-27 13:44:08,037 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109693.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:44:26,929 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:44:31,597 INFO [finetune.py:976] (3/7) Epoch 20, batch 900, loss[loss=0.1432, simple_loss=0.2061, pruned_loss=0.04015, over 4372.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2432, pruned_loss=0.05063, over 945143.19 frames. ], batch size: 19, lr: 3.25e-03, grad_scale: 64.0 +2023-04-27 13:44:33,012 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-04-27 13:44:40,147 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109741.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:44:50,920 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.301e+02 1.595e+02 1.882e+02 2.269e+02 5.455e+02, threshold=3.764e+02, percent-clipped=0.0 +2023-04-27 13:45:03,940 INFO [finetune.py:976] (3/7) Epoch 20, batch 950, loss[loss=0.153, simple_loss=0.2128, pruned_loss=0.04658, over 4811.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2409, pruned_loss=0.05018, over 949418.14 frames. ], batch size: 25, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:45:33,354 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-04-27 13:45:38,013 INFO [finetune.py:976] (3/7) Epoch 20, batch 1000, loss[loss=0.2276, simple_loss=0.2949, pruned_loss=0.08011, over 4819.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2455, pruned_loss=0.05199, over 951021.28 frames. ], batch size: 40, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:45:52,567 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:45:58,961 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.126e+02 1.623e+02 1.911e+02 2.262e+02 3.883e+02, threshold=3.822e+02, percent-clipped=2.0 +2023-04-27 13:46:11,419 INFO [finetune.py:976] (3/7) Epoch 20, batch 1050, loss[loss=0.1351, simple_loss=0.2008, pruned_loss=0.03464, over 4715.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2481, pruned_loss=0.05219, over 953780.02 frames. ], batch size: 23, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:46:25,236 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109898.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:46:25,265 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5049, 3.0465, 1.0744, 1.9422, 1.7434, 2.3186, 1.8454, 1.2745], + device='cuda:3'), covar=tensor([0.1276, 0.0903, 0.1713, 0.1083, 0.1075, 0.0893, 0.1346, 0.1926], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0245, 0.0140, 0.0120, 0.0134, 0.0154, 0.0118, 0.0122], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:46:43,733 INFO [finetune.py:976] (3/7) Epoch 20, batch 1100, loss[loss=0.1706, simple_loss=0.2226, pruned_loss=0.05931, over 4284.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2487, pruned_loss=0.05226, over 954902.42 frames. ], batch size: 18, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:46:50,146 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109935.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:46:59,848 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:05,250 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.994e+01 1.686e+02 1.942e+02 2.391e+02 4.510e+02, threshold=3.883e+02, percent-clipped=3.0 +2023-04-27 13:47:16,127 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:17,880 INFO [finetune.py:976] (3/7) Epoch 20, batch 1150, loss[loss=0.1675, simple_loss=0.2413, pruned_loss=0.04681, over 4888.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2485, pruned_loss=0.05165, over 955067.29 frames. ], batch size: 35, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:47:33,810 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 13:47:40,684 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:52,883 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110015.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:54,531 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110017.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:57,539 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110022.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:59,292 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0306, 1.7187, 1.9797, 2.3895, 2.3818, 1.9597, 1.7045, 2.1240], + device='cuda:3'), covar=tensor([0.0777, 0.1061, 0.0709, 0.0511, 0.0579, 0.0809, 0.0700, 0.0542], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0200, 0.0182, 0.0172, 0.0177, 0.0180, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:48:00,998 INFO [finetune.py:976] (3/7) Epoch 20, batch 1200, loss[loss=0.1658, simple_loss=0.234, pruned_loss=0.04883, over 4818.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2456, pruned_loss=0.05049, over 954235.81 frames. ], batch size: 39, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:48:17,798 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110044.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:48:37,465 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.685e+02 1.958e+02 2.263e+02 7.161e+02, threshold=3.916e+02, percent-clipped=1.0 +2023-04-27 13:49:01,340 INFO [finetune.py:976] (3/7) Epoch 20, batch 1250, loss[loss=0.1994, simple_loss=0.2598, pruned_loss=0.06954, over 4810.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2436, pruned_loss=0.05047, over 955724.66 frames. ], batch size: 39, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:49:02,090 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:08,927 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3497, 1.3038, 1.3691, 1.6600, 1.7141, 1.3290, 0.9714, 1.4909], + device='cuda:3'), covar=tensor([0.0833, 0.1313, 0.0873, 0.0586, 0.0622, 0.0801, 0.0857, 0.0599], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0200, 0.0182, 0.0172, 0.0178, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:49:25,669 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110105.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:28,702 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:31,771 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:37,008 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5151, 1.4502, 1.7991, 1.8477, 1.4562, 1.2235, 1.5056, 0.9609], + device='cuda:3'), covar=tensor([0.0575, 0.0623, 0.0399, 0.0547, 0.0716, 0.1055, 0.0567, 0.0643], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0074, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:49:40,408 INFO [finetune.py:976] (3/7) Epoch 20, batch 1300, loss[loss=0.1788, simple_loss=0.2376, pruned_loss=0.06004, over 4935.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2404, pruned_loss=0.04962, over 954418.98 frames. ], batch size: 38, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:50:01,795 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.513e+02 1.801e+02 2.249e+02 8.027e+02, threshold=3.601e+02, percent-clipped=2.0 +2023-04-27 13:50:09,714 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:50:13,297 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:50:13,764 INFO [finetune.py:976] (3/7) Epoch 20, batch 1350, loss[loss=0.1551, simple_loss=0.2216, pruned_loss=0.04431, over 4289.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2423, pruned_loss=0.0505, over 953415.03 frames. ], batch size: 18, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:50:16,285 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-04-27 13:50:39,146 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 13:50:47,123 INFO [finetune.py:976] (3/7) Epoch 20, batch 1400, loss[loss=0.1981, simple_loss=0.267, pruned_loss=0.06458, over 4905.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2459, pruned_loss=0.05186, over 955775.68 frames. ], batch size: 36, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:50:52,579 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110235.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:09,026 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.619e+02 1.811e+02 2.225e+02 6.587e+02, threshold=3.621e+02, percent-clipped=5.0 +2023-04-27 13:51:14,116 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.16 vs. limit=5.0 +2023-04-27 13:51:20,001 INFO [finetune.py:976] (3/7) Epoch 20, batch 1450, loss[loss=0.1565, simple_loss=0.2285, pruned_loss=0.04231, over 4715.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.247, pruned_loss=0.05193, over 955585.73 frames. ], batch size: 23, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:51:25,117 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:25,186 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:38,767 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110303.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:46,531 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110315.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:48,398 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110318.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:51,169 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 13:51:53,750 INFO [finetune.py:976] (3/7) Epoch 20, batch 1500, loss[loss=0.2042, simple_loss=0.2675, pruned_loss=0.07047, over 4821.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.248, pruned_loss=0.0522, over 955996.15 frames. ], batch size: 47, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:52:05,643 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110344.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:52:16,147 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.560e+02 1.895e+02 2.185e+02 4.822e+02, threshold=3.791e+02, percent-clipped=2.0 +2023-04-27 13:52:19,125 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110363.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:52:19,791 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110364.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:52:25,268 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110373.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:52:27,673 INFO [finetune.py:976] (3/7) Epoch 20, batch 1550, loss[loss=0.1969, simple_loss=0.2667, pruned_loss=0.06355, over 4751.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2489, pruned_loss=0.05266, over 954892.35 frames. ], batch size: 54, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:52:29,080 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110379.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:53:00,009 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110400.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:53:22,804 INFO [finetune.py:976] (3/7) Epoch 20, batch 1600, loss[loss=0.1829, simple_loss=0.253, pruned_loss=0.0564, over 4897.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2474, pruned_loss=0.05252, over 954453.43 frames. ], batch size: 35, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:54:08,139 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.546e+02 1.968e+02 2.341e+02 5.862e+02, threshold=3.936e+02, percent-clipped=2.0 +2023-04-27 13:54:17,898 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:54:21,384 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:54:30,175 INFO [finetune.py:976] (3/7) Epoch 20, batch 1650, loss[loss=0.145, simple_loss=0.2185, pruned_loss=0.03568, over 4909.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2442, pruned_loss=0.05159, over 953494.33 frames. ], batch size: 32, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:54:48,671 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5119, 1.3774, 1.7272, 1.8294, 1.3601, 1.2659, 1.4396, 0.9645], + device='cuda:3'), covar=tensor([0.0507, 0.0622, 0.0363, 0.0612, 0.0722, 0.1130, 0.0635, 0.0579], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0069, 0.0067, 0.0068, 0.0075, 0.0097, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 13:55:09,360 INFO [finetune.py:976] (3/7) Epoch 20, batch 1700, loss[loss=0.1775, simple_loss=0.238, pruned_loss=0.05853, over 4208.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2412, pruned_loss=0.05038, over 953892.04 frames. ], batch size: 18, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:55:22,345 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3421, 1.3297, 1.3597, 1.5886, 1.6802, 1.2911, 0.9024, 1.4539], + device='cuda:3'), covar=tensor([0.0879, 0.1235, 0.0921, 0.0693, 0.0700, 0.0928, 0.0877, 0.0689], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0200, 0.0183, 0.0173, 0.0178, 0.0181, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:55:31,158 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.338e+01 1.550e+02 1.876e+02 2.294e+02 4.137e+02, threshold=3.752e+02, percent-clipped=2.0 +2023-04-27 13:55:43,125 INFO [finetune.py:976] (3/7) Epoch 20, batch 1750, loss[loss=0.1683, simple_loss=0.2481, pruned_loss=0.04427, over 4861.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2433, pruned_loss=0.05085, over 952581.75 frames. ], batch size: 34, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:56:06,133 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110611.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:17,236 INFO [finetune.py:976] (3/7) Epoch 20, batch 1800, loss[loss=0.2079, simple_loss=0.2806, pruned_loss=0.06755, over 4717.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2462, pruned_loss=0.05163, over 953747.75 frames. ], batch size: 59, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:56:24,701 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110639.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:38,214 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.583e+02 1.893e+02 2.198e+02 3.811e+02, threshold=3.786e+02, percent-clipped=1.0 +2023-04-27 13:56:38,790 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:47,740 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110672.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:48,306 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110673.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:48,913 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110674.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:50,674 INFO [finetune.py:976] (3/7) Epoch 20, batch 1850, loss[loss=0.1543, simple_loss=0.2299, pruned_loss=0.03933, over 4742.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2476, pruned_loss=0.05207, over 953288.14 frames. ], batch size: 54, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:56:56,816 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9554, 2.3519, 0.9994, 1.3093, 1.7765, 1.1758, 2.9784, 1.4694], + device='cuda:3'), covar=tensor([0.0712, 0.0565, 0.0722, 0.1163, 0.0494, 0.1033, 0.0253, 0.0681], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 13:57:06,464 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110700.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:21,095 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110721.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:23,591 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2997, 3.3474, 2.3961, 3.8625, 3.3797, 3.3498, 1.4730, 3.3216], + device='cuda:3'), covar=tensor([0.2233, 0.1495, 0.3524, 0.2294, 0.3529, 0.1937, 0.5979, 0.2705], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0214, 0.0248, 0.0301, 0.0293, 0.0245, 0.0270, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 13:57:24,741 INFO [finetune.py:976] (3/7) Epoch 20, batch 1900, loss[loss=0.1953, simple_loss=0.2746, pruned_loss=0.05803, over 4827.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2497, pruned_loss=0.05286, over 954278.05 frames. ], batch size: 47, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:57:30,839 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:32,672 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8846, 1.8391, 1.5332, 1.4654, 1.7642, 1.5245, 2.1664, 1.2543], + device='cuda:3'), covar=tensor([0.3185, 0.1398, 0.4479, 0.2252, 0.1494, 0.1904, 0.1282, 0.4667], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0348, 0.0427, 0.0353, 0.0383, 0.0376, 0.0374, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 13:57:38,119 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110748.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:46,352 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.050e+01 1.575e+02 1.913e+02 2.383e+02 8.631e+02, threshold=3.825e+02, percent-clipped=5.0 +2023-04-27 13:57:51,431 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:53,298 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1043, 1.7055, 1.6248, 1.7983, 1.7792, 2.0204, 1.4935, 3.6135], + device='cuda:3'), covar=tensor([0.0618, 0.0770, 0.0703, 0.1158, 0.0583, 0.0491, 0.0708, 0.0132], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 13:57:55,036 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110771.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:59,103 INFO [finetune.py:976] (3/7) Epoch 20, batch 1950, loss[loss=0.1817, simple_loss=0.2541, pruned_loss=0.0546, over 4816.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2478, pruned_loss=0.05236, over 952431.76 frames. ], batch size: 41, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:58:29,600 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110797.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:58:52,671 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110814.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:59:01,943 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:59:11,707 INFO [finetune.py:976] (3/7) Epoch 20, batch 2000, loss[loss=0.1628, simple_loss=0.2289, pruned_loss=0.04836, over 4755.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2439, pruned_loss=0.05103, over 954988.27 frames. ], batch size: 28, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:59:28,421 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-27 13:59:33,917 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.549e+02 1.813e+02 2.155e+02 3.715e+02, threshold=3.626e+02, percent-clipped=0.0 +2023-04-27 13:59:45,757 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110868.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:59:56,921 INFO [finetune.py:976] (3/7) Epoch 20, batch 2050, loss[loss=0.1277, simple_loss=0.2027, pruned_loss=0.02638, over 4754.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2413, pruned_loss=0.05016, over 956481.26 frames. ], batch size: 28, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:00:10,169 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8556, 1.3207, 1.9495, 2.3999, 1.9517, 1.7851, 1.8484, 1.7960], + device='cuda:3'), covar=tensor([0.4732, 0.6996, 0.6882, 0.5809, 0.5827, 0.8354, 0.8525, 0.9225], + device='cuda:3'), in_proj_covar=tensor([0.0431, 0.0412, 0.0507, 0.0509, 0.0458, 0.0486, 0.0493, 0.0498], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:00:53,299 INFO [finetune.py:976] (3/7) Epoch 20, batch 2100, loss[loss=0.228, simple_loss=0.2827, pruned_loss=0.08663, over 4754.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2412, pruned_loss=0.05075, over 954414.46 frames. ], batch size: 26, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:00:54,652 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:01,123 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2068, 1.6625, 2.1196, 2.5525, 2.0558, 1.6314, 1.3998, 1.8787], + device='cuda:3'), covar=tensor([0.3023, 0.3058, 0.1580, 0.2112, 0.2429, 0.2481, 0.4077, 0.1959], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0246, 0.0227, 0.0316, 0.0219, 0.0233, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 14:01:01,702 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110939.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:13,867 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110958.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:14,356 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.601e+02 1.866e+02 2.314e+02 3.983e+02, threshold=3.732e+02, percent-clipped=4.0 +2023-04-27 14:01:14,952 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110959.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:20,708 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110967.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:25,023 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:27,215 INFO [finetune.py:976] (3/7) Epoch 20, batch 2150, loss[loss=0.2493, simple_loss=0.3171, pruned_loss=0.09079, over 4783.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.244, pruned_loss=0.05145, over 954134.92 frames. ], batch size: 54, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:01:34,417 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:47,107 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111007.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:50,729 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 14:01:55,866 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111019.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:57,637 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111022.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:02:00,612 INFO [finetune.py:976] (3/7) Epoch 20, batch 2200, loss[loss=0.1339, simple_loss=0.2025, pruned_loss=0.03263, over 3973.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2472, pruned_loss=0.05252, over 954199.85 frames. ], batch size: 17, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:02:22,144 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.590e+02 1.963e+02 2.328e+02 5.181e+02, threshold=3.927e+02, percent-clipped=3.0 +2023-04-27 14:02:31,826 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8209, 3.0374, 2.4565, 2.7170, 3.0546, 2.5449, 3.8830, 2.3722], + device='cuda:3'), covar=tensor([0.3792, 0.1864, 0.4287, 0.3059, 0.1902, 0.2475, 0.1478, 0.3882], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0349, 0.0429, 0.0354, 0.0384, 0.0376, 0.0375, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:02:34,615 INFO [finetune.py:976] (3/7) Epoch 20, batch 2250, loss[loss=0.2143, simple_loss=0.2798, pruned_loss=0.07436, over 4773.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2481, pruned_loss=0.05279, over 953916.24 frames. ], batch size: 51, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:02:40,765 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6792, 2.2186, 1.7989, 1.7197, 1.2795, 1.3195, 1.8292, 1.2328], + device='cuda:3'), covar=tensor([0.1501, 0.1288, 0.1437, 0.1599, 0.2291, 0.1882, 0.0935, 0.1930], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0211, 0.0169, 0.0203, 0.0199, 0.0184, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 14:02:44,877 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111092.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 14:03:02,562 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.02 vs. limit=5.0 +2023-04-27 14:03:08,137 INFO [finetune.py:976] (3/7) Epoch 20, batch 2300, loss[loss=0.1709, simple_loss=0.2467, pruned_loss=0.04757, over 4900.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2481, pruned_loss=0.05186, over 955337.34 frames. ], batch size: 37, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:03:29,169 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.334e+01 1.540e+02 1.823e+02 2.154e+02 3.608e+02, threshold=3.645e+02, percent-clipped=0.0 +2023-04-27 14:03:41,035 INFO [finetune.py:976] (3/7) Epoch 20, batch 2350, loss[loss=0.1974, simple_loss=0.2656, pruned_loss=0.06458, over 4894.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2466, pruned_loss=0.05138, over 956559.90 frames. ], batch size: 32, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:04:27,659 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111224.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:04:28,330 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:04:29,450 INFO [finetune.py:976] (3/7) Epoch 20, batch 2400, loss[loss=0.1647, simple_loss=0.2269, pruned_loss=0.05119, over 4279.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2441, pruned_loss=0.05103, over 954535.30 frames. ], batch size: 65, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:05:02,142 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.830e+01 1.579e+02 1.923e+02 2.296e+02 7.513e+02, threshold=3.847e+02, percent-clipped=3.0 +2023-04-27 14:05:07,602 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111267.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:05:19,544 INFO [finetune.py:976] (3/7) Epoch 20, batch 2450, loss[loss=0.2133, simple_loss=0.2766, pruned_loss=0.07497, over 4827.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2418, pruned_loss=0.05071, over 955325.66 frames. ], batch size: 40, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:05:36,567 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111286.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:06:05,501 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111314.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:06:06,091 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111315.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:06:13,838 INFO [finetune.py:976] (3/7) Epoch 20, batch 2500, loss[loss=0.1795, simple_loss=0.2575, pruned_loss=0.05077, over 4845.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2423, pruned_loss=0.05103, over 956038.64 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:06:22,336 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3885, 2.9332, 2.4768, 2.7119, 2.1027, 2.5627, 2.6209, 2.1209], + device='cuda:3'), covar=tensor([0.1950, 0.1085, 0.0731, 0.1132, 0.2998, 0.0984, 0.1816, 0.2432], + device='cuda:3'), in_proj_covar=tensor([0.0288, 0.0302, 0.0218, 0.0279, 0.0313, 0.0259, 0.0250, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1525e-04, 1.1985e-04, 8.6186e-05, 1.1053e-04, 1.2685e-04, 1.0273e-04, + 1.0099e-04, 1.0417e-04], device='cuda:3') +2023-04-27 14:06:35,392 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.671e+02 1.878e+02 2.269e+02 4.413e+02, threshold=3.755e+02, percent-clipped=2.0 +2023-04-27 14:06:46,741 INFO [finetune.py:976] (3/7) Epoch 20, batch 2550, loss[loss=0.1686, simple_loss=0.2352, pruned_loss=0.05094, over 4809.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2451, pruned_loss=0.05132, over 953528.51 frames. ], batch size: 40, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:06:57,258 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111392.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 14:07:12,880 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0223, 1.2789, 3.1246, 2.9126, 2.7919, 2.9708, 2.9489, 2.7836], + device='cuda:3'), covar=tensor([0.6759, 0.4776, 0.1281, 0.1682, 0.1476, 0.2187, 0.4060, 0.1581], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0306, 0.0409, 0.0409, 0.0354, 0.0410, 0.0315, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:07:20,615 INFO [finetune.py:976] (3/7) Epoch 20, batch 2600, loss[loss=0.1495, simple_loss=0.2225, pruned_loss=0.03829, over 4780.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2478, pruned_loss=0.05192, over 951982.70 frames. ], batch size: 26, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:07:21,004 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-27 14:07:22,130 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.67 vs. limit=5.0 +2023-04-27 14:07:29,127 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111440.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:07:42,988 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.630e+02 1.917e+02 2.393e+02 4.047e+02, threshold=3.834e+02, percent-clipped=2.0 +2023-04-27 14:07:48,665 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111468.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:07:54,483 INFO [finetune.py:976] (3/7) Epoch 20, batch 2650, loss[loss=0.1892, simple_loss=0.2643, pruned_loss=0.05702, over 4821.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2488, pruned_loss=0.05271, over 951413.68 frames. ], batch size: 30, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:08:26,346 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111524.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:08:28,081 INFO [finetune.py:976] (3/7) Epoch 20, batch 2700, loss[loss=0.1877, simple_loss=0.2608, pruned_loss=0.05732, over 4869.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2488, pruned_loss=0.05283, over 951031.01 frames. ], batch size: 34, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:08:29,411 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111529.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:08:40,381 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-04-27 14:08:49,481 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.491e+02 1.860e+02 2.271e+02 3.945e+02, threshold=3.719e+02, percent-clipped=1.0 +2023-04-27 14:08:58,400 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111572.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:09:01,904 INFO [finetune.py:976] (3/7) Epoch 20, batch 2750, loss[loss=0.1471, simple_loss=0.2239, pruned_loss=0.03514, over 4797.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.246, pruned_loss=0.05165, over 952510.74 frames. ], batch size: 29, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:09:04,326 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111581.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:09:06,754 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0007, 1.8064, 1.9926, 2.3364, 2.4223, 2.0161, 1.6151, 2.2057], + device='cuda:3'), covar=tensor([0.0757, 0.1172, 0.0690, 0.0554, 0.0521, 0.0793, 0.0747, 0.0525], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0202, 0.0183, 0.0172, 0.0178, 0.0181, 0.0154, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:09:37,595 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111614.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:09:57,463 INFO [finetune.py:976] (3/7) Epoch 20, batch 2800, loss[loss=0.1052, simple_loss=0.1833, pruned_loss=0.01356, over 4844.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2431, pruned_loss=0.05089, over 952479.69 frames. ], batch size: 47, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:10:14,522 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0287, 1.7745, 2.2635, 2.5160, 2.0373, 2.0229, 2.1482, 2.0457], + device='cuda:3'), covar=tensor([0.4913, 0.7300, 0.7051, 0.5596, 0.6388, 0.8576, 0.8929, 0.9503], + device='cuda:3'), in_proj_covar=tensor([0.0431, 0.0413, 0.0508, 0.0511, 0.0460, 0.0489, 0.0494, 0.0501], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:10:24,454 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.707e+01 1.587e+02 1.829e+02 2.177e+02 6.182e+02, threshold=3.659e+02, percent-clipped=2.0 +2023-04-27 14:10:26,839 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111662.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:10:36,927 INFO [finetune.py:976] (3/7) Epoch 20, batch 2850, loss[loss=0.1783, simple_loss=0.2538, pruned_loss=0.05142, over 4855.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2425, pruned_loss=0.05123, over 954158.42 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:10:42,914 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3973, 2.8363, 1.0180, 1.5336, 2.1560, 1.3674, 3.9980, 2.0004], + device='cuda:3'), covar=tensor([0.0647, 0.0841, 0.0888, 0.1238, 0.0506, 0.0996, 0.0189, 0.0590], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0049, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 14:11:31,380 INFO [finetune.py:976] (3/7) Epoch 20, batch 2900, loss[loss=0.228, simple_loss=0.2992, pruned_loss=0.0784, over 4905.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2437, pruned_loss=0.05099, over 954352.39 frames. ], batch size: 36, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:12:12,495 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.619e+02 1.984e+02 2.443e+02 6.162e+02, threshold=3.968e+02, percent-clipped=7.0 +2023-04-27 14:12:13,616 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.43 vs. limit=5.0 +2023-04-27 14:12:33,176 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2194, 1.4543, 1.3881, 1.6850, 1.5819, 1.8694, 1.4134, 3.3714], + device='cuda:3'), covar=tensor([0.0594, 0.0797, 0.0784, 0.1204, 0.0615, 0.0538, 0.0733, 0.0134], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 14:12:36,595 INFO [finetune.py:976] (3/7) Epoch 20, batch 2950, loss[loss=0.1369, simple_loss=0.1967, pruned_loss=0.03856, over 3975.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2451, pruned_loss=0.05113, over 951833.38 frames. ], batch size: 17, lr: 3.24e-03, grad_scale: 64.0 +2023-04-27 14:12:56,805 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111791.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 14:13:29,946 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111824.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:13:32,106 INFO [finetune.py:976] (3/7) Epoch 20, batch 3000, loss[loss=0.1686, simple_loss=0.2448, pruned_loss=0.0462, over 4841.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2471, pruned_loss=0.05226, over 951403.64 frames. ], batch size: 49, lr: 3.24e-03, grad_scale: 64.0 +2023-04-27 14:13:32,106 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 14:13:49,409 INFO [finetune.py:1010] (3/7) Epoch 20, validation: loss=0.1527, simple_loss=0.2229, pruned_loss=0.04123, over 2265189.00 frames. +2023-04-27 14:13:49,409 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 14:14:19,469 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1148, 2.4461, 2.4242, 2.8121, 2.5947, 2.6290, 2.5195, 4.9583], + device='cuda:3'), covar=tensor([0.0447, 0.0641, 0.0600, 0.0895, 0.0489, 0.0396, 0.0545, 0.0085], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 14:14:20,673 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111852.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 14:14:27,130 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 14:14:29,831 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.660e+02 1.972e+02 2.446e+02 4.506e+02, threshold=3.944e+02, percent-clipped=2.0 +2023-04-27 14:14:51,104 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-27 14:14:58,620 INFO [finetune.py:976] (3/7) Epoch 20, batch 3050, loss[loss=0.1785, simple_loss=0.2659, pruned_loss=0.04555, over 4892.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2474, pruned_loss=0.05199, over 952302.29 frames. ], batch size: 43, lr: 3.24e-03, grad_scale: 64.0 +2023-04-27 14:15:01,099 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111881.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:15:10,282 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8637, 4.3045, 0.9860, 1.9900, 2.3172, 2.7200, 2.4190, 0.9237], + device='cuda:3'), covar=tensor([0.1328, 0.0691, 0.1966, 0.1267, 0.1059, 0.1036, 0.1289, 0.2177], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0241, 0.0138, 0.0120, 0.0134, 0.0152, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 14:15:11,141 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-04-27 14:15:12,745 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4233, 0.9726, 0.5334, 1.1270, 1.1435, 1.3266, 1.2114, 1.1802], + device='cuda:3'), covar=tensor([0.0508, 0.0391, 0.0387, 0.0553, 0.0296, 0.0492, 0.0462, 0.0588], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0049, 0.0051], + device='cuda:3') +2023-04-27 14:15:48,560 INFO [finetune.py:976] (3/7) Epoch 20, batch 3100, loss[loss=0.1617, simple_loss=0.2317, pruned_loss=0.04585, over 4810.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.246, pruned_loss=0.05112, over 954484.83 frames. ], batch size: 39, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:15:50,329 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:16:16,420 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.458e+02 1.721e+02 2.109e+02 3.576e+02, threshold=3.441e+02, percent-clipped=0.0 +2023-04-27 14:16:21,359 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3183, 1.5451, 1.5836, 2.2082, 2.2915, 1.9902, 1.9111, 1.5884], + device='cuda:3'), covar=tensor([0.1825, 0.1708, 0.1868, 0.1485, 0.1411, 0.1755, 0.1900, 0.2180], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0310, 0.0350, 0.0288, 0.0326, 0.0306, 0.0300, 0.0370], + device='cuda:3'), out_proj_covar=tensor([6.3583e-05, 6.4086e-05, 7.4122e-05, 5.8170e-05, 6.7464e-05, 6.4256e-05, + 6.2841e-05, 7.8684e-05], device='cuda:3') +2023-04-27 14:16:27,776 INFO [finetune.py:976] (3/7) Epoch 20, batch 3150, loss[loss=0.1848, simple_loss=0.2489, pruned_loss=0.06032, over 4822.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2446, pruned_loss=0.05095, over 956431.22 frames. ], batch size: 30, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:16:38,437 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4922, 2.6748, 2.0412, 2.2999, 2.6389, 2.0162, 3.4091, 1.9570], + device='cuda:3'), covar=tensor([0.3649, 0.1904, 0.4242, 0.3248, 0.1940, 0.2743, 0.1718, 0.4320], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0348, 0.0425, 0.0353, 0.0380, 0.0375, 0.0372, 0.0416], + device='cuda:3'), out_proj_covar=tensor([1.0000e-04, 1.0436e-04, 1.2896e-04, 1.0660e-04, 1.1326e-04, 1.1189e-04, + 1.0968e-04, 1.2582e-04], device='cuda:3') +2023-04-27 14:16:49,220 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-04-27 14:17:02,006 INFO [finetune.py:976] (3/7) Epoch 20, batch 3200, loss[loss=0.1985, simple_loss=0.2623, pruned_loss=0.06735, over 4905.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2426, pruned_loss=0.0505, over 956686.36 frames. ], batch size: 35, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:17:13,010 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7516, 2.2755, 1.7179, 1.6801, 1.2892, 1.3674, 1.9181, 1.2440], + device='cuda:3'), covar=tensor([0.1907, 0.1579, 0.1731, 0.1986, 0.2670, 0.2333, 0.1102, 0.2280], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0211, 0.0168, 0.0203, 0.0200, 0.0185, 0.0155, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 14:17:24,080 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5350, 3.4838, 0.9503, 1.7852, 1.8878, 2.4501, 1.9905, 0.9874], + device='cuda:3'), covar=tensor([0.1364, 0.0837, 0.1920, 0.1272, 0.1093, 0.0951, 0.1323, 0.2049], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0241, 0.0137, 0.0120, 0.0133, 0.0152, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 14:17:35,523 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.584e+02 1.831e+02 2.274e+02 4.743e+02, threshold=3.663e+02, percent-clipped=4.0 +2023-04-27 14:17:39,740 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112066.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:17:46,880 INFO [finetune.py:976] (3/7) Epoch 20, batch 3250, loss[loss=0.1545, simple_loss=0.2253, pruned_loss=0.04191, over 4737.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2434, pruned_loss=0.0511, over 956330.59 frames. ], batch size: 23, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:17:48,860 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.0310, 2.3488, 2.2545, 2.3795, 2.1614, 2.2469, 2.3396, 2.2713], + device='cuda:3'), covar=tensor([0.3874, 0.6225, 0.5001, 0.4600, 0.5790, 0.7436, 0.6074, 0.5447], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0373, 0.0321, 0.0334, 0.0345, 0.0393, 0.0356, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:18:10,426 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-27 14:18:18,441 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:18:20,139 INFO [finetune.py:976] (3/7) Epoch 20, batch 3300, loss[loss=0.1751, simple_loss=0.254, pruned_loss=0.04806, over 4770.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2471, pruned_loss=0.05177, over 956780.60 frames. ], batch size: 28, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:18:20,253 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112127.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:18:42,667 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8760, 1.0260, 3.2793, 3.0352, 2.9405, 3.2402, 3.2319, 2.8785], + device='cuda:3'), covar=tensor([0.7147, 0.5777, 0.1491, 0.2235, 0.1341, 0.1760, 0.1465, 0.1869], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0307, 0.0406, 0.0406, 0.0351, 0.0408, 0.0314, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:18:44,824 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112147.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 14:18:52,672 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.675e+01 1.599e+02 1.882e+02 2.331e+02 3.289e+02, threshold=3.764e+02, percent-clipped=0.0 +2023-04-27 14:19:00,556 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112172.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:19:04,158 INFO [finetune.py:976] (3/7) Epoch 20, batch 3350, loss[loss=0.2195, simple_loss=0.2795, pruned_loss=0.07973, over 4760.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2496, pruned_loss=0.05272, over 955858.19 frames. ], batch size: 54, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:19:15,105 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8302, 2.3964, 1.8295, 1.5942, 1.3331, 1.3975, 1.8770, 1.3014], + device='cuda:3'), covar=tensor([0.1680, 0.1303, 0.1529, 0.1794, 0.2485, 0.2013, 0.1050, 0.2111], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0211, 0.0168, 0.0203, 0.0200, 0.0185, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 14:19:37,509 INFO [finetune.py:976] (3/7) Epoch 20, batch 3400, loss[loss=0.1764, simple_loss=0.2599, pruned_loss=0.04646, over 4811.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2504, pruned_loss=0.05292, over 955728.29 frames. ], batch size: 40, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:19:37,694 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 14:20:15,442 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.600e+02 1.895e+02 2.268e+02 5.639e+02, threshold=3.789e+02, percent-clipped=3.0 +2023-04-27 14:20:37,464 INFO [finetune.py:976] (3/7) Epoch 20, batch 3450, loss[loss=0.1448, simple_loss=0.2216, pruned_loss=0.03403, over 4830.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.25, pruned_loss=0.05307, over 955199.59 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:20:48,943 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2889, 1.4759, 1.8073, 1.8840, 1.8100, 1.9016, 1.8383, 1.8533], + device='cuda:3'), covar=tensor([0.3660, 0.5044, 0.4544, 0.4492, 0.5438, 0.6848, 0.4984, 0.4719], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0372, 0.0321, 0.0334, 0.0344, 0.0392, 0.0355, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:20:50,696 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8664, 2.4620, 1.9042, 1.6885, 1.3756, 1.4211, 1.9370, 1.3177], + device='cuda:3'), covar=tensor([0.1500, 0.1257, 0.1366, 0.1738, 0.2216, 0.1879, 0.0948, 0.1939], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0213, 0.0169, 0.0205, 0.0201, 0.0186, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 14:21:35,778 INFO [finetune.py:976] (3/7) Epoch 20, batch 3500, loss[loss=0.1282, simple_loss=0.2033, pruned_loss=0.02657, over 4838.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2463, pruned_loss=0.0521, over 953933.58 frames. ], batch size: 47, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:21:59,356 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.563e+02 1.868e+02 2.164e+02 4.202e+02, threshold=3.735e+02, percent-clipped=1.0 +2023-04-27 14:22:09,767 INFO [finetune.py:976] (3/7) Epoch 20, batch 3550, loss[loss=0.1604, simple_loss=0.233, pruned_loss=0.04391, over 4816.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2432, pruned_loss=0.05173, over 955031.03 frames. ], batch size: 41, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:22:15,247 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2856, 1.6123, 1.3894, 1.7234, 1.6551, 1.8031, 1.4067, 3.2107], + device='cuda:3'), covar=tensor([0.0583, 0.0710, 0.0736, 0.1074, 0.0584, 0.0551, 0.0681, 0.0158], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 14:22:18,181 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112389.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:22:30,158 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-27 14:22:37,946 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2493, 1.8117, 1.6629, 1.9853, 1.9942, 2.1441, 1.6979, 4.2382], + device='cuda:3'), covar=tensor([0.0575, 0.0782, 0.0738, 0.1190, 0.0599, 0.0492, 0.0714, 0.0118], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 14:22:40,994 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112422.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:22:44,030 INFO [finetune.py:976] (3/7) Epoch 20, batch 3600, loss[loss=0.1267, simple_loss=0.1991, pruned_loss=0.0271, over 4899.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2401, pruned_loss=0.05039, over 956962.75 frames. ], batch size: 32, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:22:57,487 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112447.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:22:59,348 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112450.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:23:06,297 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.532e+02 1.802e+02 2.141e+02 3.664e+02, threshold=3.603e+02, percent-clipped=0.0 +2023-04-27 14:23:18,444 INFO [finetune.py:976] (3/7) Epoch 20, batch 3650, loss[loss=0.1914, simple_loss=0.2756, pruned_loss=0.05357, over 4829.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2418, pruned_loss=0.05082, over 954387.06 frames. ], batch size: 39, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:23:30,564 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112495.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:23:58,093 INFO [finetune.py:976] (3/7) Epoch 20, batch 3700, loss[loss=0.2215, simple_loss=0.2893, pruned_loss=0.07687, over 4917.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2447, pruned_loss=0.05184, over 955233.77 frames. ], batch size: 38, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:24:41,780 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.566e+02 1.857e+02 2.073e+02 3.044e+02, threshold=3.713e+02, percent-clipped=0.0 +2023-04-27 14:24:48,661 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112562.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:25:05,135 INFO [finetune.py:976] (3/7) Epoch 20, batch 3750, loss[loss=0.2024, simple_loss=0.2671, pruned_loss=0.06889, over 4920.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2474, pruned_loss=0.0525, over 954684.08 frames. ], batch size: 38, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:25:05,269 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:25:26,584 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5668, 2.1165, 2.4158, 3.1014, 2.4512, 1.9665, 1.9307, 2.4532], + device='cuda:3'), covar=tensor([0.3464, 0.3185, 0.1705, 0.2306, 0.2748, 0.2668, 0.3638, 0.1985], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0245, 0.0226, 0.0312, 0.0218, 0.0232, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 14:25:56,706 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4395, 1.9602, 2.3534, 2.8636, 2.4278, 1.9241, 2.0134, 2.2903], + device='cuda:3'), covar=tensor([0.3130, 0.3015, 0.1549, 0.2282, 0.2388, 0.2561, 0.3267, 0.1848], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0245, 0.0225, 0.0311, 0.0217, 0.0232, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 14:26:00,841 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:26:11,161 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112623.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:26:19,269 INFO [finetune.py:976] (3/7) Epoch 20, batch 3800, loss[loss=0.1725, simple_loss=0.258, pruned_loss=0.0435, over 4861.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2479, pruned_loss=0.0528, over 953038.25 frames. ], batch size: 31, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:26:31,085 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112638.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 14:26:43,666 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-04-27 14:27:02,948 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.520e+02 1.863e+02 2.243e+02 5.060e+02, threshold=3.726e+02, percent-clipped=6.0 +2023-04-27 14:27:17,783 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7992, 1.5964, 1.7524, 2.0704, 2.0871, 1.6934, 1.3495, 1.8791], + device='cuda:3'), covar=tensor([0.0782, 0.1246, 0.0754, 0.0577, 0.0588, 0.0802, 0.0828, 0.0576], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0202, 0.0183, 0.0172, 0.0178, 0.0181, 0.0153, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:27:19,497 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6341, 1.9408, 1.8863, 1.9806, 1.8083, 1.9560, 2.0165, 1.9180], + device='cuda:3'), covar=tensor([0.3564, 0.5004, 0.4638, 0.4687, 0.5437, 0.7286, 0.5302, 0.4885], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0374, 0.0322, 0.0335, 0.0345, 0.0394, 0.0357, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:27:25,229 INFO [finetune.py:976] (3/7) Epoch 20, batch 3850, loss[loss=0.1875, simple_loss=0.2636, pruned_loss=0.05571, over 4816.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2469, pruned_loss=0.05227, over 954115.63 frames. ], batch size: 39, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:27:25,940 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112678.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 14:27:56,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4605, 2.3406, 2.5159, 2.8612, 2.9399, 2.3290, 1.9030, 2.5388], + device='cuda:3'), covar=tensor([0.0758, 0.0952, 0.0593, 0.0563, 0.0578, 0.0788, 0.0737, 0.0546], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0202, 0.0184, 0.0173, 0.0179, 0.0182, 0.0154, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:28:09,574 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112722.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:28:09,598 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1435, 0.6971, 0.9537, 0.8482, 1.2391, 0.9835, 0.8405, 1.0227], + device='cuda:3'), covar=tensor([0.1529, 0.1421, 0.1898, 0.1493, 0.0938, 0.1227, 0.1787, 0.2056], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0313, 0.0351, 0.0290, 0.0328, 0.0308, 0.0302, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.4336e-05, 6.4798e-05, 7.4288e-05, 5.8574e-05, 6.7831e-05, 6.4521e-05, + 6.3251e-05, 7.9515e-05], device='cuda:3') +2023-04-27 14:28:13,013 INFO [finetune.py:976] (3/7) Epoch 20, batch 3900, loss[loss=0.2002, simple_loss=0.2678, pruned_loss=0.06625, over 4829.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2448, pruned_loss=0.05173, over 954473.01 frames. ], batch size: 39, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:28:25,016 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112745.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:28:34,453 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.220e+01 1.530e+02 1.887e+02 2.253e+02 4.348e+02, threshold=3.774e+02, percent-clipped=2.0 +2023-04-27 14:28:41,691 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112770.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:28:45,926 INFO [finetune.py:976] (3/7) Epoch 20, batch 3950, loss[loss=0.1421, simple_loss=0.208, pruned_loss=0.03809, over 4830.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2411, pruned_loss=0.05022, over 956060.08 frames. ], batch size: 30, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:29:02,236 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.5764, 1.4578, 1.4046, 1.1159, 1.4270, 1.2088, 1.7645, 1.2862], + device='cuda:3'), covar=tensor([0.3363, 0.1659, 0.4815, 0.2552, 0.1560, 0.2066, 0.1615, 0.4603], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0348, 0.0425, 0.0354, 0.0380, 0.0375, 0.0372, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:29:20,147 INFO [finetune.py:976] (3/7) Epoch 20, batch 4000, loss[loss=0.2034, simple_loss=0.2814, pruned_loss=0.06266, over 4915.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2411, pruned_loss=0.05012, over 956630.84 frames. ], batch size: 37, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:29:41,607 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.2598, 2.6538, 2.5498, 2.7242, 2.5196, 2.6370, 2.6673, 2.6450], + device='cuda:3'), covar=tensor([0.3940, 0.5875, 0.5038, 0.4652, 0.5322, 0.6496, 0.5853, 0.5010], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0373, 0.0321, 0.0334, 0.0344, 0.0394, 0.0355, 0.0326], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:29:42,041 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.556e+02 1.866e+02 2.319e+02 5.001e+02, threshold=3.732e+02, percent-clipped=1.0 +2023-04-27 14:29:48,019 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 14:29:53,756 INFO [finetune.py:976] (3/7) Epoch 20, batch 4050, loss[loss=0.1819, simple_loss=0.2559, pruned_loss=0.05398, over 4754.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2437, pruned_loss=0.0508, over 955537.12 frames. ], batch size: 27, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:30:03,649 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.7302, 4.6184, 3.0656, 5.3511, 4.7429, 4.6549, 1.8820, 4.6966], + device='cuda:3'), covar=tensor([0.1670, 0.0959, 0.3585, 0.0958, 0.4866, 0.1569, 0.6160, 0.2010], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0216, 0.0252, 0.0307, 0.0297, 0.0248, 0.0275, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:30:21,078 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6610, 1.9471, 2.0431, 2.1289, 1.9743, 2.0828, 2.1276, 2.0589], + device='cuda:3'), covar=tensor([0.4158, 0.5355, 0.4723, 0.4537, 0.5334, 0.6969, 0.5398, 0.5151], + device='cuda:3'), in_proj_covar=tensor([0.0335, 0.0371, 0.0320, 0.0334, 0.0344, 0.0393, 0.0354, 0.0325], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:30:21,667 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112918.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:30:23,586 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112921.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:30:27,669 INFO [finetune.py:976] (3/7) Epoch 20, batch 4100, loss[loss=0.174, simple_loss=0.2524, pruned_loss=0.04775, over 4747.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2456, pruned_loss=0.0505, over 954417.94 frames. ], batch size: 54, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:30:31,349 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112933.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 14:31:05,061 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8130, 2.6669, 1.7503, 2.1288, 1.3893, 1.3974, 1.8321, 1.3158], + device='cuda:3'), covar=tensor([0.1810, 0.1360, 0.1670, 0.1654, 0.2455, 0.2166, 0.1107, 0.2115], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0212, 0.0169, 0.0204, 0.0200, 0.0186, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 14:31:11,352 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.236e+02 1.611e+02 1.943e+02 2.315e+02 4.443e+02, threshold=3.885e+02, percent-clipped=3.0 +2023-04-27 14:31:25,619 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112973.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 14:31:33,458 INFO [finetune.py:976] (3/7) Epoch 20, batch 4150, loss[loss=0.1724, simple_loss=0.2321, pruned_loss=0.05631, over 4383.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2484, pruned_loss=0.05195, over 953807.30 frames. ], batch size: 19, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:31:42,877 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112982.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:32:29,691 INFO [finetune.py:976] (3/7) Epoch 20, batch 4200, loss[loss=0.1686, simple_loss=0.248, pruned_loss=0.04457, over 4846.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2489, pruned_loss=0.05178, over 954809.02 frames. ], batch size: 47, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:32:41,757 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113045.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:32:57,028 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.553e+02 1.872e+02 2.220e+02 4.301e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-04-27 14:33:18,841 INFO [finetune.py:976] (3/7) Epoch 20, batch 4250, loss[loss=0.1986, simple_loss=0.2645, pruned_loss=0.06634, over 4924.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2469, pruned_loss=0.05112, over 954206.17 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:33:40,239 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113093.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:34:04,770 INFO [finetune.py:976] (3/7) Epoch 20, batch 4300, loss[loss=0.1795, simple_loss=0.2534, pruned_loss=0.05276, over 4911.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2443, pruned_loss=0.05075, over 955104.65 frames. ], batch size: 43, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:34:27,840 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.602e+02 1.982e+02 2.267e+02 5.350e+02, threshold=3.963e+02, percent-clipped=3.0 +2023-04-27 14:34:38,150 INFO [finetune.py:976] (3/7) Epoch 20, batch 4350, loss[loss=0.207, simple_loss=0.257, pruned_loss=0.07855, over 4903.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2413, pruned_loss=0.04996, over 955966.47 frames. ], batch size: 43, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:35:06,266 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113218.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:11,653 INFO [finetune.py:976] (3/7) Epoch 20, batch 4400, loss[loss=0.1569, simple_loss=0.2361, pruned_loss=0.03886, over 4853.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2418, pruned_loss=0.05043, over 955561.02 frames. ], batch size: 31, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:35:15,425 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113233.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:16,023 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113234.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:34,719 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.671e+02 1.942e+02 2.362e+02 3.791e+02, threshold=3.884e+02, percent-clipped=0.0 +2023-04-27 14:35:38,477 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113266.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:42,679 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113273.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:35:45,062 INFO [finetune.py:976] (3/7) Epoch 20, batch 4450, loss[loss=0.176, simple_loss=0.2588, pruned_loss=0.0466, over 4854.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2446, pruned_loss=0.05068, over 955447.17 frames. ], batch size: 44, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:35:45,130 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113277.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:47,575 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113281.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:56,993 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113295.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:36:04,603 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0715, 2.4475, 2.2278, 2.2728, 1.7081, 2.0566, 2.0753, 1.6511], + device='cuda:3'), covar=tensor([0.1828, 0.1135, 0.0667, 0.1275, 0.3071, 0.1073, 0.1693, 0.2502], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0300, 0.0215, 0.0277, 0.0311, 0.0257, 0.0247, 0.0260], + device='cuda:3'), out_proj_covar=tensor([1.1475e-04, 1.1907e-04, 8.5162e-05, 1.0950e-04, 1.2617e-04, 1.0175e-04, + 1.0006e-04, 1.0301e-04], device='cuda:3') +2023-04-27 14:36:07,988 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6173, 2.7780, 2.2519, 2.4724, 2.7470, 2.3129, 3.7689, 2.0946], + device='cuda:3'), covar=tensor([0.3905, 0.1840, 0.3817, 0.3442, 0.1968, 0.2490, 0.1299, 0.3776], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0344, 0.0420, 0.0349, 0.0375, 0.0369, 0.0367, 0.0412], + device='cuda:3'), out_proj_covar=tensor([9.9779e-05, 1.0297e-04, 1.2760e-04, 1.0544e-04, 1.1184e-04, 1.1020e-04, + 1.0787e-04, 1.2443e-04], device='cuda:3') +2023-04-27 14:36:15,056 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:36:18,629 INFO [finetune.py:976] (3/7) Epoch 20, batch 4500, loss[loss=0.2095, simple_loss=0.2663, pruned_loss=0.07638, over 4886.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2458, pruned_loss=0.05089, over 956429.58 frames. ], batch size: 35, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:36:21,135 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6220, 3.5533, 2.6744, 4.1685, 3.5813, 3.5511, 1.6579, 3.5683], + device='cuda:3'), covar=tensor([0.1652, 0.1255, 0.4047, 0.1726, 0.2722, 0.1902, 0.5550, 0.2271], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0216, 0.0251, 0.0307, 0.0296, 0.0248, 0.0275, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:37:03,585 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.082e+01 1.625e+02 1.957e+02 2.322e+02 4.205e+02, threshold=3.914e+02, percent-clipped=1.0 +2023-04-27 14:37:25,450 INFO [finetune.py:976] (3/7) Epoch 20, batch 4550, loss[loss=0.2138, simple_loss=0.2859, pruned_loss=0.07086, over 4819.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2472, pruned_loss=0.05137, over 954940.19 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:37:49,679 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.35 vs. limit=5.0 +2023-04-27 14:38:02,717 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9899, 1.5511, 1.5592, 1.7357, 2.1416, 1.8048, 1.4764, 1.4350], + device='cuda:3'), covar=tensor([0.1620, 0.1505, 0.1972, 0.1171, 0.0879, 0.1480, 0.2076, 0.2344], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0314, 0.0352, 0.0290, 0.0328, 0.0308, 0.0302, 0.0375], + device='cuda:3'), out_proj_covar=tensor([6.4325e-05, 6.5051e-05, 7.4380e-05, 5.8552e-05, 6.7711e-05, 6.4590e-05, + 6.3312e-05, 7.9683e-05], device='cuda:3') +2023-04-27 14:38:21,112 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 14:38:24,622 INFO [finetune.py:976] (3/7) Epoch 20, batch 4600, loss[loss=0.1722, simple_loss=0.2577, pruned_loss=0.04336, over 4713.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2474, pruned_loss=0.05141, over 955522.16 frames. ], batch size: 54, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:39:00,866 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.917e+01 1.466e+02 1.708e+02 2.051e+02 4.040e+02, threshold=3.416e+02, percent-clipped=1.0 +2023-04-27 14:39:13,176 INFO [finetune.py:976] (3/7) Epoch 20, batch 4650, loss[loss=0.1658, simple_loss=0.2367, pruned_loss=0.0475, over 4918.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2446, pruned_loss=0.05053, over 956119.04 frames. ], batch size: 37, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:39:47,028 INFO [finetune.py:976] (3/7) Epoch 20, batch 4700, loss[loss=0.1147, simple_loss=0.186, pruned_loss=0.02166, over 4783.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2413, pruned_loss=0.04937, over 955604.75 frames. ], batch size: 26, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:40:08,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.331e+01 1.541e+02 1.809e+02 2.201e+02 3.607e+02, threshold=3.618e+02, percent-clipped=1.0 +2023-04-27 14:40:19,994 INFO [finetune.py:976] (3/7) Epoch 20, batch 4750, loss[loss=0.1843, simple_loss=0.2452, pruned_loss=0.06167, over 4861.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2398, pruned_loss=0.04922, over 955493.60 frames. ], batch size: 44, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:40:20,117 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:29,021 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113590.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:29,639 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4264, 3.4013, 2.6148, 4.0358, 3.5195, 3.4083, 1.3934, 3.4656], + device='cuda:3'), covar=tensor([0.2115, 0.1367, 0.3993, 0.2215, 0.3299, 0.2114, 0.6418, 0.3031], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0216, 0.0252, 0.0306, 0.0296, 0.0248, 0.0275, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:40:46,820 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:50,943 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1637, 2.9588, 0.9082, 1.5513, 1.5785, 2.1297, 1.6223, 0.9886], + device='cuda:3'), covar=tensor([0.1510, 0.0947, 0.1847, 0.1313, 0.1153, 0.0957, 0.1566, 0.1886], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0240, 0.0137, 0.0119, 0.0133, 0.0151, 0.0116, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 14:40:52,167 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:53,789 INFO [finetune.py:976] (3/7) Epoch 20, batch 4800, loss[loss=0.1566, simple_loss=0.2086, pruned_loss=0.05229, over 4176.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2415, pruned_loss=0.05014, over 952887.56 frames. ], batch size: 18, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:41:06,519 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113646.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:41:15,402 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.571e+02 1.898e+02 2.197e+02 3.707e+02, threshold=3.796e+02, percent-clipped=1.0 +2023-04-27 14:41:27,082 INFO [finetune.py:976] (3/7) Epoch 20, batch 4850, loss[loss=0.1867, simple_loss=0.2662, pruned_loss=0.05359, over 4766.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2455, pruned_loss=0.05109, over 953434.78 frames. ], batch size: 26, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:41:27,789 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113678.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:41:46,458 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:42:00,310 INFO [finetune.py:976] (3/7) Epoch 20, batch 4900, loss[loss=0.2244, simple_loss=0.287, pruned_loss=0.08089, over 4880.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.247, pruned_loss=0.05189, over 952945.03 frames. ], batch size: 35, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:42:19,553 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-27 14:42:27,118 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.699e+02 1.953e+02 2.350e+02 6.319e+02, threshold=3.906e+02, percent-clipped=4.0 +2023-04-27 14:42:38,975 INFO [finetune.py:976] (3/7) Epoch 20, batch 4950, loss[loss=0.1721, simple_loss=0.2481, pruned_loss=0.04801, over 4816.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2504, pruned_loss=0.05356, over 953497.33 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:43:30,855 INFO [finetune.py:976] (3/7) Epoch 20, batch 5000, loss[loss=0.1772, simple_loss=0.2375, pruned_loss=0.05847, over 4822.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2475, pruned_loss=0.05237, over 953080.40 frames. ], batch size: 30, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:44:03,789 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-27 14:44:14,003 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.541e+02 1.777e+02 2.201e+02 3.232e+02, threshold=3.554e+02, percent-clipped=0.0 +2023-04-27 14:44:29,662 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0194, 2.0745, 1.8834, 1.7241, 2.0890, 1.7409, 2.6518, 1.5617], + device='cuda:3'), covar=tensor([0.3641, 0.1706, 0.4131, 0.3120, 0.1810, 0.2344, 0.1287, 0.4169], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0347, 0.0424, 0.0352, 0.0380, 0.0373, 0.0370, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:44:32,033 INFO [finetune.py:976] (3/7) Epoch 20, batch 5050, loss[loss=0.1845, simple_loss=0.2548, pruned_loss=0.0571, over 4897.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2442, pruned_loss=0.05108, over 952304.40 frames. ], batch size: 43, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:44:52,889 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113890.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:45:21,588 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 14:45:31,588 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 14:45:34,211 INFO [finetune.py:976] (3/7) Epoch 20, batch 5100, loss[loss=0.1576, simple_loss=0.2222, pruned_loss=0.04654, over 4935.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2407, pruned_loss=0.04972, over 953945.42 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:45:42,068 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113938.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:45:50,950 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-04-27 14:45:57,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.243e+01 1.576e+02 1.837e+02 2.238e+02 4.174e+02, threshold=3.673e+02, percent-clipped=2.0 +2023-04-27 14:46:05,379 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113973.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:46:08,236 INFO [finetune.py:976] (3/7) Epoch 20, batch 5150, loss[loss=0.1248, simple_loss=0.2088, pruned_loss=0.02037, over 4742.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2395, pruned_loss=0.04937, over 954363.18 frames. ], batch size: 27, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:46:27,907 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114002.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:46:27,996 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9541, 1.6664, 2.1556, 2.3490, 2.0222, 1.8907, 1.9983, 1.9600], + device='cuda:3'), covar=tensor([0.4696, 0.6989, 0.6684, 0.5804, 0.5890, 0.8188, 0.8316, 0.9931], + device='cuda:3'), in_proj_covar=tensor([0.0431, 0.0413, 0.0508, 0.0509, 0.0460, 0.0488, 0.0497, 0.0502], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:46:43,541 INFO [finetune.py:976] (3/7) Epoch 20, batch 5200, loss[loss=0.1925, simple_loss=0.2609, pruned_loss=0.06209, over 4896.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2426, pruned_loss=0.05065, over 953215.23 frames. ], batch size: 36, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:47:06,606 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.860e+01 1.682e+02 1.942e+02 2.331e+02 4.447e+02, threshold=3.884e+02, percent-clipped=1.0 +2023-04-27 14:47:16,110 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-27 14:47:16,856 INFO [finetune.py:976] (3/7) Epoch 20, batch 5250, loss[loss=0.2078, simple_loss=0.2758, pruned_loss=0.06993, over 4886.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2454, pruned_loss=0.05141, over 954622.25 frames. ], batch size: 35, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:47:27,546 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6648, 2.1074, 1.8236, 2.0126, 1.5908, 1.7731, 1.7848, 1.3497], + device='cuda:3'), covar=tensor([0.1823, 0.0962, 0.0741, 0.1064, 0.2953, 0.1030, 0.1601, 0.2374], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0300, 0.0216, 0.0277, 0.0312, 0.0256, 0.0248, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1458e-04, 1.1928e-04, 8.5568e-05, 1.0957e-04, 1.2671e-04, 1.0171e-04, + 1.0047e-04, 1.0364e-04], device='cuda:3') +2023-04-27 14:47:50,696 INFO [finetune.py:976] (3/7) Epoch 20, batch 5300, loss[loss=0.2007, simple_loss=0.2671, pruned_loss=0.06716, over 4821.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2468, pruned_loss=0.05167, over 953932.99 frames. ], batch size: 39, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:47:50,819 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114127.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:48:10,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7094, 1.1296, 1.7410, 2.1797, 1.7675, 1.5984, 1.6670, 1.6532], + device='cuda:3'), covar=tensor([0.4472, 0.6689, 0.6024, 0.5397, 0.5971, 0.7842, 0.7792, 0.8968], + device='cuda:3'), in_proj_covar=tensor([0.0430, 0.0413, 0.0506, 0.0507, 0.0458, 0.0487, 0.0495, 0.0500], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:48:13,280 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.642e+01 1.650e+02 1.908e+02 2.204e+02 4.976e+02, threshold=3.816e+02, percent-clipped=2.0 +2023-04-27 14:48:24,239 INFO [finetune.py:976] (3/7) Epoch 20, batch 5350, loss[loss=0.1578, simple_loss=0.2274, pruned_loss=0.04413, over 4901.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2477, pruned_loss=0.05173, over 953077.86 frames. ], batch size: 36, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:48:31,434 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114188.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:48:52,734 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-04-27 14:48:58,067 INFO [finetune.py:976] (3/7) Epoch 20, batch 5400, loss[loss=0.1964, simple_loss=0.2606, pruned_loss=0.06609, over 4888.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2459, pruned_loss=0.05135, over 954597.73 frames. ], batch size: 35, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:49:33,868 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.776e+01 1.553e+02 1.821e+02 2.286e+02 4.099e+02, threshold=3.642e+02, percent-clipped=1.0 +2023-04-27 14:49:53,933 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:49:55,190 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7626, 1.6404, 1.8023, 2.0803, 2.1531, 1.6957, 1.2659, 1.8801], + device='cuda:3'), covar=tensor([0.0718, 0.1081, 0.0653, 0.0476, 0.0482, 0.0732, 0.0736, 0.0490], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0202, 0.0183, 0.0171, 0.0177, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:49:56,281 INFO [finetune.py:976] (3/7) Epoch 20, batch 5450, loss[loss=0.1613, simple_loss=0.2328, pruned_loss=0.04491, over 4837.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.244, pruned_loss=0.0507, over 955896.29 frames. ], batch size: 33, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:49:57,633 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0348, 1.9778, 2.4664, 2.6367, 1.9059, 1.6568, 2.0348, 1.1759], + device='cuda:3'), covar=tensor([0.0555, 0.0683, 0.0363, 0.0651, 0.0735, 0.1093, 0.0673, 0.0720], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0069, 0.0067, 0.0068, 0.0076, 0.0096, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 14:50:28,205 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114302.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:50:58,438 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:51:02,104 INFO [finetune.py:976] (3/7) Epoch 20, batch 5500, loss[loss=0.2186, simple_loss=0.2776, pruned_loss=0.07981, over 4923.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2409, pruned_loss=0.04975, over 957274.84 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:51:22,118 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114350.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:51:28,689 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.538e+02 1.938e+02 2.403e+02 5.552e+02, threshold=3.877e+02, percent-clipped=2.0 +2023-04-27 14:51:41,082 INFO [finetune.py:976] (3/7) Epoch 20, batch 5550, loss[loss=0.1245, simple_loss=0.2013, pruned_loss=0.02383, over 4746.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.242, pruned_loss=0.04967, over 955546.30 frames. ], batch size: 28, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:51:44,275 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3456, 3.1175, 2.6037, 2.6967, 2.0451, 2.5195, 2.7114, 2.1377], + device='cuda:3'), covar=tensor([0.2252, 0.1182, 0.0767, 0.1381, 0.3320, 0.1304, 0.1988, 0.2813], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0304, 0.0219, 0.0280, 0.0316, 0.0259, 0.0252, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1618e-04, 1.2081e-04, 8.6729e-05, 1.1092e-04, 1.2812e-04, 1.0264e-04, + 1.0208e-04, 1.0508e-04], device='cuda:3') +2023-04-27 14:52:06,604 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3574, 3.2987, 0.7466, 1.7879, 1.9303, 2.2663, 1.8495, 1.0776], + device='cuda:3'), covar=tensor([0.1373, 0.1071, 0.2130, 0.1227, 0.0980, 0.1020, 0.1476, 0.1982], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0242, 0.0137, 0.0119, 0.0133, 0.0152, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 14:52:13,272 INFO [finetune.py:976] (3/7) Epoch 20, batch 5600, loss[loss=0.1561, simple_loss=0.2318, pruned_loss=0.04015, over 4754.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2449, pruned_loss=0.05012, over 955912.57 frames. ], batch size: 26, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:52:31,154 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 14:52:32,511 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.241e+02 1.607e+02 1.931e+02 2.416e+02 4.718e+02, threshold=3.861e+02, percent-clipped=3.0 +2023-04-27 14:52:42,451 INFO [finetune.py:976] (3/7) Epoch 20, batch 5650, loss[loss=0.2062, simple_loss=0.2757, pruned_loss=0.0683, over 4838.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2458, pruned_loss=0.04987, over 955307.42 frames. ], batch size: 47, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:52:46,428 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:52:58,908 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9552, 1.4903, 1.8160, 1.7446, 1.7960, 1.5126, 0.8186, 1.4457], + device='cuda:3'), covar=tensor([0.3951, 0.3308, 0.1941, 0.2158, 0.2515, 0.2763, 0.4121, 0.2095], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0245, 0.0226, 0.0313, 0.0220, 0.0233, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 14:53:01,027 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-04-27 14:53:12,741 INFO [finetune.py:976] (3/7) Epoch 20, batch 5700, loss[loss=0.1536, simple_loss=0.2016, pruned_loss=0.05277, over 4108.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2425, pruned_loss=0.04979, over 942721.42 frames. ], batch size: 18, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:53:23,789 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.3600, 1.2104, 1.2675, 0.9421, 1.2011, 1.0458, 1.4780, 1.2686], + device='cuda:3'), covar=tensor([0.3103, 0.1599, 0.4097, 0.2148, 0.1305, 0.1701, 0.1478, 0.3949], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0348, 0.0426, 0.0352, 0.0381, 0.0373, 0.0372, 0.0418], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:53:23,856 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.43 vs. limit=5.0 +2023-04-27 14:53:39,075 INFO [finetune.py:976] (3/7) Epoch 21, batch 0, loss[loss=0.1724, simple_loss=0.2372, pruned_loss=0.05376, over 4746.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2372, pruned_loss=0.05376, over 4746.00 frames. ], batch size: 27, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:53:39,075 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 14:53:45,838 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6004, 1.6046, 3.6611, 3.4336, 3.3226, 3.4380, 3.5878, 3.2587], + device='cuda:3'), covar=tensor([0.5822, 0.4332, 0.1160, 0.1717, 0.1047, 0.1297, 0.0759, 0.1360], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0306, 0.0403, 0.0403, 0.0347, 0.0407, 0.0311, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:53:56,092 INFO [finetune.py:1010] (3/7) Epoch 21, validation: loss=0.1544, simple_loss=0.2245, pruned_loss=0.04212, over 2265189.00 frames. +2023-04-27 14:53:56,093 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 14:54:02,564 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 7.383e+01 1.462e+02 1.753e+02 2.110e+02 4.375e+02, threshold=3.507e+02, percent-clipped=1.0 +2023-04-27 14:54:24,908 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7488, 1.3057, 1.8134, 2.2527, 1.8708, 1.7492, 1.7949, 1.7349], + device='cuda:3'), covar=tensor([0.4356, 0.6270, 0.6096, 0.4831, 0.5486, 0.6867, 0.7380, 0.8713], + device='cuda:3'), in_proj_covar=tensor([0.0427, 0.0410, 0.0504, 0.0503, 0.0455, 0.0483, 0.0493, 0.0496], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:54:37,746 INFO [finetune.py:976] (3/7) Epoch 21, batch 50, loss[loss=0.1492, simple_loss=0.2328, pruned_loss=0.03277, over 4767.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2492, pruned_loss=0.05266, over 215579.17 frames. ], batch size: 28, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:55:05,781 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6860, 2.0848, 2.1234, 2.2506, 2.1458, 2.2978, 2.2476, 2.2469], + device='cuda:3'), covar=tensor([0.3561, 0.4602, 0.3808, 0.4158, 0.4683, 0.5841, 0.4587, 0.4285], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0376, 0.0322, 0.0337, 0.0348, 0.0395, 0.0358, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 14:55:27,047 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 14:55:37,893 INFO [finetune.py:976] (3/7) Epoch 21, batch 100, loss[loss=0.1768, simple_loss=0.2563, pruned_loss=0.04866, over 4835.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2397, pruned_loss=0.04919, over 378827.42 frames. ], batch size: 47, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:55:46,618 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 14:55:48,246 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.526e+02 1.765e+02 2.101e+02 5.147e+02, threshold=3.531e+02, percent-clipped=4.0 +2023-04-27 14:56:44,430 INFO [finetune.py:976] (3/7) Epoch 21, batch 150, loss[loss=0.1005, simple_loss=0.1753, pruned_loss=0.01281, over 4753.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2358, pruned_loss=0.04802, over 506444.18 frames. ], batch size: 27, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:56:57,802 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114716.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:56:58,808 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-27 14:57:12,222 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6259, 1.7286, 1.6173, 1.2772, 1.7154, 1.4287, 2.2099, 1.4354], + device='cuda:3'), covar=tensor([0.3885, 0.1887, 0.4787, 0.3091, 0.1911, 0.2412, 0.1544, 0.4727], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0350, 0.0427, 0.0354, 0.0383, 0.0374, 0.0374, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 14:57:22,715 INFO [finetune.py:976] (3/7) Epoch 21, batch 200, loss[loss=0.1702, simple_loss=0.2406, pruned_loss=0.04986, over 4826.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2345, pruned_loss=0.04761, over 604364.51 frames. ], batch size: 38, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:57:26,735 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.480e+02 1.757e+02 1.981e+02 3.579e+02, threshold=3.513e+02, percent-clipped=1.0 +2023-04-27 14:57:38,722 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114777.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:57:42,348 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114783.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:57:52,806 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 14:57:56,199 INFO [finetune.py:976] (3/7) Epoch 21, batch 250, loss[loss=0.1502, simple_loss=0.2241, pruned_loss=0.03813, over 4771.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2378, pruned_loss=0.04863, over 682719.31 frames. ], batch size: 26, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:58:15,122 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114831.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:58:30,013 INFO [finetune.py:976] (3/7) Epoch 21, batch 300, loss[loss=0.2306, simple_loss=0.3009, pruned_loss=0.08012, over 4162.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2421, pruned_loss=0.04954, over 744089.60 frames. ], batch size: 65, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:58:34,661 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.745e+01 1.667e+02 1.887e+02 2.264e+02 4.948e+02, threshold=3.774e+02, percent-clipped=2.0 +2023-04-27 14:58:42,259 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114871.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:58:48,238 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8095, 1.4246, 1.3698, 1.5570, 1.9668, 1.6269, 1.3263, 1.3355], + device='cuda:3'), covar=tensor([0.1558, 0.1456, 0.1823, 0.1291, 0.0927, 0.1319, 0.1974, 0.2019], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0315, 0.0353, 0.0291, 0.0328, 0.0311, 0.0305, 0.0375], + device='cuda:3'), out_proj_covar=tensor([6.4399e-05, 6.5230e-05, 7.4684e-05, 5.8819e-05, 6.7677e-05, 6.5177e-05, + 6.3840e-05, 7.9805e-05], device='cuda:3') +2023-04-27 14:58:52,211 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 14:59:03,088 INFO [finetune.py:976] (3/7) Epoch 21, batch 350, loss[loss=0.1817, simple_loss=0.2556, pruned_loss=0.05384, over 4786.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2447, pruned_loss=0.05052, over 791295.56 frames. ], batch size: 45, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:59:22,187 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114932.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:59:36,225 INFO [finetune.py:976] (3/7) Epoch 21, batch 400, loss[loss=0.1951, simple_loss=0.2644, pruned_loss=0.06286, over 4821.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2474, pruned_loss=0.05135, over 827012.11 frames. ], batch size: 30, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:59:40,904 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.660e+02 1.983e+02 2.165e+02 4.861e+02, threshold=3.966e+02, percent-clipped=1.0 +2023-04-27 15:00:10,132 INFO [finetune.py:976] (3/7) Epoch 21, batch 450, loss[loss=0.137, simple_loss=0.209, pruned_loss=0.03253, over 4819.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.245, pruned_loss=0.0502, over 854400.20 frames. ], batch size: 25, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:00:59,034 INFO [finetune.py:976] (3/7) Epoch 21, batch 500, loss[loss=0.1376, simple_loss=0.2166, pruned_loss=0.02928, over 4894.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2433, pruned_loss=0.05004, over 875405.46 frames. ], batch size: 32, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:01:09,614 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.974e+01 1.571e+02 1.812e+02 2.235e+02 3.288e+02, threshold=3.623e+02, percent-clipped=0.0 +2023-04-27 15:01:16,395 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 15:01:16,901 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115072.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:01:32,716 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115086.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:01:48,247 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7236, 1.5806, 1.7476, 2.0176, 2.1017, 1.5826, 1.2172, 1.8575], + device='cuda:3'), covar=tensor([0.0752, 0.1151, 0.0715, 0.0502, 0.0530, 0.0797, 0.0756, 0.0494], + device='cuda:3'), in_proj_covar=tensor([0.0190, 0.0204, 0.0185, 0.0173, 0.0179, 0.0182, 0.0153, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:01:54,520 INFO [finetune.py:976] (3/7) Epoch 21, batch 550, loss[loss=0.1614, simple_loss=0.2235, pruned_loss=0.04961, over 4891.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2413, pruned_loss=0.04955, over 895175.49 frames. ], batch size: 35, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:02:53,403 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115147.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:03:02,663 INFO [finetune.py:976] (3/7) Epoch 21, batch 600, loss[loss=0.1177, simple_loss=0.1925, pruned_loss=0.02147, over 4766.00 frames. ], tot_loss[loss=0.171, simple_loss=0.242, pruned_loss=0.05004, over 908438.59 frames. ], batch size: 27, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:03:12,532 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.494e+02 1.729e+02 2.315e+02 4.392e+02, threshold=3.458e+02, percent-clipped=4.0 +2023-04-27 15:03:27,112 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 15:03:57,839 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 15:04:02,674 INFO [finetune.py:976] (3/7) Epoch 21, batch 650, loss[loss=0.1288, simple_loss=0.199, pruned_loss=0.0293, over 4769.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2452, pruned_loss=0.05109, over 919960.65 frames. ], batch size: 26, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:04:17,775 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115227.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:04:36,594 INFO [finetune.py:976] (3/7) Epoch 21, batch 700, loss[loss=0.1508, simple_loss=0.2321, pruned_loss=0.03478, over 4815.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2467, pruned_loss=0.05083, over 929360.74 frames. ], batch size: 51, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:04:40,862 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.680e+02 1.934e+02 2.254e+02 3.960e+02, threshold=3.868e+02, percent-clipped=2.0 +2023-04-27 15:04:53,552 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-27 15:05:10,552 INFO [finetune.py:976] (3/7) Epoch 21, batch 750, loss[loss=0.1413, simple_loss=0.231, pruned_loss=0.02576, over 4751.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2475, pruned_loss=0.05056, over 936147.79 frames. ], batch size: 27, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:05:44,407 INFO [finetune.py:976] (3/7) Epoch 21, batch 800, loss[loss=0.2261, simple_loss=0.2934, pruned_loss=0.07935, over 4818.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2474, pruned_loss=0.05068, over 941039.42 frames. ], batch size: 38, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:05:48,602 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.653e+01 1.488e+02 1.739e+02 2.068e+02 3.121e+02, threshold=3.478e+02, percent-clipped=0.0 +2023-04-27 15:05:54,622 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115370.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:05:55,232 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115371.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:05:55,819 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115372.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:01,795 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5760, 1.4370, 1.8946, 1.9075, 1.4627, 1.4331, 1.5005, 0.9500], + device='cuda:3'), covar=tensor([0.0522, 0.0639, 0.0402, 0.0535, 0.0762, 0.1112, 0.0607, 0.0632], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0069, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:06:18,145 INFO [finetune.py:976] (3/7) Epoch 21, batch 850, loss[loss=0.1553, simple_loss=0.229, pruned_loss=0.04081, over 4823.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2456, pruned_loss=0.05031, over 944262.69 frames. ], batch size: 40, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:06:24,140 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 15:06:26,772 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1085, 2.0763, 1.8268, 1.7752, 2.1841, 1.6942, 2.6941, 1.6423], + device='cuda:3'), covar=tensor([0.3761, 0.2075, 0.4613, 0.3347, 0.1800, 0.2550, 0.1391, 0.4361], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0354, 0.0432, 0.0358, 0.0384, 0.0378, 0.0377, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:06:27,931 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115420.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:29,716 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.15 vs. limit=5.0 +2023-04-27 15:06:35,552 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115431.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:36,158 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:37,985 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.4827, 1.4050, 1.4366, 1.0559, 1.4170, 1.2843, 1.7389, 1.3275], + device='cuda:3'), covar=tensor([0.3259, 0.1887, 0.4773, 0.2471, 0.1441, 0.1900, 0.1621, 0.4895], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0353, 0.0432, 0.0358, 0.0384, 0.0378, 0.0377, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:06:43,132 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115442.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:07:01,732 INFO [finetune.py:976] (3/7) Epoch 21, batch 900, loss[loss=0.1112, simple_loss=0.1864, pruned_loss=0.01801, over 4934.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2425, pruned_loss=0.04921, over 945179.50 frames. ], batch size: 33, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:07:06,008 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.637e+01 1.508e+02 1.757e+02 2.120e+02 3.037e+02, threshold=3.515e+02, percent-clipped=0.0 +2023-04-27 15:07:34,517 INFO [finetune.py:976] (3/7) Epoch 21, batch 950, loss[loss=0.2059, simple_loss=0.2714, pruned_loss=0.07017, over 4808.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2415, pruned_loss=0.0495, over 948940.11 frames. ], batch size: 39, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:07:46,097 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-27 15:08:06,316 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115527.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:08:39,544 INFO [finetune.py:976] (3/7) Epoch 21, batch 1000, loss[loss=0.2008, simple_loss=0.2749, pruned_loss=0.06332, over 4712.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2437, pruned_loss=0.05027, over 950059.41 frames. ], batch size: 59, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:08:49,079 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.678e+02 2.025e+02 2.586e+02 4.511e+02, threshold=4.050e+02, percent-clipped=4.0 +2023-04-27 15:09:08,298 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115575.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:09:45,484 INFO [finetune.py:976] (3/7) Epoch 21, batch 1050, loss[loss=0.1807, simple_loss=0.246, pruned_loss=0.05771, over 4761.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2466, pruned_loss=0.0512, over 950643.02 frames. ], batch size: 26, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:10:18,956 INFO [finetune.py:976] (3/7) Epoch 21, batch 1100, loss[loss=0.148, simple_loss=0.2212, pruned_loss=0.03739, over 4784.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2466, pruned_loss=0.05111, over 950416.20 frames. ], batch size: 29, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:10:22,339 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 15:10:23,752 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.551e+02 1.797e+02 2.277e+02 4.571e+02, threshold=3.594e+02, percent-clipped=1.0 +2023-04-27 15:10:52,442 INFO [finetune.py:976] (3/7) Epoch 21, batch 1150, loss[loss=0.1819, simple_loss=0.2569, pruned_loss=0.05339, over 4839.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2476, pruned_loss=0.05092, over 950019.38 frames. ], batch size: 44, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:11:07,916 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115726.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:11:08,528 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115727.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:11:18,224 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115742.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:11:18,880 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9731, 2.3578, 2.1320, 2.8327, 3.0028, 2.6144, 2.5235, 2.0519], + device='cuda:3'), covar=tensor([0.1252, 0.1446, 0.1824, 0.1217, 0.0955, 0.1282, 0.1693, 0.2136], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0311, 0.0350, 0.0288, 0.0327, 0.0309, 0.0303, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.3852e-05, 6.4409e-05, 7.3832e-05, 5.8180e-05, 6.7635e-05, 6.4855e-05, + 6.3540e-05, 7.9163e-05], device='cuda:3') +2023-04-27 15:11:19,538 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 15:11:25,944 INFO [finetune.py:976] (3/7) Epoch 21, batch 1200, loss[loss=0.1485, simple_loss=0.2232, pruned_loss=0.03687, over 4712.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2464, pruned_loss=0.05079, over 951738.01 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:11:31,122 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.519e+02 1.757e+02 2.035e+02 5.645e+02, threshold=3.514e+02, percent-clipped=1.0 +2023-04-27 15:11:50,226 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115790.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:11:53,853 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 15:11:56,090 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2463, 1.4937, 1.3528, 1.7179, 1.5875, 1.7040, 1.3447, 2.9234], + device='cuda:3'), covar=tensor([0.0600, 0.0753, 0.0749, 0.1149, 0.0577, 0.0540, 0.0712, 0.0175], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 15:11:59,754 INFO [finetune.py:976] (3/7) Epoch 21, batch 1250, loss[loss=0.1462, simple_loss=0.2128, pruned_loss=0.03976, over 4826.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2439, pruned_loss=0.05008, over 949865.11 frames. ], batch size: 41, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:12:55,528 INFO [finetune.py:976] (3/7) Epoch 21, batch 1300, loss[loss=0.1751, simple_loss=0.2463, pruned_loss=0.05191, over 4805.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2417, pruned_loss=0.04964, over 950527.52 frames. ], batch size: 51, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:12:59,786 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.963e+01 1.535e+02 1.763e+02 2.248e+02 3.829e+02, threshold=3.527e+02, percent-clipped=1.0 +2023-04-27 15:13:21,240 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115892.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 15:13:29,446 INFO [finetune.py:976] (3/7) Epoch 21, batch 1350, loss[loss=0.1725, simple_loss=0.2555, pruned_loss=0.04479, over 4765.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2417, pruned_loss=0.05002, over 951019.23 frames. ], batch size: 28, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:14:23,450 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5489, 0.6826, 1.4458, 1.9206, 1.6283, 1.4445, 1.4631, 1.4779], + device='cuda:3'), covar=tensor([0.3854, 0.5971, 0.5206, 0.4903, 0.4877, 0.6356, 0.6369, 0.7542], + device='cuda:3'), in_proj_covar=tensor([0.0429, 0.0413, 0.0506, 0.0506, 0.0458, 0.0487, 0.0496, 0.0500], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:14:34,405 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115953.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 15:14:34,902 INFO [finetune.py:976] (3/7) Epoch 21, batch 1400, loss[loss=0.1647, simple_loss=0.246, pruned_loss=0.04167, over 4896.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2447, pruned_loss=0.05068, over 952966.82 frames. ], batch size: 35, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:14:44,906 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.645e+02 1.882e+02 2.243e+02 4.994e+02, threshold=3.764e+02, percent-clipped=1.0 +2023-04-27 15:15:10,045 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1296, 2.4753, 0.9114, 1.4001, 1.4347, 1.9145, 1.5102, 0.8746], + device='cuda:3'), covar=tensor([0.1504, 0.1058, 0.1796, 0.1385, 0.1222, 0.0872, 0.1640, 0.1925], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0239, 0.0136, 0.0119, 0.0132, 0.0151, 0.0116, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:15:14,228 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1567, 2.5154, 1.0738, 1.3744, 2.0253, 1.2952, 3.5337, 1.8109], + device='cuda:3'), covar=tensor([0.0651, 0.0835, 0.0803, 0.1193, 0.0506, 0.0964, 0.0280, 0.0650], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 15:15:26,317 INFO [finetune.py:976] (3/7) Epoch 21, batch 1450, loss[loss=0.173, simple_loss=0.2392, pruned_loss=0.05337, over 4931.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2461, pruned_loss=0.05108, over 952490.91 frames. ], batch size: 33, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:15:33,446 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116014.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:15:42,220 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116026.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:15:42,839 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:00,015 INFO [finetune.py:976] (3/7) Epoch 21, batch 1500, loss[loss=0.1744, simple_loss=0.2487, pruned_loss=0.05009, over 4890.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2485, pruned_loss=0.05223, over 952333.41 frames. ], batch size: 43, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:16:05,187 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.610e+02 1.919e+02 2.360e+02 3.995e+02, threshold=3.837e+02, percent-clipped=2.0 +2023-04-27 15:16:05,880 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.6885, 4.6317, 2.9498, 5.3445, 4.7034, 4.5761, 2.0143, 4.5765], + device='cuda:3'), covar=tensor([0.1589, 0.0947, 0.3662, 0.0972, 0.2586, 0.1515, 0.5769, 0.2312], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0216, 0.0253, 0.0309, 0.0298, 0.0247, 0.0275, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:16:13,679 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116074.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:14,788 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116075.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:14,854 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116075.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:16:33,606 INFO [finetune.py:976] (3/7) Epoch 21, batch 1550, loss[loss=0.1436, simple_loss=0.2229, pruned_loss=0.03217, over 4906.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2477, pruned_loss=0.05172, over 954790.27 frames. ], batch size: 37, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:16:34,344 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116105.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:41,966 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:52,267 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-04-27 15:16:54,133 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-04-27 15:17:06,682 INFO [finetune.py:976] (3/7) Epoch 21, batch 1600, loss[loss=0.1549, simple_loss=0.2275, pruned_loss=0.04114, over 4145.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2455, pruned_loss=0.05088, over 953629.59 frames. ], batch size: 18, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:17:10,952 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.579e+02 1.851e+02 2.317e+02 5.378e+02, threshold=3.702e+02, percent-clipped=3.0 +2023-04-27 15:17:14,640 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:17:21,309 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116177.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:17:39,784 INFO [finetune.py:976] (3/7) Epoch 21, batch 1650, loss[loss=0.1689, simple_loss=0.2351, pruned_loss=0.05131, over 4750.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2429, pruned_loss=0.05003, over 954197.92 frames. ], batch size: 27, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:18:14,847 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116248.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:18:16,162 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-04-27 15:18:18,417 INFO [finetune.py:976] (3/7) Epoch 21, batch 1700, loss[loss=0.2084, simple_loss=0.2809, pruned_loss=0.06789, over 4738.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2415, pruned_loss=0.04988, over 956074.63 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:18:28,119 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.499e+02 1.782e+02 2.142e+02 3.522e+02, threshold=3.563e+02, percent-clipped=0.0 +2023-04-27 15:19:31,551 INFO [finetune.py:976] (3/7) Epoch 21, batch 1750, loss[loss=0.1838, simple_loss=0.2621, pruned_loss=0.05278, over 4738.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2422, pruned_loss=0.04981, over 957262.59 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:19:55,833 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116323.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:20:33,794 INFO [finetune.py:976] (3/7) Epoch 21, batch 1800, loss[loss=0.2035, simple_loss=0.2703, pruned_loss=0.06833, over 4810.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2458, pruned_loss=0.05095, over 956826.62 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:20:37,558 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4381, 2.1216, 2.3808, 2.8147, 2.7202, 2.2226, 1.7587, 2.3760], + device='cuda:3'), covar=tensor([0.0729, 0.0908, 0.0613, 0.0539, 0.0648, 0.0805, 0.0820, 0.0550], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0200, 0.0183, 0.0172, 0.0177, 0.0180, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:20:38,070 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.681e+02 1.987e+02 2.405e+02 5.932e+02, threshold=3.974e+02, percent-clipped=5.0 +2023-04-27 15:20:43,620 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116370.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 15:20:53,285 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:21:07,564 INFO [finetune.py:976] (3/7) Epoch 21, batch 1850, loss[loss=0.1738, simple_loss=0.2471, pruned_loss=0.05021, over 4823.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.246, pruned_loss=0.05074, over 956065.93 frames. ], batch size: 33, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:21:34,677 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.6669, 4.4787, 3.0451, 5.2553, 4.5105, 4.6185, 1.7628, 4.5071], + device='cuda:3'), covar=tensor([0.1565, 0.1004, 0.3414, 0.0829, 0.2037, 0.1396, 0.5861, 0.2156], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0217, 0.0254, 0.0309, 0.0300, 0.0248, 0.0277, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:21:35,319 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116446.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:21:40,091 INFO [finetune.py:976] (3/7) Epoch 21, batch 1900, loss[loss=0.12, simple_loss=0.195, pruned_loss=0.02253, over 4696.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2479, pruned_loss=0.05144, over 957153.74 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:21:45,215 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.938e+01 1.604e+02 1.932e+02 2.429e+02 3.655e+02, threshold=3.864e+02, percent-clipped=0.0 +2023-04-27 15:21:45,316 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116461.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:21:52,109 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116472.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:22:13,561 INFO [finetune.py:976] (3/7) Epoch 21, batch 1950, loss[loss=0.1488, simple_loss=0.2263, pruned_loss=0.03565, over 4897.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2458, pruned_loss=0.05029, over 955689.26 frames. ], batch size: 36, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:22:14,295 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116505.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:22:15,452 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3358, 2.0592, 2.5102, 2.8481, 2.2941, 2.2422, 2.3926, 2.3464], + device='cuda:3'), covar=tensor([0.4755, 0.6912, 0.7407, 0.5298, 0.6331, 0.8654, 0.9171, 0.8449], + device='cuda:3'), in_proj_covar=tensor([0.0431, 0.0413, 0.0509, 0.0508, 0.0459, 0.0489, 0.0498, 0.0502], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:22:16,021 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116507.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:22:43,312 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116548.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 15:22:46,841 INFO [finetune.py:976] (3/7) Epoch 21, batch 2000, loss[loss=0.1369, simple_loss=0.2031, pruned_loss=0.03539, over 4685.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2421, pruned_loss=0.04891, over 955144.73 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:22:51,557 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.496e+02 1.815e+02 2.157e+02 3.594e+02, threshold=3.630e+02, percent-clipped=0.0 +2023-04-27 15:22:55,222 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116566.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:23:02,489 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7765, 4.0424, 0.7231, 2.0074, 2.2460, 2.6738, 2.3406, 0.9713], + device='cuda:3'), covar=tensor([0.1412, 0.1107, 0.2323, 0.1380, 0.1129, 0.1160, 0.1595, 0.2299], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0239, 0.0136, 0.0119, 0.0132, 0.0151, 0.0116, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:23:06,596 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2398, 1.7274, 2.0421, 2.4574, 2.0848, 1.6167, 1.2494, 1.8901], + device='cuda:3'), covar=tensor([0.3392, 0.3140, 0.1745, 0.2413, 0.2531, 0.2702, 0.4239, 0.2074], + device='cuda:3'), in_proj_covar=tensor([0.0289, 0.0242, 0.0224, 0.0311, 0.0216, 0.0230, 0.0225, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 15:23:11,438 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9142, 1.6330, 1.8221, 2.2262, 2.2730, 1.7501, 1.4580, 1.9170], + device='cuda:3'), covar=tensor([0.0767, 0.1165, 0.0817, 0.0601, 0.0598, 0.0870, 0.0843, 0.0613], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0201, 0.0184, 0.0173, 0.0177, 0.0180, 0.0152, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:23:14,413 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116596.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:23:20,569 INFO [finetune.py:976] (3/7) Epoch 21, batch 2050, loss[loss=0.174, simple_loss=0.2353, pruned_loss=0.05637, over 4834.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.239, pruned_loss=0.0479, over 956877.93 frames. ], batch size: 33, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:23:21,875 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8589, 2.1770, 1.1590, 1.5960, 2.2076, 1.7369, 1.6877, 1.6751], + device='cuda:3'), covar=tensor([0.0468, 0.0328, 0.0300, 0.0525, 0.0245, 0.0478, 0.0471, 0.0547], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 15:23:30,136 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116618.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:23:59,066 INFO [finetune.py:976] (3/7) Epoch 21, batch 2100, loss[loss=0.1871, simple_loss=0.247, pruned_loss=0.06359, over 4856.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2387, pruned_loss=0.04794, over 956538.11 frames. ], batch size: 44, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:24:03,921 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.802e+01 1.616e+02 1.834e+02 2.363e+02 4.673e+02, threshold=3.668e+02, percent-clipped=1.0 +2023-04-27 15:24:21,188 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116670.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:24:32,901 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:24:32,955 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:25:00,267 INFO [finetune.py:976] (3/7) Epoch 21, batch 2150, loss[loss=0.1364, simple_loss=0.212, pruned_loss=0.03046, over 4683.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2428, pruned_loss=0.04948, over 955691.04 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:25:05,382 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:25:21,300 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:25:28,189 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3853, 1.3759, 1.6386, 1.6445, 1.3756, 1.2048, 1.4584, 0.9550], + device='cuda:3'), covar=tensor([0.0539, 0.0500, 0.0403, 0.0529, 0.0610, 0.0811, 0.0490, 0.0579], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0067, 0.0066, 0.0067, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:26:05,622 INFO [finetune.py:976] (3/7) Epoch 21, batch 2200, loss[loss=0.1141, simple_loss=0.1936, pruned_loss=0.01734, over 4711.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2452, pruned_loss=0.05048, over 951054.23 frames. ], batch size: 23, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:26:10,849 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.212e+01 1.600e+02 1.964e+02 2.414e+02 3.602e+02, threshold=3.928e+02, percent-clipped=0.0 +2023-04-27 15:26:10,976 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116761.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:13,351 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 15:26:13,893 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:18,740 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116772.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:38,512 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116802.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 15:26:38,543 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116802.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:39,689 INFO [finetune.py:976] (3/7) Epoch 21, batch 2250, loss[loss=0.1828, simple_loss=0.2582, pruned_loss=0.05367, over 4811.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2471, pruned_loss=0.05168, over 951225.23 frames. ], batch size: 39, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:26:42,837 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116809.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:50,606 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:27:04,068 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-04-27 15:27:04,485 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.0955, 2.3610, 2.2930, 2.4550, 2.1386, 2.3835, 2.3160, 2.2915], + device='cuda:3'), covar=tensor([0.3948, 0.6043, 0.5099, 0.4447, 0.5938, 0.7086, 0.6412, 0.5583], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0375, 0.0323, 0.0336, 0.0347, 0.0395, 0.0358, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:27:13,183 INFO [finetune.py:976] (3/7) Epoch 21, batch 2300, loss[loss=0.165, simple_loss=0.2402, pruned_loss=0.04487, over 4834.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2464, pruned_loss=0.05092, over 951917.29 frames. ], batch size: 30, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:27:17,445 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.557e+02 1.911e+02 2.274e+02 3.749e+02, threshold=3.822e+02, percent-clipped=0.0 +2023-04-27 15:27:17,533 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116861.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:27:18,830 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116863.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:27:24,782 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 15:27:37,266 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-27 15:27:46,965 INFO [finetune.py:976] (3/7) Epoch 21, batch 2350, loss[loss=0.145, simple_loss=0.2216, pruned_loss=0.03421, over 4762.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2443, pruned_loss=0.05024, over 951294.70 frames. ], batch size: 27, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:27:52,050 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-04-27 15:28:20,895 INFO [finetune.py:976] (3/7) Epoch 21, batch 2400, loss[loss=0.1564, simple_loss=0.2261, pruned_loss=0.04333, over 4899.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2416, pruned_loss=0.04947, over 951628.59 frames. ], batch size: 35, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:28:25,121 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.594e+02 1.806e+02 2.173e+02 4.519e+02, threshold=3.612e+02, percent-clipped=1.0 +2023-04-27 15:28:34,736 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:28:37,805 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116979.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:28:54,617 INFO [finetune.py:976] (3/7) Epoch 21, batch 2450, loss[loss=0.1906, simple_loss=0.2671, pruned_loss=0.0571, over 4815.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2386, pruned_loss=0.04883, over 951639.40 frames. ], batch size: 40, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:28:55,783 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-27 15:29:03,176 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6378, 3.6454, 2.6091, 4.2244, 3.7125, 3.5867, 1.5236, 3.6685], + device='cuda:3'), covar=tensor([0.1783, 0.1251, 0.3162, 0.1749, 0.3175, 0.1953, 0.5785, 0.2526], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0215, 0.0253, 0.0307, 0.0298, 0.0247, 0.0276, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:29:10,687 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:29:28,082 INFO [finetune.py:976] (3/7) Epoch 21, batch 2500, loss[loss=0.1816, simple_loss=0.2609, pruned_loss=0.05113, over 4795.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2399, pruned_loss=0.04959, over 950710.34 frames. ], batch size: 45, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:29:32,764 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117060.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:29:33,286 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.592e+02 1.814e+02 2.143e+02 3.626e+02, threshold=3.628e+02, percent-clipped=1.0 +2023-04-27 15:30:29,012 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117102.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:30:29,865 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 15:30:30,116 INFO [finetune.py:976] (3/7) Epoch 21, batch 2550, loss[loss=0.1788, simple_loss=0.2599, pruned_loss=0.04882, over 4803.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2438, pruned_loss=0.05023, over 951036.96 frames. ], batch size: 41, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:31:00,897 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6594, 1.5223, 1.7173, 1.9679, 2.0583, 1.6185, 1.2317, 1.7803], + device='cuda:3'), covar=tensor([0.0830, 0.1154, 0.0745, 0.0646, 0.0577, 0.0815, 0.0835, 0.0590], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0202, 0.0184, 0.0173, 0.0177, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:31:33,708 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117150.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:31:41,897 INFO [finetune.py:976] (3/7) Epoch 21, batch 2600, loss[loss=0.18, simple_loss=0.2609, pruned_loss=0.04949, over 4818.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2456, pruned_loss=0.05082, over 950236.75 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:31:44,416 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:31:52,555 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.651e+02 1.953e+02 2.341e+02 4.282e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-04-27 15:31:52,671 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117161.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:32:09,247 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 15:32:27,548 INFO [finetune.py:976] (3/7) Epoch 21, batch 2650, loss[loss=0.1481, simple_loss=0.2396, pruned_loss=0.02834, over 4816.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2462, pruned_loss=0.05053, over 952077.86 frames. ], batch size: 39, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:32:30,679 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117209.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:32:39,152 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=117221.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:33:01,272 INFO [finetune.py:976] (3/7) Epoch 21, batch 2700, loss[loss=0.1866, simple_loss=0.2601, pruned_loss=0.05658, over 4835.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2453, pruned_loss=0.04991, over 954686.20 frames. ], batch size: 47, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:33:06,002 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.896e+01 1.368e+02 1.680e+02 2.087e+02 5.894e+02, threshold=3.359e+02, percent-clipped=1.0 +2023-04-27 15:33:15,093 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117274.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:33:19,990 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117282.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:33:31,670 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2521, 1.4723, 1.6667, 1.8156, 1.7570, 1.8544, 1.7720, 1.8106], + device='cuda:3'), covar=tensor([0.3556, 0.4690, 0.4150, 0.4101, 0.5054, 0.6462, 0.4523, 0.4265], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0371, 0.0321, 0.0334, 0.0345, 0.0393, 0.0354, 0.0327], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:33:35,175 INFO [finetune.py:976] (3/7) Epoch 21, batch 2750, loss[loss=0.1619, simple_loss=0.2347, pruned_loss=0.0445, over 4898.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2429, pruned_loss=0.04924, over 956486.34 frames. ], batch size: 32, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:33:44,952 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9632, 1.6740, 1.8649, 2.2768, 2.2802, 1.9374, 1.5593, 2.0296], + device='cuda:3'), covar=tensor([0.0760, 0.1059, 0.0659, 0.0508, 0.0549, 0.0754, 0.0768, 0.0516], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0202, 0.0185, 0.0174, 0.0177, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:33:47,710 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117322.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:33:52,064 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=117329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:34:08,696 INFO [finetune.py:976] (3/7) Epoch 21, batch 2800, loss[loss=0.1345, simple_loss=0.2084, pruned_loss=0.03026, over 4767.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2407, pruned_loss=0.04868, over 956555.76 frames. ], batch size: 28, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:34:12,917 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117360.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:34:13,413 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.550e+02 1.821e+02 2.376e+02 5.325e+02, threshold=3.642e+02, percent-clipped=3.0 +2023-04-27 15:34:23,116 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 15:34:31,551 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:34:33,418 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6073, 2.5192, 1.9319, 2.3227, 2.4737, 2.0905, 3.2112, 1.8382], + device='cuda:3'), covar=tensor([0.3572, 0.2081, 0.4548, 0.3214, 0.1987, 0.2654, 0.1612, 0.4296], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0345, 0.0422, 0.0350, 0.0379, 0.0372, 0.0367, 0.0417], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:34:39,251 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3948, 2.8976, 0.9861, 1.6709, 1.6702, 2.1916, 1.6771, 0.9777], + device='cuda:3'), covar=tensor([0.1192, 0.1011, 0.1580, 0.1125, 0.1024, 0.0843, 0.1359, 0.1757], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0238, 0.0135, 0.0118, 0.0132, 0.0151, 0.0115, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:34:41,632 INFO [finetune.py:976] (3/7) Epoch 21, batch 2850, loss[loss=0.1491, simple_loss=0.2244, pruned_loss=0.03689, over 4817.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2394, pruned_loss=0.04872, over 956900.75 frames. ], batch size: 38, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:34:41,738 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6479, 1.5242, 0.8443, 1.3389, 1.6882, 1.4987, 1.4059, 1.4573], + device='cuda:3'), covar=tensor([0.0448, 0.0334, 0.0349, 0.0487, 0.0276, 0.0473, 0.0441, 0.0511], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0051], + device='cuda:3') +2023-04-27 15:34:44,560 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117408.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:35:14,768 INFO [finetune.py:976] (3/7) Epoch 21, batch 2900, loss[loss=0.1211, simple_loss=0.2003, pruned_loss=0.02094, over 4721.00 frames. ], tot_loss[loss=0.172, simple_loss=0.243, pruned_loss=0.05047, over 955364.81 frames. ], batch size: 23, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:35:17,804 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117458.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:35:19,564 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.585e+02 1.858e+02 2.394e+02 5.975e+02, threshold=3.717e+02, percent-clipped=5.0 +2023-04-27 15:36:04,140 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8525, 1.6050, 1.7634, 2.1454, 2.1245, 1.7016, 1.3974, 1.8769], + device='cuda:3'), covar=tensor([0.0779, 0.1159, 0.0760, 0.0540, 0.0574, 0.0819, 0.0775, 0.0553], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0203, 0.0185, 0.0175, 0.0178, 0.0182, 0.0153, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:36:16,932 INFO [finetune.py:976] (3/7) Epoch 21, batch 2950, loss[loss=0.1775, simple_loss=0.2515, pruned_loss=0.05174, over 4826.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2459, pruned_loss=0.0512, over 955497.24 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:36:18,222 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117506.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:37:23,432 INFO [finetune.py:976] (3/7) Epoch 21, batch 3000, loss[loss=0.166, simple_loss=0.2355, pruned_loss=0.0483, over 4923.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.246, pruned_loss=0.05133, over 953907.11 frames. ], batch size: 42, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:37:23,432 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 15:37:31,923 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4280, 3.4544, 2.5335, 3.8805, 3.4935, 3.4975, 1.5010, 3.4552], + device='cuda:3'), covar=tensor([0.1731, 0.1283, 0.2583, 0.1942, 0.2328, 0.1783, 0.5284, 0.2120], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0213, 0.0249, 0.0303, 0.0294, 0.0244, 0.0271, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:37:43,671 INFO [finetune.py:1010] (3/7) Epoch 21, validation: loss=0.1531, simple_loss=0.2228, pruned_loss=0.04164, over 2265189.00 frames. +2023-04-27 15:37:43,671 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 15:37:54,335 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.590e+02 1.926e+02 2.493e+02 6.945e+02, threshold=3.852e+02, percent-clipped=2.0 +2023-04-27 15:38:15,520 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:38:49,964 INFO [finetune.py:976] (3/7) Epoch 21, batch 3050, loss[loss=0.178, simple_loss=0.2586, pruned_loss=0.04869, over 4893.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2471, pruned_loss=0.05135, over 954291.81 frames. ], batch size: 43, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:39:16,595 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9057, 1.3317, 4.9266, 4.6231, 4.2859, 4.6463, 4.3682, 4.2936], + device='cuda:3'), covar=tensor([0.6677, 0.5917, 0.0944, 0.1800, 0.1110, 0.1528, 0.1591, 0.1645], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0306, 0.0403, 0.0403, 0.0346, 0.0408, 0.0310, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:39:28,405 INFO [finetune.py:976] (3/7) Epoch 21, batch 3100, loss[loss=0.1691, simple_loss=0.2436, pruned_loss=0.04729, over 4813.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2448, pruned_loss=0.05051, over 954191.12 frames. ], batch size: 41, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:39:33,614 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.782e+01 1.629e+02 1.831e+02 2.140e+02 4.594e+02, threshold=3.661e+02, percent-clipped=1.0 +2023-04-27 15:39:49,469 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:39:51,871 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8694, 1.2004, 1.4597, 1.5769, 1.5968, 1.6546, 1.5230, 1.5056], + device='cuda:3'), covar=tensor([0.3101, 0.3693, 0.3397, 0.3311, 0.4175, 0.5383, 0.3789, 0.3633], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0375, 0.0324, 0.0337, 0.0347, 0.0395, 0.0357, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:40:01,889 INFO [finetune.py:976] (3/7) Epoch 21, batch 3150, loss[loss=0.1934, simple_loss=0.2504, pruned_loss=0.06822, over 4822.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2417, pruned_loss=0.04946, over 953569.25 frames. ], batch size: 30, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:40:10,249 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9143, 1.1879, 4.9756, 4.7080, 4.3431, 4.6535, 4.4328, 4.3877], + device='cuda:3'), covar=tensor([0.7207, 0.6681, 0.1231, 0.1923, 0.1080, 0.1308, 0.1520, 0.1685], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0308, 0.0405, 0.0406, 0.0349, 0.0411, 0.0312, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:40:34,883 INFO [finetune.py:976] (3/7) Epoch 21, batch 3200, loss[loss=0.1465, simple_loss=0.22, pruned_loss=0.03649, over 4834.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.239, pruned_loss=0.04871, over 954010.69 frames. ], batch size: 47, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:40:40,630 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.488e+02 1.740e+02 2.134e+02 4.816e+02, threshold=3.479e+02, percent-clipped=1.0 +2023-04-27 15:41:22,274 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6591, 3.6856, 2.7167, 4.2383, 3.7327, 3.6687, 1.6049, 3.7184], + device='cuda:3'), covar=tensor([0.1692, 0.1240, 0.2922, 0.1799, 0.4210, 0.1652, 0.5830, 0.2081], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0215, 0.0250, 0.0304, 0.0296, 0.0245, 0.0273, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 15:41:31,202 INFO [finetune.py:976] (3/7) Epoch 21, batch 3250, loss[loss=0.1489, simple_loss=0.2267, pruned_loss=0.03557, over 4783.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2401, pruned_loss=0.0494, over 954821.82 frames. ], batch size: 29, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:42:33,311 INFO [finetune.py:976] (3/7) Epoch 21, batch 3300, loss[loss=0.2247, simple_loss=0.2977, pruned_loss=0.07585, over 4828.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2434, pruned_loss=0.05068, over 952853.40 frames. ], batch size: 49, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:42:45,022 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.168e+02 1.642e+02 1.977e+02 2.287e+02 4.163e+02, threshold=3.954e+02, percent-clipped=4.0 +2023-04-27 15:43:01,382 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117877.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:43:18,017 INFO [finetune.py:976] (3/7) Epoch 21, batch 3350, loss[loss=0.1659, simple_loss=0.2458, pruned_loss=0.04296, over 4857.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.245, pruned_loss=0.05089, over 953889.69 frames. ], batch size: 44, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:43:37,870 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:43:54,529 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 15:44:01,273 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-04-27 15:44:02,873 INFO [finetune.py:976] (3/7) Epoch 21, batch 3400, loss[loss=0.1968, simple_loss=0.2677, pruned_loss=0.063, over 4922.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2471, pruned_loss=0.05171, over 954276.34 frames. ], batch size: 42, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:44:13,576 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.581e+02 1.978e+02 2.353e+02 4.308e+02, threshold=3.955e+02, percent-clipped=3.0 +2023-04-27 15:44:44,657 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117985.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:44:44,864 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 15:44:57,989 INFO [finetune.py:976] (3/7) Epoch 21, batch 3450, loss[loss=0.1366, simple_loss=0.2063, pruned_loss=0.03344, over 4760.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2477, pruned_loss=0.05146, over 955907.90 frames. ], batch size: 27, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:45:17,643 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118033.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:45:22,457 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:45:31,329 INFO [finetune.py:976] (3/7) Epoch 21, batch 3500, loss[loss=0.1952, simple_loss=0.2674, pruned_loss=0.06152, over 4818.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2454, pruned_loss=0.05065, over 956477.95 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:45:36,183 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.106e+01 1.494e+02 1.798e+02 2.113e+02 5.768e+02, threshold=3.596e+02, percent-clipped=1.0 +2023-04-27 15:45:42,465 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 15:46:03,422 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118101.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:46:05,140 INFO [finetune.py:976] (3/7) Epoch 21, batch 3550, loss[loss=0.1797, simple_loss=0.2487, pruned_loss=0.05535, over 4822.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2419, pruned_loss=0.04942, over 957130.02 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:46:39,044 INFO [finetune.py:976] (3/7) Epoch 21, batch 3600, loss[loss=0.1535, simple_loss=0.2278, pruned_loss=0.03962, over 4823.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2393, pruned_loss=0.04854, over 956021.77 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:46:43,924 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.483e+02 1.865e+02 2.338e+02 3.870e+02, threshold=3.730e+02, percent-clipped=2.0 +2023-04-27 15:46:47,649 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118168.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:47:39,451 INFO [finetune.py:976] (3/7) Epoch 21, batch 3650, loss[loss=0.1681, simple_loss=0.2413, pruned_loss=0.04744, over 4787.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2419, pruned_loss=0.04978, over 954984.27 frames. ], batch size: 29, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:48:09,970 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118229.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:48:43,767 INFO [finetune.py:976] (3/7) Epoch 21, batch 3700, loss[loss=0.175, simple_loss=0.2535, pruned_loss=0.04829, over 4849.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2455, pruned_loss=0.05109, over 953241.71 frames. ], batch size: 44, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:48:54,005 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.556e+02 1.839e+02 2.205e+02 3.757e+02, threshold=3.678e+02, percent-clipped=1.0 +2023-04-27 15:48:58,378 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118269.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:49:16,503 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:49:27,414 INFO [finetune.py:976] (3/7) Epoch 21, batch 3750, loss[loss=0.1267, simple_loss=0.2068, pruned_loss=0.02335, over 4774.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.247, pruned_loss=0.05198, over 953895.77 frames. ], batch size: 26, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:49:40,288 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0171, 1.0235, 1.1542, 1.1392, 1.0101, 0.9356, 0.9997, 0.4391], + device='cuda:3'), covar=tensor([0.0461, 0.0458, 0.0399, 0.0427, 0.0655, 0.0986, 0.0412, 0.0586], + device='cuda:3'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:49:43,201 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:49:43,829 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118330.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:50:10,270 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118350.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:50:12,621 INFO [finetune.py:976] (3/7) Epoch 21, batch 3800, loss[loss=0.18, simple_loss=0.2524, pruned_loss=0.05379, over 4899.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2481, pruned_loss=0.05229, over 953563.44 frames. ], batch size: 36, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:50:23,125 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.645e+02 2.008e+02 2.325e+02 4.479e+02, threshold=4.015e+02, percent-clipped=4.0 +2023-04-27 15:50:24,178 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 15:50:52,585 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:50:53,797 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1705, 1.4702, 1.3367, 1.7040, 1.6145, 1.6714, 1.3190, 3.1043], + device='cuda:3'), covar=tensor([0.0641, 0.0818, 0.0821, 0.1235, 0.0616, 0.0506, 0.0748, 0.0168], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 15:50:56,203 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:50:56,341 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-04-27 15:51:02,178 INFO [finetune.py:976] (3/7) Epoch 21, batch 3850, loss[loss=0.1822, simple_loss=0.2584, pruned_loss=0.05298, over 4815.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2467, pruned_loss=0.05151, over 954597.30 frames. ], batch size: 39, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:51:35,148 INFO [finetune.py:976] (3/7) Epoch 21, batch 3900, loss[loss=0.1519, simple_loss=0.2235, pruned_loss=0.04016, over 4888.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2432, pruned_loss=0.05022, over 954084.75 frames. ], batch size: 32, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:51:40,427 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.638e+02 1.891e+02 2.265e+02 7.777e+02, threshold=3.782e+02, percent-clipped=1.0 +2023-04-27 15:51:56,295 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2538, 4.4298, 0.8556, 2.3285, 2.6450, 2.8040, 2.6100, 1.0523], + device='cuda:3'), covar=tensor([0.1118, 0.0832, 0.2064, 0.1143, 0.0871, 0.1076, 0.1309, 0.1965], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0237, 0.0135, 0.0119, 0.0132, 0.0151, 0.0115, 0.0117], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:52:07,427 INFO [finetune.py:976] (3/7) Epoch 21, batch 3950, loss[loss=0.1491, simple_loss=0.2258, pruned_loss=0.03617, over 4786.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2416, pruned_loss=0.05014, over 954847.02 frames. ], batch size: 28, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:52:10,917 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8039, 2.7229, 2.9633, 3.2124, 3.0362, 2.5744, 2.1607, 2.8928], + device='cuda:3'), covar=tensor([0.0899, 0.0867, 0.0500, 0.0649, 0.0562, 0.0885, 0.0806, 0.0535], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0203, 0.0186, 0.0175, 0.0179, 0.0182, 0.0154, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:52:21,107 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118524.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 15:52:24,935 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 15:52:40,874 INFO [finetune.py:976] (3/7) Epoch 21, batch 4000, loss[loss=0.1749, simple_loss=0.2368, pruned_loss=0.05652, over 4889.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2417, pruned_loss=0.05044, over 957085.01 frames. ], batch size: 32, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:52:47,269 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.678e+01 1.545e+02 1.901e+02 2.367e+02 3.489e+02, threshold=3.803e+02, percent-clipped=0.0 +2023-04-27 15:52:48,622 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0452, 2.4882, 2.0662, 2.0504, 1.4299, 1.4843, 2.0677, 1.4223], + device='cuda:3'), covar=tensor([0.1622, 0.1360, 0.1352, 0.1554, 0.2310, 0.1936, 0.1008, 0.2022], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0212, 0.0170, 0.0205, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 15:52:52,749 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5933, 1.9696, 1.6623, 2.3405, 2.5670, 2.1306, 2.0346, 1.8798], + device='cuda:3'), covar=tensor([0.1788, 0.1613, 0.2267, 0.1855, 0.1207, 0.1999, 0.2140, 0.2171], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0309, 0.0349, 0.0287, 0.0324, 0.0306, 0.0299, 0.0368], + device='cuda:3'), out_proj_covar=tensor([6.3360e-05, 6.3901e-05, 7.3640e-05, 5.7955e-05, 6.6859e-05, 6.4177e-05, + 6.2599e-05, 7.8023e-05], device='cuda:3') +2023-04-27 15:53:30,868 INFO [finetune.py:976] (3/7) Epoch 21, batch 4050, loss[loss=0.1845, simple_loss=0.2735, pruned_loss=0.04776, over 4836.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2443, pruned_loss=0.05117, over 956630.00 frames. ], batch size: 49, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:53:56,778 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:54:06,609 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118631.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:54:26,640 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118645.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:54:33,133 INFO [finetune.py:976] (3/7) Epoch 21, batch 4100, loss[loss=0.1717, simple_loss=0.2427, pruned_loss=0.05035, over 4807.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2456, pruned_loss=0.05112, over 956765.01 frames. ], batch size: 45, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:54:38,506 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.657e+02 1.981e+02 2.296e+02 3.754e+02, threshold=3.963e+02, percent-clipped=0.0 +2023-04-27 15:54:54,353 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:54:58,634 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:55:00,961 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118696.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:55:06,697 INFO [finetune.py:976] (3/7) Epoch 21, batch 4150, loss[loss=0.1824, simple_loss=0.2588, pruned_loss=0.05305, over 4800.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2488, pruned_loss=0.05318, over 954384.61 frames. ], batch size: 45, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:55:49,008 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6593, 1.9525, 1.7212, 1.4906, 1.2401, 1.2465, 1.7477, 1.2173], + device='cuda:3'), covar=tensor([0.1739, 0.1340, 0.1445, 0.1655, 0.2399, 0.2094, 0.1041, 0.2125], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0170, 0.0204, 0.0199, 0.0185, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 15:55:56,015 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118744.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:56:00,563 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 15:56:07,498 INFO [finetune.py:976] (3/7) Epoch 21, batch 4200, loss[loss=0.1346, simple_loss=0.1956, pruned_loss=0.03677, over 3948.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2482, pruned_loss=0.05194, over 954207.80 frames. ], batch size: 17, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:56:19,453 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.568e+02 1.859e+02 2.228e+02 3.642e+02, threshold=3.719e+02, percent-clipped=0.0 +2023-04-27 15:56:21,902 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8492, 1.9142, 1.7809, 1.4606, 2.0664, 1.6498, 2.5706, 1.5119], + device='cuda:3'), covar=tensor([0.3427, 0.1666, 0.4828, 0.2894, 0.1419, 0.2131, 0.1065, 0.4458], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0347, 0.0422, 0.0351, 0.0378, 0.0373, 0.0367, 0.0415], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 15:56:58,126 INFO [finetune.py:976] (3/7) Epoch 21, batch 4250, loss[loss=0.1225, simple_loss=0.1925, pruned_loss=0.02623, over 4838.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2456, pruned_loss=0.05082, over 954558.20 frames. ], batch size: 25, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:57:13,421 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118824.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:57:32,153 INFO [finetune.py:976] (3/7) Epoch 21, batch 4300, loss[loss=0.1282, simple_loss=0.209, pruned_loss=0.02372, over 4799.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2429, pruned_loss=0.05002, over 955819.66 frames. ], batch size: 29, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:57:37,516 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.744e+01 1.518e+02 1.726e+02 2.140e+02 3.725e+02, threshold=3.451e+02, percent-clipped=1.0 +2023-04-27 15:57:44,622 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118872.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:58:06,089 INFO [finetune.py:976] (3/7) Epoch 21, batch 4350, loss[loss=0.1615, simple_loss=0.2337, pruned_loss=0.04468, over 4934.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2401, pruned_loss=0.04911, over 956623.51 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:58:20,437 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:58:39,669 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:58:51,399 INFO [finetune.py:976] (3/7) Epoch 21, batch 4400, loss[loss=0.1522, simple_loss=0.2301, pruned_loss=0.03713, over 4788.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.242, pruned_loss=0.05017, over 955911.23 frames. ], batch size: 29, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:59:00,773 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.550e+02 1.866e+02 2.301e+02 5.364e+02, threshold=3.732e+02, percent-clipped=5.0 +2023-04-27 15:59:14,269 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118973.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:24,788 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9031, 1.7143, 2.1539, 2.2913, 1.7484, 1.5641, 1.8431, 1.2071], + device='cuda:3'), covar=tensor([0.0491, 0.0674, 0.0389, 0.0540, 0.0643, 0.1200, 0.0599, 0.0744], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:59:35,148 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118985.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:36,317 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:45,378 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118993.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:45,452 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3174, 2.0551, 2.6203, 3.0404, 2.1528, 1.7742, 2.2756, 1.4753], + device='cuda:3'), covar=tensor([0.0519, 0.0755, 0.0443, 0.0458, 0.0624, 0.1256, 0.0668, 0.0744], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0097, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 15:59:52,142 INFO [finetune.py:976] (3/7) Epoch 21, batch 4450, loss[loss=0.1698, simple_loss=0.2407, pruned_loss=0.04947, over 4748.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2447, pruned_loss=0.05115, over 955122.70 frames. ], batch size: 23, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:00:11,986 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119033.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:00:25,718 INFO [finetune.py:976] (3/7) Epoch 21, batch 4500, loss[loss=0.189, simple_loss=0.2617, pruned_loss=0.0582, over 4866.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2471, pruned_loss=0.05226, over 956232.91 frames. ], batch size: 34, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:00:30,591 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.010e+02 1.788e+02 2.087e+02 2.542e+02 6.471e+02, threshold=4.174e+02, percent-clipped=4.0 +2023-04-27 16:00:30,762 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5075, 0.7155, 1.4769, 1.7739, 1.5120, 1.4080, 1.4228, 1.5049], + device='cuda:3'), covar=tensor([0.5785, 0.7505, 0.7244, 0.8406, 0.7191, 0.9396, 0.9186, 1.0281], + device='cuda:3'), in_proj_covar=tensor([0.0428, 0.0410, 0.0502, 0.0503, 0.0456, 0.0486, 0.0493, 0.0499], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:00:39,546 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119075.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:00:58,672 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119093.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:01:15,260 INFO [finetune.py:976] (3/7) Epoch 21, batch 4550, loss[loss=0.1752, simple_loss=0.2568, pruned_loss=0.04679, over 4906.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2474, pruned_loss=0.05194, over 955743.16 frames. ], batch size: 46, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:01:26,573 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9782, 1.8885, 2.3338, 2.5544, 1.9134, 1.5772, 1.9736, 0.9643], + device='cuda:3'), covar=tensor([0.0611, 0.0752, 0.0503, 0.0670, 0.0667, 0.1169, 0.0723, 0.0792], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0068, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:01:59,498 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:02:16,875 INFO [finetune.py:976] (3/7) Epoch 21, batch 4600, loss[loss=0.1553, simple_loss=0.2139, pruned_loss=0.04837, over 4704.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2449, pruned_loss=0.05048, over 953416.12 frames. ], batch size: 23, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:02:17,007 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119154.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 16:02:21,811 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.477e+02 1.832e+02 2.213e+02 3.294e+02, threshold=3.663e+02, percent-clipped=0.0 +2023-04-27 16:02:26,410 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.91 vs. limit=5.0 +2023-04-27 16:02:50,901 INFO [finetune.py:976] (3/7) Epoch 21, batch 4650, loss[loss=0.1842, simple_loss=0.2539, pruned_loss=0.0572, over 4808.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2435, pruned_loss=0.05008, over 954118.62 frames. ], batch size: 25, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:03:24,684 INFO [finetune.py:976] (3/7) Epoch 21, batch 4700, loss[loss=0.1504, simple_loss=0.2163, pruned_loss=0.04226, over 4818.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2406, pruned_loss=0.04888, over 956168.78 frames. ], batch size: 51, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:03:29,613 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.611e+02 1.939e+02 2.351e+02 3.791e+02, threshold=3.879e+02, percent-clipped=1.0 +2023-04-27 16:03:42,116 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.52 vs. limit=2.0 +2023-04-27 16:03:45,982 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119287.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:03:58,705 INFO [finetune.py:976] (3/7) Epoch 21, batch 4750, loss[loss=0.1942, simple_loss=0.2724, pruned_loss=0.05797, over 4749.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2396, pruned_loss=0.04904, over 958244.18 frames. ], batch size: 59, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:04:19,362 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8352, 2.3154, 1.7843, 1.6791, 1.3154, 1.3327, 1.8785, 1.2410], + device='cuda:3'), covar=tensor([0.1625, 0.1382, 0.1468, 0.1787, 0.2398, 0.1960, 0.1026, 0.2160], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0210, 0.0169, 0.0203, 0.0199, 0.0184, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 16:04:39,495 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119335.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:04:44,769 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6308, 1.7527, 0.8430, 1.3552, 1.8976, 1.4891, 1.4000, 1.4988], + device='cuda:3'), covar=tensor([0.0487, 0.0337, 0.0349, 0.0521, 0.0268, 0.0468, 0.0457, 0.0551], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 16:04:47,203 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-04-27 16:04:54,473 INFO [finetune.py:976] (3/7) Epoch 21, batch 4800, loss[loss=0.1662, simple_loss=0.2485, pruned_loss=0.04198, over 4804.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2428, pruned_loss=0.05029, over 956325.82 frames. ], batch size: 45, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:04:54,748 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 16:04:59,370 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.524e+02 1.812e+02 2.153e+02 4.144e+02, threshold=3.624e+02, percent-clipped=1.0 +2023-04-27 16:05:26,791 INFO [finetune.py:976] (3/7) Epoch 21, batch 4850, loss[loss=0.1634, simple_loss=0.2387, pruned_loss=0.04402, over 4917.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2458, pruned_loss=0.05094, over 957033.73 frames. ], batch size: 38, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:05:43,555 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119431.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:05:55,518 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119449.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 16:05:56,842 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-04-27 16:05:58,991 INFO [finetune.py:976] (3/7) Epoch 21, batch 4900, loss[loss=0.2539, simple_loss=0.308, pruned_loss=0.09987, over 4200.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2485, pruned_loss=0.05232, over 955993.29 frames. ], batch size: 66, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:06:04,797 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.637e+02 1.876e+02 2.288e+02 4.482e+02, threshold=3.752e+02, percent-clipped=1.0 +2023-04-27 16:06:22,136 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-04-27 16:06:35,841 INFO [finetune.py:976] (3/7) Epoch 21, batch 4950, loss[loss=0.1238, simple_loss=0.2067, pruned_loss=0.0205, over 4753.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2495, pruned_loss=0.0522, over 957114.53 frames. ], batch size: 26, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:06:49,595 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119514.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:07:42,892 INFO [finetune.py:976] (3/7) Epoch 21, batch 5000, loss[loss=0.1792, simple_loss=0.2571, pruned_loss=0.0506, over 4807.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2469, pruned_loss=0.05163, over 954484.28 frames. ], batch size: 45, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:07:56,236 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.721e+01 1.533e+02 1.932e+02 2.312e+02 5.244e+02, threshold=3.864e+02, percent-clipped=4.0 +2023-04-27 16:08:02,455 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8236, 2.2832, 1.8622, 1.7050, 1.3787, 1.3819, 1.9120, 1.3091], + device='cuda:3'), covar=tensor([0.1451, 0.1217, 0.1335, 0.1612, 0.2179, 0.1843, 0.0957, 0.1929], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0211, 0.0168, 0.0203, 0.0199, 0.0185, 0.0155, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 16:08:15,217 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119575.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:08:44,469 INFO [finetune.py:976] (3/7) Epoch 21, batch 5050, loss[loss=0.1697, simple_loss=0.2505, pruned_loss=0.04447, over 4857.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2449, pruned_loss=0.05133, over 953333.77 frames. ], batch size: 49, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:09:22,561 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 16:09:28,916 INFO [finetune.py:976] (3/7) Epoch 21, batch 5100, loss[loss=0.159, simple_loss=0.2232, pruned_loss=0.04741, over 4916.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2407, pruned_loss=0.04967, over 952945.12 frames. ], batch size: 37, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:09:41,161 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7847, 3.7338, 2.8321, 4.3784, 3.8309, 3.8356, 1.8171, 3.7735], + device='cuda:3'), covar=tensor([0.1964, 0.1268, 0.3925, 0.1422, 0.4467, 0.1736, 0.5721, 0.2522], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0215, 0.0251, 0.0304, 0.0295, 0.0245, 0.0275, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 16:09:41,672 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.943e+01 1.499e+02 1.766e+02 2.161e+02 3.760e+02, threshold=3.532e+02, percent-clipped=0.0 +2023-04-27 16:09:43,046 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119665.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:10:03,830 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 16:10:17,338 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8830, 2.3838, 2.6539, 3.2497, 3.0857, 2.6544, 2.2609, 2.8533], + device='cuda:3'), covar=tensor([0.0746, 0.1022, 0.0578, 0.0546, 0.0491, 0.0819, 0.0719, 0.0528], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0203, 0.0184, 0.0174, 0.0178, 0.0181, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:10:31,234 INFO [finetune.py:976] (3/7) Epoch 21, batch 5150, loss[loss=0.1781, simple_loss=0.2473, pruned_loss=0.05443, over 4858.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.241, pruned_loss=0.04984, over 950413.13 frames. ], batch size: 31, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:11:05,036 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119726.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:11:08,027 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119731.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:11:30,485 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119749.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:11:33,424 INFO [finetune.py:976] (3/7) Epoch 21, batch 5200, loss[loss=0.2274, simple_loss=0.2996, pruned_loss=0.07758, over 4728.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2451, pruned_loss=0.05146, over 950263.98 frames. ], batch size: 54, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:11:39,367 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.908e+01 1.576e+02 2.065e+02 2.326e+02 4.790e+02, threshold=4.130e+02, percent-clipped=1.0 +2023-04-27 16:11:51,108 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119779.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:12:02,215 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:12:02,886 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4982, 1.6820, 1.4827, 1.1529, 1.1880, 1.1490, 1.4591, 1.1452], + device='cuda:3'), covar=tensor([0.1820, 0.1373, 0.1559, 0.1788, 0.2509, 0.2106, 0.1108, 0.2150], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0169, 0.0204, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 16:12:06,570 INFO [finetune.py:976] (3/7) Epoch 21, batch 5250, loss[loss=0.1825, simple_loss=0.2481, pruned_loss=0.0584, over 4826.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2466, pruned_loss=0.05138, over 948474.36 frames. ], batch size: 33, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:12:15,534 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4056, 1.0840, 1.2444, 1.1860, 1.5999, 1.2803, 1.0902, 1.1451], + device='cuda:3'), covar=tensor([0.1978, 0.1511, 0.2162, 0.1478, 0.0868, 0.1550, 0.2030, 0.2552], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0312, 0.0352, 0.0289, 0.0328, 0.0309, 0.0302, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.4267e-05, 6.4480e-05, 7.4264e-05, 5.8241e-05, 6.7605e-05, 6.4672e-05, + 6.3297e-05, 7.8998e-05], device='cuda:3') +2023-04-27 16:12:42,056 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5677, 1.3443, 4.4738, 4.2029, 3.9148, 4.2262, 4.1435, 3.9770], + device='cuda:3'), covar=tensor([0.6816, 0.5869, 0.0979, 0.1680, 0.1150, 0.1328, 0.1351, 0.1441], + device='cuda:3'), in_proj_covar=tensor([0.0307, 0.0304, 0.0402, 0.0402, 0.0345, 0.0406, 0.0309, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:12:51,197 INFO [finetune.py:976] (3/7) Epoch 21, batch 5300, loss[loss=0.189, simple_loss=0.2563, pruned_loss=0.06087, over 4760.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2484, pruned_loss=0.05173, over 950286.65 frames. ], batch size: 26, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:13:02,329 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.628e+02 1.910e+02 2.445e+02 5.972e+02, threshold=3.820e+02, percent-clipped=3.0 +2023-04-27 16:13:02,480 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1898, 1.6169, 2.0088, 2.3474, 2.0164, 1.6258, 1.3875, 1.8660], + device='cuda:3'), covar=tensor([0.2679, 0.2785, 0.1401, 0.2035, 0.2290, 0.2261, 0.3939, 0.1786], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0244, 0.0225, 0.0313, 0.0218, 0.0232, 0.0227, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 16:13:13,804 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119870.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:13:37,790 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3692, 3.3683, 1.3593, 1.7564, 1.7302, 2.5028, 1.8581, 0.9949], + device='cuda:3'), covar=tensor([0.1492, 0.0882, 0.1684, 0.1281, 0.1171, 0.0895, 0.1504, 0.2033], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0119, 0.0132, 0.0151, 0.0115, 0.0117], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:13:53,927 INFO [finetune.py:976] (3/7) Epoch 21, batch 5350, loss[loss=0.1734, simple_loss=0.2538, pruned_loss=0.04648, over 4837.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2478, pruned_loss=0.05129, over 950700.91 frames. ], batch size: 49, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:14:43,495 INFO [finetune.py:976] (3/7) Epoch 21, batch 5400, loss[loss=0.1407, simple_loss=0.219, pruned_loss=0.03117, over 4823.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2454, pruned_loss=0.05123, over 950995.33 frames. ], batch size: 25, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:14:48,954 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.397e+02 1.683e+02 2.017e+02 3.693e+02, threshold=3.366e+02, percent-clipped=0.0 +2023-04-27 16:15:18,157 INFO [finetune.py:976] (3/7) Epoch 21, batch 5450, loss[loss=0.1511, simple_loss=0.2311, pruned_loss=0.03552, over 4870.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2424, pruned_loss=0.05003, over 952923.39 frames. ], batch size: 31, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:15:34,277 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=120021.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:16:01,522 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 16:16:03,188 INFO [finetune.py:976] (3/7) Epoch 21, batch 5500, loss[loss=0.1669, simple_loss=0.2455, pruned_loss=0.04415, over 4816.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2396, pruned_loss=0.04931, over 954022.24 frames. ], batch size: 40, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:16:14,182 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.019e+02 1.665e+02 1.823e+02 2.184e+02 3.265e+02, threshold=3.646e+02, percent-clipped=0.0 +2023-04-27 16:17:07,728 INFO [finetune.py:976] (3/7) Epoch 21, batch 5550, loss[loss=0.1826, simple_loss=0.2717, pruned_loss=0.04676, over 4815.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2403, pruned_loss=0.04909, over 952804.25 frames. ], batch size: 39, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:17:07,838 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3092, 1.4704, 1.3232, 1.6852, 1.5817, 1.8974, 1.3612, 3.2844], + device='cuda:3'), covar=tensor([0.0587, 0.0789, 0.0789, 0.1197, 0.0598, 0.0558, 0.0765, 0.0156], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0037, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 16:17:59,284 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0535, 2.1866, 0.7194, 1.3994, 1.3448, 1.6683, 1.4542, 0.7646], + device='cuda:3'), covar=tensor([0.1265, 0.1201, 0.1679, 0.1052, 0.0973, 0.0924, 0.1596, 0.1441], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0241, 0.0137, 0.0119, 0.0133, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:18:05,779 INFO [finetune.py:976] (3/7) Epoch 21, batch 5600, loss[loss=0.1559, simple_loss=0.2252, pruned_loss=0.04335, over 4836.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2442, pruned_loss=0.05013, over 953845.55 frames. ], batch size: 30, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:18:11,017 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.536e+02 1.871e+02 2.277e+02 4.229e+02, threshold=3.743e+02, percent-clipped=5.0 +2023-04-27 16:18:14,689 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 16:18:15,199 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=120170.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:18:36,563 INFO [finetune.py:976] (3/7) Epoch 21, batch 5650, loss[loss=0.1704, simple_loss=0.2598, pruned_loss=0.04054, over 4807.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2467, pruned_loss=0.05056, over 952389.25 frames. ], batch size: 45, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:18:37,754 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8952, 1.0704, 1.6308, 1.7583, 1.6527, 1.7710, 1.6100, 1.6290], + device='cuda:3'), covar=tensor([0.4045, 0.5087, 0.4083, 0.3951, 0.5224, 0.6866, 0.4460, 0.4394], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0374, 0.0323, 0.0337, 0.0348, 0.0395, 0.0358, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 16:18:45,337 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=120218.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:19:07,452 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3601, 2.7211, 1.2694, 1.6148, 2.3141, 1.4581, 3.7122, 2.2199], + device='cuda:3'), covar=tensor([0.0628, 0.0608, 0.0746, 0.1246, 0.0454, 0.1002, 0.0198, 0.0540], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0047, 0.0046, 0.0049, 0.0052, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 16:19:30,420 INFO [finetune.py:976] (3/7) Epoch 21, batch 5700, loss[loss=0.1388, simple_loss=0.2002, pruned_loss=0.03871, over 4354.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2415, pruned_loss=0.0494, over 932707.07 frames. ], batch size: 19, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:19:41,630 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.444e+02 1.777e+02 2.267e+02 3.469e+02, threshold=3.555e+02, percent-clipped=0.0 +2023-04-27 16:20:17,303 INFO [finetune.py:976] (3/7) Epoch 22, batch 0, loss[loss=0.1702, simple_loss=0.2529, pruned_loss=0.04377, over 4910.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2529, pruned_loss=0.04377, over 4910.00 frames. ], batch size: 38, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:20:17,303 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 16:20:33,674 INFO [finetune.py:1010] (3/7) Epoch 22, validation: loss=0.1546, simple_loss=0.2251, pruned_loss=0.04204, over 2265189.00 frames. +2023-04-27 16:20:33,674 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 16:20:37,875 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9881, 1.6679, 1.9082, 2.3941, 2.3830, 1.8896, 1.6694, 2.0945], + device='cuda:3'), covar=tensor([0.0808, 0.1298, 0.0765, 0.0607, 0.0611, 0.0894, 0.0832, 0.0592], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0201, 0.0183, 0.0174, 0.0177, 0.0180, 0.0152, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:20:59,177 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=120321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:21:05,111 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-04-27 16:21:06,359 INFO [finetune.py:976] (3/7) Epoch 22, batch 50, loss[loss=0.2099, simple_loss=0.2747, pruned_loss=0.07255, over 4874.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2461, pruned_loss=0.05148, over 213879.03 frames. ], batch size: 34, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:21:21,923 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9826, 1.7211, 2.1731, 2.3662, 2.0098, 1.9501, 2.0490, 1.9675], + device='cuda:3'), covar=tensor([0.4961, 0.7647, 0.7172, 0.5874, 0.6230, 0.8087, 0.9430, 1.1426], + device='cuda:3'), in_proj_covar=tensor([0.0429, 0.0412, 0.0505, 0.0504, 0.0457, 0.0487, 0.0495, 0.0502], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:21:23,745 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4535, 1.5858, 1.7721, 1.8899, 1.6597, 1.8425, 1.8260, 1.8830], + device='cuda:3'), covar=tensor([0.3855, 0.5369, 0.4711, 0.4440, 0.6190, 0.7535, 0.5323, 0.5053], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0375, 0.0323, 0.0338, 0.0348, 0.0395, 0.0359, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 16:21:24,327 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9478, 2.5845, 0.8407, 1.2632, 1.8568, 1.1133, 3.4277, 1.6355], + device='cuda:3'), covar=tensor([0.0941, 0.0802, 0.1029, 0.1767, 0.0682, 0.1484, 0.0311, 0.0895], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0047, 0.0046, 0.0049, 0.0052, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 16:21:27,312 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.523e+01 1.622e+02 1.941e+02 2.362e+02 4.057e+02, threshold=3.882e+02, percent-clipped=2.0 +2023-04-27 16:21:31,538 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=120369.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:21:51,383 INFO [finetune.py:976] (3/7) Epoch 22, batch 100, loss[loss=0.1704, simple_loss=0.2485, pruned_loss=0.04618, over 4828.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2407, pruned_loss=0.04932, over 380991.03 frames. ], batch size: 40, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:22:49,222 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 16:22:51,081 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5372, 1.1004, 0.4439, 1.2619, 1.0855, 1.4096, 1.3535, 1.3553], + device='cuda:3'), covar=tensor([0.0510, 0.0426, 0.0417, 0.0565, 0.0303, 0.0531, 0.0496, 0.0602], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 16:22:55,503 INFO [finetune.py:976] (3/7) Epoch 22, batch 150, loss[loss=0.1857, simple_loss=0.2516, pruned_loss=0.05988, over 4874.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2367, pruned_loss=0.04862, over 509114.74 frames. ], batch size: 31, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:23:25,740 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 16:23:29,000 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.459e+02 1.791e+02 2.121e+02 3.336e+02, threshold=3.582e+02, percent-clipped=0.0 +2023-04-27 16:23:58,243 INFO [finetune.py:976] (3/7) Epoch 22, batch 200, loss[loss=0.1733, simple_loss=0.2314, pruned_loss=0.05761, over 4324.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2349, pruned_loss=0.04814, over 604420.79 frames. ], batch size: 19, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:25:06,005 INFO [finetune.py:976] (3/7) Epoch 22, batch 250, loss[loss=0.1599, simple_loss=0.2415, pruned_loss=0.03909, over 4746.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2378, pruned_loss=0.04909, over 682491.75 frames. ], batch size: 59, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:25:37,775 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 16:25:49,438 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.595e+02 1.873e+02 2.274e+02 4.484e+02, threshold=3.746e+02, percent-clipped=2.0 +2023-04-27 16:26:12,262 INFO [finetune.py:976] (3/7) Epoch 22, batch 300, loss[loss=0.1721, simple_loss=0.2622, pruned_loss=0.04104, over 4905.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2426, pruned_loss=0.05054, over 742054.64 frames. ], batch size: 37, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:27:06,862 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8562, 1.7476, 1.6019, 1.2966, 1.7897, 1.5214, 2.1186, 1.4488], + device='cuda:3'), covar=tensor([0.3086, 0.1611, 0.4006, 0.2560, 0.1263, 0.1906, 0.1367, 0.3919], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0351, 0.0427, 0.0354, 0.0382, 0.0374, 0.0369, 0.0419], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:27:19,461 INFO [finetune.py:976] (3/7) Epoch 22, batch 350, loss[loss=0.188, simple_loss=0.268, pruned_loss=0.05403, over 4794.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2462, pruned_loss=0.05206, over 788902.34 frames. ], batch size: 51, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:28:01,877 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.528e+02 1.802e+02 2.377e+02 4.881e+02, threshold=3.603e+02, percent-clipped=3.0 +2023-04-27 16:28:20,625 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=120676.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:28:24,207 INFO [finetune.py:976] (3/7) Epoch 22, batch 400, loss[loss=0.163, simple_loss=0.2471, pruned_loss=0.03948, over 4746.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2465, pruned_loss=0.05095, over 826712.68 frames. ], batch size: 27, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:28:44,117 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7450, 1.1431, 1.4426, 1.6050, 1.5023, 1.2548, 0.7319, 1.1443], + device='cuda:3'), covar=tensor([0.3693, 0.4448, 0.2349, 0.2469, 0.2940, 0.2938, 0.4761, 0.2357], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0243, 0.0225, 0.0312, 0.0218, 0.0232, 0.0227, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 16:29:32,129 INFO [finetune.py:976] (3/7) Epoch 22, batch 450, loss[loss=0.2104, simple_loss=0.2877, pruned_loss=0.06657, over 4890.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2465, pruned_loss=0.05071, over 854919.65 frames. ], batch size: 35, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:29:40,504 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=120737.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:29:56,710 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1827, 1.8931, 2.2275, 2.5647, 2.5304, 2.1568, 1.7838, 2.3123], + device='cuda:3'), covar=tensor([0.0758, 0.1041, 0.0587, 0.0493, 0.0517, 0.0788, 0.0720, 0.0508], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0198, 0.0181, 0.0172, 0.0175, 0.0177, 0.0149, 0.0175], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:30:03,210 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2290, 1.3983, 1.7133, 1.8189, 1.6794, 1.8074, 1.7395, 1.7536], + device='cuda:3'), covar=tensor([0.3357, 0.4959, 0.3906, 0.3986, 0.5023, 0.6389, 0.4287, 0.4030], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0372, 0.0323, 0.0337, 0.0347, 0.0394, 0.0356, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 16:30:03,658 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.511e+02 1.855e+02 2.296e+02 3.408e+02, threshold=3.711e+02, percent-clipped=0.0 +2023-04-27 16:30:15,273 INFO [finetune.py:976] (3/7) Epoch 22, batch 500, loss[loss=0.1285, simple_loss=0.1954, pruned_loss=0.03082, over 4741.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2427, pruned_loss=0.04933, over 877152.90 frames. ], batch size: 54, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:30:22,590 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0792, 0.6271, 0.9016, 0.8321, 1.1799, 0.9358, 0.8582, 0.9403], + device='cuda:3'), covar=tensor([0.1921, 0.1816, 0.2388, 0.1618, 0.1201, 0.1742, 0.1930, 0.2666], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0310, 0.0349, 0.0287, 0.0324, 0.0308, 0.0300, 0.0370], + device='cuda:3'), out_proj_covar=tensor([6.3866e-05, 6.4161e-05, 7.3787e-05, 5.7746e-05, 6.6868e-05, 6.4541e-05, + 6.2753e-05, 7.8634e-05], device='cuda:3') +2023-04-27 16:30:28,388 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-27 16:30:49,307 INFO [finetune.py:976] (3/7) Epoch 22, batch 550, loss[loss=0.1555, simple_loss=0.2365, pruned_loss=0.03723, over 4917.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2398, pruned_loss=0.0482, over 893483.64 frames. ], batch size: 36, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:30:59,021 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8311, 1.6498, 4.0161, 3.7740, 3.5022, 3.7216, 3.5927, 3.5750], + device='cuda:3'), covar=tensor([0.6023, 0.4448, 0.0909, 0.1298, 0.1011, 0.1748, 0.3472, 0.1202], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0307, 0.0409, 0.0408, 0.0350, 0.0412, 0.0314, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:31:15,726 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.573e+02 1.876e+02 2.291e+02 4.467e+02, threshold=3.751e+02, percent-clipped=2.0 +2023-04-27 16:31:38,950 INFO [finetune.py:976] (3/7) Epoch 22, batch 600, loss[loss=0.1747, simple_loss=0.2413, pruned_loss=0.05408, over 4790.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2408, pruned_loss=0.04893, over 905478.72 frames. ], batch size: 29, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:32:09,827 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5732, 2.1035, 2.5030, 2.9477, 2.5485, 2.0837, 1.9400, 2.3427], + device='cuda:3'), covar=tensor([0.3253, 0.3044, 0.1640, 0.2046, 0.2489, 0.2578, 0.3490, 0.2070], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0245, 0.0226, 0.0315, 0.0220, 0.0234, 0.0228, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 16:32:45,425 INFO [finetune.py:976] (3/7) Epoch 22, batch 650, loss[loss=0.1535, simple_loss=0.23, pruned_loss=0.03846, over 4706.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2427, pruned_loss=0.04963, over 915846.56 frames. ], batch size: 23, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:32:55,842 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5430, 2.9969, 0.9820, 1.7263, 2.1690, 1.4734, 4.0934, 2.1671], + device='cuda:3'), covar=tensor([0.0599, 0.0743, 0.0981, 0.1220, 0.0507, 0.0975, 0.0265, 0.0599], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 16:33:16,620 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5269, 1.6639, 1.6043, 2.1168, 1.9717, 2.1552, 1.6641, 4.4470], + device='cuda:3'), covar=tensor([0.0558, 0.0805, 0.0798, 0.1169, 0.0602, 0.0477, 0.0708, 0.0113], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0037, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 16:33:26,706 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.656e+02 1.853e+02 2.297e+02 3.999e+02, threshold=3.705e+02, percent-clipped=2.0 +2023-04-27 16:33:52,001 INFO [finetune.py:976] (3/7) Epoch 22, batch 700, loss[loss=0.2236, simple_loss=0.2865, pruned_loss=0.08029, over 4037.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2462, pruned_loss=0.05144, over 923846.98 frames. ], batch size: 65, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:34:21,528 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121003.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:34:35,163 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0118, 2.3978, 0.8625, 1.2680, 1.4145, 1.7879, 1.5887, 0.8495], + device='cuda:3'), covar=tensor([0.1881, 0.2012, 0.1981, 0.1900, 0.1353, 0.1188, 0.1739, 0.1906], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0119, 0.0132, 0.0152, 0.0116, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:34:45,578 INFO [finetune.py:976] (3/7) Epoch 22, batch 750, loss[loss=0.1851, simple_loss=0.266, pruned_loss=0.05216, over 4892.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2476, pruned_loss=0.0519, over 932339.12 frames. ], batch size: 43, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:34:45,653 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:35:04,853 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.944e+01 1.522e+02 1.865e+02 2.464e+02 7.582e+02, threshold=3.731e+02, percent-clipped=3.0 +2023-04-27 16:35:05,562 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121064.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:35:19,166 INFO [finetune.py:976] (3/7) Epoch 22, batch 800, loss[loss=0.1909, simple_loss=0.2601, pruned_loss=0.06088, over 4871.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2468, pruned_loss=0.05109, over 936590.72 frames. ], batch size: 34, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:35:52,469 INFO [finetune.py:976] (3/7) Epoch 22, batch 850, loss[loss=0.2013, simple_loss=0.2641, pruned_loss=0.06926, over 4903.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2459, pruned_loss=0.05127, over 940725.46 frames. ], batch size: 35, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:36:11,707 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.493e+02 1.782e+02 2.235e+02 3.771e+02, threshold=3.565e+02, percent-clipped=1.0 +2023-04-27 16:36:12,462 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121164.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:36:17,916 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4867, 1.1219, 0.3847, 1.2201, 1.1210, 1.3727, 1.3048, 1.3029], + device='cuda:3'), covar=tensor([0.0490, 0.0382, 0.0409, 0.0567, 0.0300, 0.0486, 0.0492, 0.0550], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 16:36:25,801 INFO [finetune.py:976] (3/7) Epoch 22, batch 900, loss[loss=0.1667, simple_loss=0.2473, pruned_loss=0.04307, over 4907.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2428, pruned_loss=0.05019, over 944663.62 frames. ], batch size: 37, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:36:32,587 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5858, 1.3949, 1.3312, 1.3877, 1.8017, 1.4649, 1.2302, 1.2518], + device='cuda:3'), covar=tensor([0.1704, 0.1293, 0.1826, 0.1311, 0.0793, 0.1685, 0.1979, 0.2332], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0314, 0.0353, 0.0290, 0.0327, 0.0312, 0.0304, 0.0375], + device='cuda:3'), out_proj_covar=tensor([6.4606e-05, 6.4882e-05, 7.4589e-05, 5.8363e-05, 6.7505e-05, 6.5293e-05, + 6.3554e-05, 7.9724e-05], device='cuda:3') +2023-04-27 16:36:39,363 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-27 16:36:52,972 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:36:59,092 INFO [finetune.py:976] (3/7) Epoch 22, batch 950, loss[loss=0.2306, simple_loss=0.2999, pruned_loss=0.0807, over 4931.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2415, pruned_loss=0.05025, over 946474.66 frames. ], batch size: 38, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:37:35,638 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.216e+01 1.432e+02 1.893e+02 2.215e+02 5.796e+02, threshold=3.785e+02, percent-clipped=6.0 +2023-04-27 16:37:48,376 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8856, 1.3923, 1.6875, 1.7447, 1.6816, 1.3983, 0.8690, 1.3798], + device='cuda:3'), covar=tensor([0.3005, 0.3220, 0.1580, 0.1876, 0.2352, 0.2565, 0.4087, 0.2063], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0243, 0.0225, 0.0312, 0.0219, 0.0232, 0.0227, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 16:38:01,688 INFO [finetune.py:976] (3/7) Epoch 22, batch 1000, loss[loss=0.194, simple_loss=0.2674, pruned_loss=0.06024, over 4817.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2453, pruned_loss=0.05146, over 948644.83 frames. ], batch size: 38, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:38:41,528 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121311.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:38:53,678 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4386, 1.9662, 2.3519, 2.8672, 2.3597, 1.9052, 1.8878, 2.2356], + device='cuda:3'), covar=tensor([0.3052, 0.2986, 0.1481, 0.2239, 0.2633, 0.2532, 0.3567, 0.2043], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0243, 0.0225, 0.0312, 0.0219, 0.0232, 0.0227, 0.0182], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 16:39:01,159 INFO [finetune.py:976] (3/7) Epoch 22, batch 1050, loss[loss=0.1946, simple_loss=0.269, pruned_loss=0.0601, over 4828.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2481, pruned_loss=0.0521, over 951134.34 frames. ], batch size: 30, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:39:01,274 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121332.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:39:04,679 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5694, 1.2101, 1.4029, 1.3276, 1.7148, 1.4455, 1.2461, 1.3563], + device='cuda:3'), covar=tensor([0.1788, 0.1772, 0.2102, 0.1807, 0.1188, 0.1713, 0.1942, 0.2673], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0314, 0.0353, 0.0289, 0.0327, 0.0311, 0.0304, 0.0375], + device='cuda:3'), out_proj_covar=tensor([6.4642e-05, 6.4902e-05, 7.4540e-05, 5.8234e-05, 6.7377e-05, 6.5127e-05, + 6.3585e-05, 7.9678e-05], device='cuda:3') +2023-04-27 16:39:18,517 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121359.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:39:20,860 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.706e+02 2.101e+02 2.455e+02 4.347e+02, threshold=4.202e+02, percent-clipped=2.0 +2023-04-27 16:39:22,182 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1046, 2.2791, 2.0249, 1.8454, 2.3612, 1.9587, 2.9594, 1.7985], + device='cuda:3'), covar=tensor([0.3640, 0.1793, 0.4107, 0.3170, 0.1522, 0.2425, 0.1204, 0.4452], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0346, 0.0422, 0.0350, 0.0379, 0.0372, 0.0365, 0.0415], + device='cuda:3'), out_proj_covar=tensor([9.9829e-05, 1.0373e-04, 1.2814e-04, 1.0534e-04, 1.1266e-04, 1.1073e-04, + 1.0724e-04, 1.2515e-04], device='cuda:3') +2023-04-27 16:39:25,704 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5989, 3.5708, 2.6848, 4.1318, 3.6498, 3.6454, 1.6011, 3.5450], + device='cuda:3'), covar=tensor([0.1915, 0.1294, 0.3302, 0.2001, 0.3550, 0.1753, 0.5651, 0.2533], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0216, 0.0251, 0.0305, 0.0294, 0.0245, 0.0274, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 16:39:26,937 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121372.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:39:31,744 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:39:33,388 INFO [finetune.py:976] (3/7) Epoch 22, batch 1100, loss[loss=0.1593, simple_loss=0.2426, pruned_loss=0.03802, over 4918.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2493, pruned_loss=0.05265, over 951545.62 frames. ], batch size: 38, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:39:45,243 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7694, 3.9212, 0.9146, 2.0597, 2.1008, 2.6676, 2.3202, 0.9353], + device='cuda:3'), covar=tensor([0.1332, 0.1051, 0.2048, 0.1355, 0.1083, 0.1097, 0.1476, 0.2324], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0120, 0.0132, 0.0153, 0.0116, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:39:50,918 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121407.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:40:06,338 INFO [finetune.py:976] (3/7) Epoch 22, batch 1150, loss[loss=0.2087, simple_loss=0.2768, pruned_loss=0.07026, over 4806.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2502, pruned_loss=0.05306, over 952532.03 frames. ], batch size: 40, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:40:27,719 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.622e+02 1.904e+02 2.379e+02 4.461e+02, threshold=3.808e+02, percent-clipped=1.0 +2023-04-27 16:40:30,882 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121468.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:40:39,780 INFO [finetune.py:976] (3/7) Epoch 22, batch 1200, loss[loss=0.137, simple_loss=0.2111, pruned_loss=0.03149, over 4229.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2472, pruned_loss=0.05172, over 952771.56 frames. ], batch size: 66, lr: 3.15e-03, grad_scale: 16.0 +2023-04-27 16:40:45,090 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1890, 1.4715, 1.3466, 1.7357, 1.5902, 1.6444, 1.3957, 2.9484], + device='cuda:3'), covar=tensor([0.0650, 0.0765, 0.0769, 0.1125, 0.0598, 0.0529, 0.0713, 0.0169], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 16:41:04,811 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 16:41:05,070 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:41:12,826 INFO [finetune.py:976] (3/7) Epoch 22, batch 1250, loss[loss=0.1194, simple_loss=0.1932, pruned_loss=0.02281, over 4831.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.244, pruned_loss=0.05075, over 953151.52 frames. ], batch size: 30, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:41:34,045 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.460e+02 1.755e+02 2.215e+02 4.561e+02, threshold=3.510e+02, percent-clipped=2.0 +2023-04-27 16:41:40,321 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5853, 1.3070, 1.2858, 1.3360, 1.7736, 1.4887, 1.1565, 1.2432], + device='cuda:3'), covar=tensor([0.1599, 0.1431, 0.1936, 0.1489, 0.0848, 0.1500, 0.2190, 0.2513], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0314, 0.0353, 0.0290, 0.0327, 0.0311, 0.0304, 0.0376], + device='cuda:3'), out_proj_covar=tensor([6.4730e-05, 6.4954e-05, 7.4562e-05, 5.8347e-05, 6.7454e-05, 6.5171e-05, + 6.3528e-05, 7.9881e-05], device='cuda:3') +2023-04-27 16:41:46,169 INFO [finetune.py:976] (3/7) Epoch 22, batch 1300, loss[loss=0.2002, simple_loss=0.2664, pruned_loss=0.06704, over 4903.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2407, pruned_loss=0.04934, over 952115.91 frames. ], batch size: 37, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:42:19,075 INFO [finetune.py:976] (3/7) Epoch 22, batch 1350, loss[loss=0.1835, simple_loss=0.2537, pruned_loss=0.05668, over 4901.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2409, pruned_loss=0.04967, over 951767.23 frames. ], batch size: 35, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:42:44,889 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4791, 1.4054, 1.7764, 1.7958, 1.3683, 1.2078, 1.5498, 0.9437], + device='cuda:3'), covar=tensor([0.0618, 0.0616, 0.0384, 0.0574, 0.0771, 0.1190, 0.0545, 0.0597], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0072, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:42:55,349 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6934, 1.5689, 1.9094, 1.9965, 1.5708, 1.3075, 1.7221, 1.0815], + device='cuda:3'), covar=tensor([0.0738, 0.0522, 0.0508, 0.0673, 0.0719, 0.0945, 0.0543, 0.0679], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0072, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:42:55,952 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:42:58,253 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.587e+02 1.863e+02 2.182e+02 3.847e+02, threshold=3.726e+02, percent-clipped=2.0 +2023-04-27 16:43:04,719 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121667.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:43:07,787 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8084, 2.2834, 0.9728, 1.1281, 1.5442, 1.0872, 2.4763, 1.2878], + device='cuda:3'), covar=tensor([0.0750, 0.0636, 0.0655, 0.1394, 0.0497, 0.1152, 0.0367, 0.0767], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0052, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 16:43:19,351 INFO [finetune.py:976] (3/7) Epoch 22, batch 1400, loss[loss=0.135, simple_loss=0.2171, pruned_loss=0.02643, over 4783.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2433, pruned_loss=0.05033, over 951950.90 frames. ], batch size: 26, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:43:59,849 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:44:10,148 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 16:44:24,395 INFO [finetune.py:976] (3/7) Epoch 22, batch 1450, loss[loss=0.1821, simple_loss=0.2598, pruned_loss=0.05218, over 4915.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2455, pruned_loss=0.05094, over 953677.28 frames. ], batch size: 42, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:45:04,293 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.748e+02 1.956e+02 2.570e+02 3.972e+02, threshold=3.912e+02, percent-clipped=2.0 +2023-04-27 16:45:04,382 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121763.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:45:13,519 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3260, 1.3561, 1.3970, 1.6243, 1.5899, 1.3594, 0.9725, 1.4967], + device='cuda:3'), covar=tensor([0.0864, 0.1248, 0.0933, 0.0647, 0.0685, 0.0830, 0.0921, 0.0580], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0199, 0.0182, 0.0172, 0.0175, 0.0178, 0.0151, 0.0175], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:45:27,550 INFO [finetune.py:976] (3/7) Epoch 22, batch 1500, loss[loss=0.1603, simple_loss=0.2401, pruned_loss=0.04027, over 4776.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2459, pruned_loss=0.05065, over 954719.31 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:45:43,084 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0922, 1.7591, 2.2678, 2.5259, 2.1041, 2.0283, 2.1343, 2.0926], + device='cuda:3'), covar=tensor([0.4981, 0.7886, 0.7500, 0.5697, 0.6427, 0.9043, 0.9102, 1.0365], + device='cuda:3'), in_proj_covar=tensor([0.0430, 0.0413, 0.0505, 0.0505, 0.0457, 0.0488, 0.0494, 0.0502], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:45:52,965 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4591, 1.7738, 1.7893, 1.8877, 1.7119, 1.8213, 1.8987, 1.8615], + device='cuda:3'), covar=tensor([0.4030, 0.5389, 0.4234, 0.4234, 0.5681, 0.7502, 0.4777, 0.4579], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0375, 0.0326, 0.0339, 0.0348, 0.0395, 0.0358, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 16:45:58,989 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:46:06,251 INFO [finetune.py:976] (3/7) Epoch 22, batch 1550, loss[loss=0.1654, simple_loss=0.252, pruned_loss=0.03941, over 4888.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2458, pruned_loss=0.05027, over 954848.62 frames. ], batch size: 43, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:46:28,045 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.267e+01 1.607e+02 1.890e+02 2.272e+02 5.935e+02, threshold=3.780e+02, percent-clipped=2.0 +2023-04-27 16:46:31,189 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121868.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:46:39,811 INFO [finetune.py:976] (3/7) Epoch 22, batch 1600, loss[loss=0.1584, simple_loss=0.2394, pruned_loss=0.03874, over 4842.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.244, pruned_loss=0.04968, over 956437.33 frames. ], batch size: 44, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:46:47,809 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1430, 2.6429, 1.1534, 1.4365, 2.0963, 1.2551, 3.4044, 1.7985], + device='cuda:3'), covar=tensor([0.0644, 0.0653, 0.0742, 0.1168, 0.0473, 0.0981, 0.0239, 0.0606], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 16:47:00,761 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-27 16:47:13,761 INFO [finetune.py:976] (3/7) Epoch 22, batch 1650, loss[loss=0.1428, simple_loss=0.2133, pruned_loss=0.03614, over 4782.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2414, pruned_loss=0.04951, over 957743.23 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:47:24,827 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:47:33,996 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.538e+02 1.825e+02 2.150e+02 3.786e+02, threshold=3.649e+02, percent-clipped=1.0 +2023-04-27 16:47:38,075 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121967.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:47:47,187 INFO [finetune.py:976] (3/7) Epoch 22, batch 1700, loss[loss=0.1431, simple_loss=0.2085, pruned_loss=0.03878, over 4764.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2388, pruned_loss=0.04858, over 957306.35 frames. ], batch size: 27, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:48:06,909 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122011.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:48:16,071 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122015.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:48:37,956 INFO [finetune.py:976] (3/7) Epoch 22, batch 1750, loss[loss=0.2053, simple_loss=0.2796, pruned_loss=0.06548, over 4909.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2416, pruned_loss=0.04963, over 958213.22 frames. ], batch size: 37, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:49:13,875 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.616e+02 1.845e+02 2.409e+02 4.396e+02, threshold=3.689e+02, percent-clipped=3.0 +2023-04-27 16:49:13,991 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122063.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:49:43,794 INFO [finetune.py:976] (3/7) Epoch 22, batch 1800, loss[loss=0.1961, simple_loss=0.2591, pruned_loss=0.06661, over 4903.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2444, pruned_loss=0.05046, over 956791.08 frames. ], batch size: 36, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:50:18,107 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122111.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:50:27,653 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122117.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:50:49,858 INFO [finetune.py:976] (3/7) Epoch 22, batch 1850, loss[loss=0.1868, simple_loss=0.2631, pruned_loss=0.05523, over 4928.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2461, pruned_loss=0.05094, over 957826.66 frames. ], batch size: 38, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:51:27,102 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.233e+01 1.753e+02 1.978e+02 2.407e+02 4.460e+02, threshold=3.956e+02, percent-clipped=4.0 +2023-04-27 16:51:48,067 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122178.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 16:51:56,335 INFO [finetune.py:976] (3/7) Epoch 22, batch 1900, loss[loss=0.1548, simple_loss=0.2202, pruned_loss=0.0447, over 4168.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2478, pruned_loss=0.05141, over 955990.70 frames. ], batch size: 18, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:51:56,480 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2532, 1.6893, 2.1076, 2.4248, 2.0541, 1.6237, 1.3948, 1.8187], + device='cuda:3'), covar=tensor([0.2981, 0.3191, 0.1472, 0.2092, 0.2506, 0.2615, 0.4068, 0.2033], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0244, 0.0227, 0.0314, 0.0221, 0.0234, 0.0228, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 16:52:10,098 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122195.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:52:40,530 INFO [finetune.py:976] (3/7) Epoch 22, batch 1950, loss[loss=0.204, simple_loss=0.2715, pruned_loss=0.06823, over 4840.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2462, pruned_loss=0.05079, over 957305.01 frames. ], batch size: 49, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:52:55,278 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122256.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:52:59,360 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.614e+02 1.828e+02 2.204e+02 4.650e+02, threshold=3.656e+02, percent-clipped=1.0 +2023-04-27 16:53:13,158 INFO [finetune.py:976] (3/7) Epoch 22, batch 2000, loss[loss=0.1569, simple_loss=0.225, pruned_loss=0.04444, over 4783.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2431, pruned_loss=0.04964, over 957585.20 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:53:29,026 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122306.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:53:29,136 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 16:53:32,798 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-27 16:53:37,540 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:53:45,139 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7741, 1.9631, 0.8005, 1.4426, 1.7503, 1.5781, 1.5084, 1.5927], + device='cuda:3'), covar=tensor([0.0489, 0.0353, 0.0379, 0.0554, 0.0292, 0.0531, 0.0519, 0.0576], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 16:53:46,289 INFO [finetune.py:976] (3/7) Epoch 22, batch 2050, loss[loss=0.1796, simple_loss=0.2593, pruned_loss=0.04993, over 4868.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2394, pruned_loss=0.04824, over 956829.16 frames. ], batch size: 31, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:53:50,031 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0202, 2.3421, 2.0266, 2.3788, 1.7414, 2.0925, 2.1763, 1.5707], + device='cuda:3'), covar=tensor([0.1688, 0.0964, 0.0765, 0.0954, 0.2872, 0.1068, 0.1559, 0.2345], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0302, 0.0218, 0.0278, 0.0316, 0.0256, 0.0249, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1452e-04, 1.1956e-04, 8.6168e-05, 1.0976e-04, 1.2790e-04, 1.0141e-04, + 1.0054e-04, 1.0410e-04], device='cuda:3') +2023-04-27 16:54:12,101 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.559e+02 1.860e+02 2.262e+02 4.767e+02, threshold=3.720e+02, percent-clipped=3.0 +2023-04-27 16:54:35,063 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2900, 1.5236, 1.4737, 1.7664, 1.6033, 1.8594, 1.3965, 3.6165], + device='cuda:3'), covar=tensor([0.0634, 0.0829, 0.0842, 0.1283, 0.0681, 0.0527, 0.0784, 0.0158], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 16:54:35,089 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:54:36,181 INFO [finetune.py:976] (3/7) Epoch 22, batch 2100, loss[loss=0.172, simple_loss=0.2367, pruned_loss=0.05364, over 4872.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2398, pruned_loss=0.04861, over 955379.62 frames. ], batch size: 31, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:54:44,310 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2548, 1.7338, 2.1240, 2.7919, 2.2405, 1.7143, 1.9014, 2.1454], + device='cuda:3'), covar=tensor([0.3214, 0.3603, 0.1880, 0.2419, 0.2653, 0.2608, 0.3746, 0.2012], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0244, 0.0227, 0.0313, 0.0221, 0.0233, 0.0228, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 16:54:45,396 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8589, 2.2807, 0.9038, 1.2079, 1.6516, 1.1330, 2.4928, 1.3330], + device='cuda:3'), covar=tensor([0.0706, 0.0543, 0.0656, 0.1294, 0.0430, 0.1048, 0.0334, 0.0682], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 16:55:06,291 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 16:55:19,159 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 16:55:40,953 INFO [finetune.py:976] (3/7) Epoch 22, batch 2150, loss[loss=0.1822, simple_loss=0.2562, pruned_loss=0.05411, over 4770.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2438, pruned_loss=0.0505, over 956099.65 frames. ], batch size: 54, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:56:19,330 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.667e+02 1.982e+02 2.424e+02 1.094e+03, threshold=3.964e+02, percent-clipped=3.0 +2023-04-27 16:56:22,546 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122468.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:56:30,977 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122473.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 16:56:31,620 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1777, 2.1972, 1.8795, 1.8592, 2.3976, 1.8277, 2.9054, 1.7062], + device='cuda:3'), covar=tensor([0.3712, 0.2268, 0.4494, 0.3428, 0.1817, 0.2746, 0.1288, 0.4456], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0350, 0.0427, 0.0354, 0.0382, 0.0376, 0.0368, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:56:42,043 INFO [finetune.py:976] (3/7) Epoch 22, batch 2200, loss[loss=0.186, simple_loss=0.2737, pruned_loss=0.04918, over 4813.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2447, pruned_loss=0.05002, over 956399.84 frames. ], batch size: 40, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:57:03,752 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5914, 1.6615, 0.8224, 1.3051, 1.7822, 1.4293, 1.3591, 1.4665], + device='cuda:3'), covar=tensor([0.0485, 0.0386, 0.0344, 0.0548, 0.0275, 0.0522, 0.0490, 0.0565], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 16:57:48,019 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122529.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:57:49,741 INFO [finetune.py:976] (3/7) Epoch 22, batch 2250, loss[loss=0.1559, simple_loss=0.2457, pruned_loss=0.03305, over 4796.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2468, pruned_loss=0.0507, over 957376.46 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:58:19,798 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122551.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:58:34,009 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.620e+02 1.922e+02 2.285e+02 4.968e+02, threshold=3.845e+02, percent-clipped=1.0 +2023-04-27 16:58:51,721 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9099, 1.7229, 2.2877, 2.3996, 1.6862, 1.5338, 1.9447, 1.0687], + device='cuda:3'), covar=tensor([0.0648, 0.0782, 0.0392, 0.0748, 0.0775, 0.1138, 0.0671, 0.0728], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0067, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 16:58:56,480 INFO [finetune.py:976] (3/7) Epoch 22, batch 2300, loss[loss=0.1566, simple_loss=0.237, pruned_loss=0.03811, over 4803.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2466, pruned_loss=0.05015, over 957569.79 frames. ], batch size: 41, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:59:15,489 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1569, 1.8306, 2.3828, 2.6160, 2.1767, 2.0270, 2.2162, 2.2074], + device='cuda:3'), covar=tensor([0.4866, 0.7773, 0.7334, 0.5998, 0.6317, 0.9949, 0.9687, 0.9837], + device='cuda:3'), in_proj_covar=tensor([0.0433, 0.0415, 0.0510, 0.0508, 0.0461, 0.0492, 0.0499, 0.0506], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 16:59:36,001 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122606.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:00:08,291 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8868, 2.1056, 1.9433, 2.2022, 1.6660, 1.9849, 1.9832, 1.5537], + device='cuda:3'), covar=tensor([0.1632, 0.1171, 0.0770, 0.1035, 0.3183, 0.0981, 0.1554, 0.2115], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0304, 0.0218, 0.0277, 0.0315, 0.0256, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1395e-04, 1.2038e-04, 8.5977e-05, 1.0956e-04, 1.2768e-04, 1.0139e-04, + 1.0017e-04, 1.0410e-04], device='cuda:3') +2023-04-27 17:00:08,779 INFO [finetune.py:976] (3/7) Epoch 22, batch 2350, loss[loss=0.1813, simple_loss=0.2472, pruned_loss=0.05769, over 4310.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2456, pruned_loss=0.05026, over 955693.49 frames. ], batch size: 65, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 17:00:34,411 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3717, 1.7427, 1.6527, 2.0812, 1.9478, 1.9643, 1.6380, 4.3354], + device='cuda:3'), covar=tensor([0.0507, 0.0748, 0.0712, 0.1097, 0.0567, 0.0545, 0.0678, 0.0095], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 17:00:40,761 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122654.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:00:46,226 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.405e+01 1.589e+02 1.908e+02 2.329e+02 5.605e+02, threshold=3.816e+02, percent-clipped=4.0 +2023-04-27 17:01:04,138 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122675.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:01:14,127 INFO [finetune.py:976] (3/7) Epoch 22, batch 2400, loss[loss=0.1705, simple_loss=0.2387, pruned_loss=0.0511, over 4828.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2432, pruned_loss=0.04942, over 955519.72 frames. ], batch size: 33, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 17:02:00,228 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 17:02:21,394 INFO [finetune.py:976] (3/7) Epoch 22, batch 2450, loss[loss=0.1641, simple_loss=0.2326, pruned_loss=0.04781, over 4777.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.24, pruned_loss=0.04865, over 954900.74 frames. ], batch size: 28, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:02:27,068 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-27 17:02:38,157 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7958, 1.4802, 1.3519, 1.6026, 2.0046, 1.6311, 1.4349, 1.2276], + device='cuda:3'), covar=tensor([0.1676, 0.1697, 0.1990, 0.1597, 0.0969, 0.1783, 0.2405, 0.2607], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0310, 0.0349, 0.0287, 0.0324, 0.0308, 0.0299, 0.0370], + device='cuda:3'), out_proj_covar=tensor([6.3771e-05, 6.4205e-05, 7.3435e-05, 5.7879e-05, 6.6751e-05, 6.4558e-05, + 6.2586e-05, 7.8590e-05], device='cuda:3') +2023-04-27 17:02:42,827 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.578e+02 1.755e+02 2.241e+02 3.458e+02, threshold=3.510e+02, percent-clipped=0.0 +2023-04-27 17:02:48,957 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122773.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 17:02:54,349 INFO [finetune.py:976] (3/7) Epoch 22, batch 2500, loss[loss=0.1759, simple_loss=0.2478, pruned_loss=0.05201, over 4768.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2424, pruned_loss=0.04994, over 955614.56 frames. ], batch size: 28, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:03:21,720 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:03:23,559 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122824.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:03:28,387 INFO [finetune.py:976] (3/7) Epoch 22, batch 2550, loss[loss=0.1747, simple_loss=0.2449, pruned_loss=0.05221, over 4787.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2461, pruned_loss=0.05103, over 954689.92 frames. ], batch size: 29, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:03:40,599 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122851.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:03:48,778 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.572e+02 1.988e+02 2.306e+02 4.105e+02, threshold=3.976e+02, percent-clipped=1.0 +2023-04-27 17:03:48,864 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4542, 3.3572, 2.4688, 3.8880, 3.4555, 3.4067, 1.5572, 3.3833], + device='cuda:3'), covar=tensor([0.1907, 0.1431, 0.3420, 0.2384, 0.2792, 0.1855, 0.5314, 0.2603], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0217, 0.0252, 0.0306, 0.0295, 0.0248, 0.0277, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:03:48,987 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-04-27 17:04:00,163 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9670, 2.5089, 1.0664, 1.3569, 2.1322, 1.2186, 3.0994, 1.6181], + device='cuda:3'), covar=tensor([0.0697, 0.0571, 0.0789, 0.1192, 0.0428, 0.0961, 0.0230, 0.0639], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 17:04:00,765 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1051, 2.4890, 0.9650, 1.4599, 1.4748, 1.9095, 1.5806, 0.8571], + device='cuda:3'), covar=tensor([0.1358, 0.1046, 0.1534, 0.1217, 0.1056, 0.0861, 0.1429, 0.1680], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0241, 0.0139, 0.0120, 0.0133, 0.0153, 0.0118, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 17:04:01,376 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122872.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:04:13,222 INFO [finetune.py:976] (3/7) Epoch 22, batch 2600, loss[loss=0.1667, simple_loss=0.2405, pruned_loss=0.04639, over 4896.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2464, pruned_loss=0.05082, over 953866.53 frames. ], batch size: 43, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:04:34,933 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122899.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:05:19,790 INFO [finetune.py:976] (3/7) Epoch 22, batch 2650, loss[loss=0.2031, simple_loss=0.2695, pruned_loss=0.06839, over 4825.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2491, pruned_loss=0.05164, over 954676.83 frames. ], batch size: 30, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:05:19,911 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5368, 1.5154, 0.5429, 1.2426, 1.6044, 1.3748, 1.3032, 1.3944], + device='cuda:3'), covar=tensor([0.0506, 0.0391, 0.0381, 0.0566, 0.0275, 0.0523, 0.0489, 0.0604], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 17:05:20,520 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:06:00,530 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.522e+02 1.838e+02 2.219e+02 4.134e+02, threshold=3.677e+02, percent-clipped=1.0 +2023-04-27 17:06:15,908 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122975.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:06:26,130 INFO [finetune.py:976] (3/7) Epoch 22, batch 2700, loss[loss=0.1624, simple_loss=0.2411, pruned_loss=0.04187, over 4758.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2484, pruned_loss=0.05134, over 952556.12 frames. ], batch size: 54, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:06:47,004 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 17:06:56,566 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123023.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:07:02,525 INFO [finetune.py:976] (3/7) Epoch 22, batch 2750, loss[loss=0.1806, simple_loss=0.2459, pruned_loss=0.05762, over 4312.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2457, pruned_loss=0.05084, over 953514.97 frames. ], batch size: 18, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:07:28,110 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.026e+01 1.558e+02 1.841e+02 2.200e+02 4.189e+02, threshold=3.681e+02, percent-clipped=2.0 +2023-04-27 17:07:41,201 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-27 17:07:53,158 INFO [finetune.py:976] (3/7) Epoch 22, batch 2800, loss[loss=0.1331, simple_loss=0.2042, pruned_loss=0.03095, over 4822.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2417, pruned_loss=0.04913, over 955410.17 frames. ], batch size: 38, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:08:48,190 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:08:53,535 INFO [finetune.py:976] (3/7) Epoch 22, batch 2850, loss[loss=0.2039, simple_loss=0.2495, pruned_loss=0.07919, over 4491.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2397, pruned_loss=0.04856, over 956395.40 frames. ], batch size: 20, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:09:07,565 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4457, 1.1697, 0.4342, 1.1147, 1.1023, 1.3105, 1.2090, 1.1948], + device='cuda:3'), covar=tensor([0.0463, 0.0377, 0.0408, 0.0560, 0.0302, 0.0490, 0.0473, 0.0545], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 17:09:12,898 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.588e+02 1.846e+02 2.327e+02 3.843e+02, threshold=3.691e+02, percent-clipped=1.0 +2023-04-27 17:09:19,396 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123172.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:09:27,498 INFO [finetune.py:976] (3/7) Epoch 22, batch 2900, loss[loss=0.1765, simple_loss=0.2388, pruned_loss=0.0571, over 4837.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.243, pruned_loss=0.05033, over 955402.04 frames. ], batch size: 25, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:09:35,650 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 17:10:09,345 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123228.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:10:11,740 INFO [finetune.py:976] (3/7) Epoch 22, batch 2950, loss[loss=0.2025, simple_loss=0.2746, pruned_loss=0.06521, over 4824.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2455, pruned_loss=0.05052, over 955829.78 frames. ], batch size: 40, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:10:20,046 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1168, 2.5687, 0.9796, 1.4398, 2.2563, 1.1526, 3.4586, 1.9490], + device='cuda:3'), covar=tensor([0.0642, 0.0714, 0.0892, 0.1231, 0.0474, 0.1052, 0.0228, 0.0583], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0048, 0.0046, 0.0049, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 17:10:55,130 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.651e+02 2.036e+02 2.473e+02 7.289e+02, threshold=4.071e+02, percent-clipped=4.0 +2023-04-27 17:11:18,370 INFO [finetune.py:976] (3/7) Epoch 22, batch 3000, loss[loss=0.2202, simple_loss=0.2834, pruned_loss=0.07852, over 4871.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2472, pruned_loss=0.05126, over 955597.62 frames. ], batch size: 31, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:11:18,370 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 17:11:20,898 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3960, 3.4147, 2.5643, 3.8833, 3.4737, 3.5033, 1.4383, 3.3988], + device='cuda:3'), covar=tensor([0.2082, 0.1516, 0.3761, 0.2335, 0.2702, 0.1708, 0.5657, 0.2575], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0217, 0.0251, 0.0306, 0.0296, 0.0246, 0.0277, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:11:29,116 INFO [finetune.py:1010] (3/7) Epoch 22, validation: loss=0.1537, simple_loss=0.2227, pruned_loss=0.04237, over 2265189.00 frames. +2023-04-27 17:11:29,117 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 17:11:39,683 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-04-27 17:12:11,382 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123314.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:12:33,517 INFO [finetune.py:976] (3/7) Epoch 22, batch 3050, loss[loss=0.1698, simple_loss=0.2465, pruned_loss=0.04655, over 4860.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2488, pruned_loss=0.05114, over 957228.15 frames. ], batch size: 31, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:13:01,591 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.350e+01 1.609e+02 1.855e+02 2.357e+02 3.763e+02, threshold=3.710e+02, percent-clipped=0.0 +2023-04-27 17:13:14,330 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123375.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:13:23,090 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 17:13:24,029 INFO [finetune.py:976] (3/7) Epoch 22, batch 3100, loss[loss=0.1684, simple_loss=0.2408, pruned_loss=0.04797, over 4801.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2469, pruned_loss=0.05065, over 957798.38 frames. ], batch size: 41, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:14:29,456 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123428.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:14:31,789 INFO [finetune.py:976] (3/7) Epoch 22, batch 3150, loss[loss=0.176, simple_loss=0.2445, pruned_loss=0.05372, over 4906.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2436, pruned_loss=0.04984, over 957449.53 frames. ], batch size: 35, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:15:16,033 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.572e+02 1.779e+02 2.164e+02 3.617e+02, threshold=3.558e+02, percent-clipped=0.0 +2023-04-27 17:15:33,942 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 17:15:38,628 INFO [finetune.py:976] (3/7) Epoch 22, batch 3200, loss[loss=0.1642, simple_loss=0.2321, pruned_loss=0.0481, over 4900.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.24, pruned_loss=0.04851, over 957075.53 frames. ], batch size: 35, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:15:48,880 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123489.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:15:56,408 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7603, 1.5199, 1.7264, 2.0825, 2.0849, 1.6762, 1.3423, 1.7284], + device='cuda:3'), covar=tensor([0.0648, 0.1143, 0.0693, 0.0475, 0.0546, 0.0690, 0.0724, 0.0566], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0199, 0.0182, 0.0173, 0.0175, 0.0178, 0.0151, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 17:16:40,547 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-27 17:16:43,112 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123528.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:16:51,256 INFO [finetune.py:976] (3/7) Epoch 22, batch 3250, loss[loss=0.1491, simple_loss=0.2279, pruned_loss=0.03514, over 4866.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2407, pruned_loss=0.04932, over 956907.62 frames. ], batch size: 34, lr: 3.14e-03, grad_scale: 64.0 +2023-04-27 17:17:26,395 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.572e+02 1.866e+02 2.251e+02 4.300e+02, threshold=3.733e+02, percent-clipped=4.0 +2023-04-27 17:17:40,416 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123576.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:17:44,045 INFO [finetune.py:976] (3/7) Epoch 22, batch 3300, loss[loss=0.1423, simple_loss=0.2401, pruned_loss=0.02222, over 4898.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2452, pruned_loss=0.0511, over 957088.36 frames. ], batch size: 35, lr: 3.14e-03, grad_scale: 64.0 +2023-04-27 17:17:44,950 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-27 17:17:54,168 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-04-27 17:17:58,263 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0524, 1.0313, 1.2501, 1.2010, 1.0316, 0.9549, 1.0133, 0.5321], + device='cuda:3'), covar=tensor([0.0551, 0.0583, 0.0439, 0.0597, 0.0823, 0.1264, 0.0513, 0.0656], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0066, 0.0067, 0.0075, 0.0095, 0.0072, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 17:18:17,762 INFO [finetune.py:976] (3/7) Epoch 22, batch 3350, loss[loss=0.1899, simple_loss=0.2457, pruned_loss=0.06711, over 4132.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2456, pruned_loss=0.05059, over 952896.38 frames. ], batch size: 65, lr: 3.14e-03, grad_scale: 64.0 +2023-04-27 17:18:28,300 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2762, 1.7041, 2.1006, 2.6962, 2.1045, 1.6752, 1.5631, 2.0531], + device='cuda:3'), covar=tensor([0.3007, 0.3145, 0.1628, 0.1998, 0.2618, 0.2574, 0.3805, 0.1936], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0245, 0.0226, 0.0315, 0.0221, 0.0233, 0.0228, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 17:18:40,220 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.630e+02 1.835e+02 2.174e+02 3.522e+02, threshold=3.670e+02, percent-clipped=0.0 +2023-04-27 17:18:43,918 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123670.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:18:51,586 INFO [finetune.py:976] (3/7) Epoch 22, batch 3400, loss[loss=0.184, simple_loss=0.2617, pruned_loss=0.05317, over 4809.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2472, pruned_loss=0.0514, over 954426.54 frames. ], batch size: 41, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:18:57,286 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-04-27 17:19:25,131 INFO [finetune.py:976] (3/7) Epoch 22, batch 3450, loss[loss=0.1804, simple_loss=0.2603, pruned_loss=0.05021, over 4923.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2466, pruned_loss=0.05067, over 954951.73 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:19:57,044 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.610e+02 1.907e+02 2.404e+02 3.474e+02, threshold=3.815e+02, percent-clipped=0.0 +2023-04-27 17:20:21,186 INFO [finetune.py:976] (3/7) Epoch 22, batch 3500, loss[loss=0.1423, simple_loss=0.2087, pruned_loss=0.03798, over 4831.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2436, pruned_loss=0.04981, over 954751.19 frames. ], batch size: 38, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:20:28,352 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:20:31,441 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3597, 2.0969, 2.3672, 2.6780, 2.6486, 2.2113, 1.9548, 2.2421], + device='cuda:3'), covar=tensor([0.0798, 0.0958, 0.0624, 0.0563, 0.0575, 0.0802, 0.0777, 0.0577], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0202, 0.0184, 0.0176, 0.0178, 0.0181, 0.0153, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 17:21:11,880 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5305, 1.8204, 1.9241, 2.0826, 1.9053, 1.9315, 1.9752, 1.9340], + device='cuda:3'), covar=tensor([0.3967, 0.5676, 0.4627, 0.4285, 0.5411, 0.7623, 0.5197, 0.4946], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0376, 0.0328, 0.0341, 0.0349, 0.0397, 0.0359, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:21:22,864 INFO [finetune.py:976] (3/7) Epoch 22, batch 3550, loss[loss=0.1629, simple_loss=0.2344, pruned_loss=0.0457, over 4857.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2423, pruned_loss=0.04992, over 955619.81 frames. ], batch size: 47, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:21:43,412 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.536e+02 1.822e+02 2.195e+02 3.918e+02, threshold=3.645e+02, percent-clipped=1.0 +2023-04-27 17:21:49,374 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123872.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:21:56,775 INFO [finetune.py:976] (3/7) Epoch 22, batch 3600, loss[loss=0.1901, simple_loss=0.2594, pruned_loss=0.06039, over 4855.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2388, pruned_loss=0.0485, over 955329.21 frames. ], batch size: 44, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:22:05,381 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.30 vs. limit=5.0 +2023-04-27 17:22:30,655 INFO [finetune.py:976] (3/7) Epoch 22, batch 3650, loss[loss=0.1863, simple_loss=0.267, pruned_loss=0.0528, over 4807.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2411, pruned_loss=0.04942, over 953822.54 frames. ], batch size: 45, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:22:31,423 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:22:53,543 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-04-27 17:23:14,431 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.652e+02 1.968e+02 2.235e+02 6.540e+02, threshold=3.937e+02, percent-clipped=5.0 +2023-04-27 17:23:23,358 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123970.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:23:34,655 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4076, 0.9813, 0.3730, 1.1389, 1.0491, 1.2866, 1.2161, 1.1967], + device='cuda:3'), covar=tensor([0.0514, 0.0428, 0.0440, 0.0582, 0.0331, 0.0548, 0.0509, 0.0600], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 17:23:37,656 INFO [finetune.py:976] (3/7) Epoch 22, batch 3700, loss[loss=0.2129, simple_loss=0.2822, pruned_loss=0.07185, over 4765.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2439, pruned_loss=0.04997, over 953666.35 frames. ], batch size: 59, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:24:22,551 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:24:29,785 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8215, 3.7881, 2.6722, 4.3645, 3.8026, 3.7668, 1.5325, 3.7759], + device='cuda:3'), covar=tensor([0.1652, 0.1120, 0.3110, 0.1623, 0.2845, 0.1782, 0.5599, 0.2194], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0218, 0.0254, 0.0307, 0.0298, 0.0247, 0.0278, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:24:44,200 INFO [finetune.py:976] (3/7) Epoch 22, batch 3750, loss[loss=0.2053, simple_loss=0.2721, pruned_loss=0.06921, over 4732.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2452, pruned_loss=0.05075, over 951460.15 frames. ], batch size: 54, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:24:52,271 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:24:59,627 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0424, 1.7169, 2.1703, 2.4133, 2.0728, 1.9529, 2.0568, 1.9876], + device='cuda:3'), covar=tensor([0.4423, 0.6893, 0.6842, 0.5447, 0.5685, 0.8086, 0.8218, 1.0133], + device='cuda:3'), in_proj_covar=tensor([0.0430, 0.0413, 0.0509, 0.0504, 0.0460, 0.0489, 0.0495, 0.0505], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 17:25:08,744 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124062.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:25:09,866 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.543e+02 1.791e+02 2.127e+02 4.626e+02, threshold=3.581e+02, percent-clipped=2.0 +2023-04-27 17:25:16,254 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 17:25:22,233 INFO [finetune.py:976] (3/7) Epoch 22, batch 3800, loss[loss=0.1641, simple_loss=0.2274, pruned_loss=0.05041, over 4723.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2465, pruned_loss=0.05118, over 950079.70 frames. ], batch size: 23, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:25:23,527 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124084.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:25:32,954 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:25:44,457 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4001, 1.4894, 1.5560, 1.9576, 1.8278, 2.0404, 1.5367, 4.2359], + device='cuda:3'), covar=tensor([0.0568, 0.0924, 0.0834, 0.1249, 0.0677, 0.0573, 0.0840, 0.0150], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 17:25:49,414 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124123.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:25:56,294 INFO [finetune.py:976] (3/7) Epoch 22, batch 3850, loss[loss=0.1726, simple_loss=0.2405, pruned_loss=0.05233, over 4806.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2453, pruned_loss=0.05036, over 951525.21 frames. ], batch size: 41, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:25:56,353 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124132.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:26:03,958 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3180, 3.3679, 0.8863, 1.8074, 1.8666, 2.4540, 1.9200, 1.0646], + device='cuda:3'), covar=tensor([0.1393, 0.0888, 0.1996, 0.1192, 0.1063, 0.0930, 0.1401, 0.1862], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0241, 0.0138, 0.0120, 0.0133, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 17:26:13,753 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.16 vs. limit=5.0 +2023-04-27 17:26:17,752 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.596e+02 1.902e+02 2.251e+02 4.955e+02, threshold=3.804e+02, percent-clipped=1.0 +2023-04-27 17:26:33,009 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1828, 2.1007, 1.7762, 1.7254, 2.2002, 1.8192, 2.5782, 1.6087], + device='cuda:3'), covar=tensor([0.3424, 0.1932, 0.4505, 0.2756, 0.1654, 0.2270, 0.1566, 0.4463], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0346, 0.0423, 0.0352, 0.0377, 0.0372, 0.0367, 0.0415], + device='cuda:3'), out_proj_covar=tensor([9.9693e-05, 1.0358e-04, 1.2828e-04, 1.0588e-04, 1.1223e-04, 1.1074e-04, + 1.0800e-04, 1.2498e-04], device='cuda:3') +2023-04-27 17:26:35,211 INFO [finetune.py:976] (3/7) Epoch 22, batch 3900, loss[loss=0.1444, simple_loss=0.2141, pruned_loss=0.03737, over 4680.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2426, pruned_loss=0.04986, over 952906.98 frames. ], batch size: 23, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:26:57,626 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 17:27:38,101 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124228.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:27:40,956 INFO [finetune.py:976] (3/7) Epoch 22, batch 3950, loss[loss=0.1909, simple_loss=0.2499, pruned_loss=0.0659, over 4728.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2411, pruned_loss=0.04952, over 953995.30 frames. ], batch size: 54, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:28:07,731 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.810e+01 1.371e+02 1.651e+02 2.165e+02 4.664e+02, threshold=3.302e+02, percent-clipped=1.0 +2023-04-27 17:28:13,207 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:28:19,117 INFO [finetune.py:976] (3/7) Epoch 22, batch 4000, loss[loss=0.1826, simple_loss=0.2477, pruned_loss=0.05871, over 4673.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2389, pruned_loss=0.0488, over 954199.65 frames. ], batch size: 23, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:28:48,506 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3153, 3.3624, 0.9585, 1.7279, 1.7246, 2.5601, 1.8623, 1.0594], + device='cuda:3'), covar=tensor([0.1520, 0.0790, 0.1989, 0.1362, 0.1213, 0.0966, 0.1385, 0.1991], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0241, 0.0138, 0.0120, 0.0133, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 17:29:09,163 INFO [finetune.py:976] (3/7) Epoch 22, batch 4050, loss[loss=0.2137, simple_loss=0.2925, pruned_loss=0.06747, over 4804.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2424, pruned_loss=0.04982, over 955890.81 frames. ], batch size: 45, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:29:09,910 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124333.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:29:38,966 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124348.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 17:29:54,265 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.579e+02 1.902e+02 2.335e+02 4.180e+02, threshold=3.804e+02, percent-clipped=3.0 +2023-04-27 17:30:17,168 INFO [finetune.py:976] (3/7) Epoch 22, batch 4100, loss[loss=0.1499, simple_loss=0.2253, pruned_loss=0.03725, over 4825.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2449, pruned_loss=0.05052, over 954534.45 frames. ], batch size: 47, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:30:34,612 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124392.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:30:47,379 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1934, 1.5531, 1.4699, 1.6995, 1.7180, 1.8508, 1.4833, 3.3864], + device='cuda:3'), covar=tensor([0.0595, 0.0782, 0.0740, 0.1185, 0.0607, 0.0591, 0.0733, 0.0167], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 17:30:58,084 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124409.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 17:31:08,996 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124418.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:31:19,712 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5724, 2.1297, 2.5230, 2.8697, 2.8810, 2.2026, 2.1082, 2.3636], + device='cuda:3'), covar=tensor([0.0767, 0.0988, 0.0630, 0.0536, 0.0546, 0.0843, 0.0702, 0.0541], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0201, 0.0183, 0.0176, 0.0177, 0.0180, 0.0152, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 17:31:27,632 INFO [finetune.py:976] (3/7) Epoch 22, batch 4150, loss[loss=0.1596, simple_loss=0.2415, pruned_loss=0.03881, over 4924.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2444, pruned_loss=0.04952, over 955302.80 frames. ], batch size: 42, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:31:27,747 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:31:32,482 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 17:32:10,276 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.606e+01 1.564e+02 1.862e+02 2.379e+02 7.363e+02, threshold=3.724e+02, percent-clipped=1.0 +2023-04-27 17:32:33,359 INFO [finetune.py:976] (3/7) Epoch 22, batch 4200, loss[loss=0.1918, simple_loss=0.2695, pruned_loss=0.05706, over 4889.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2452, pruned_loss=0.04958, over 956094.75 frames. ], batch size: 36, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:32:36,588 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8904, 1.1367, 1.4323, 1.5266, 1.5505, 1.6313, 1.4974, 1.4711], + device='cuda:3'), covar=tensor([0.2936, 0.4083, 0.3386, 0.3297, 0.4465, 0.5771, 0.3841, 0.3668], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0375, 0.0328, 0.0343, 0.0350, 0.0398, 0.0360, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:32:45,639 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124493.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:33:09,254 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124528.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:33:11,585 INFO [finetune.py:976] (3/7) Epoch 22, batch 4250, loss[loss=0.1847, simple_loss=0.256, pruned_loss=0.05668, over 4708.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2437, pruned_loss=0.0495, over 956335.95 frames. ], batch size: 54, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:33:13,016 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.99 vs. limit=5.0 +2023-04-27 17:33:19,620 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6377, 1.7113, 0.7708, 1.3775, 1.7800, 1.5096, 1.4353, 1.5139], + device='cuda:3'), covar=tensor([0.0482, 0.0351, 0.0353, 0.0528, 0.0277, 0.0485, 0.0488, 0.0555], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 17:33:33,183 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.562e+02 1.869e+02 2.245e+02 4.302e+02, threshold=3.738e+02, percent-clipped=2.0 +2023-04-27 17:33:40,671 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124576.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:33:44,720 INFO [finetune.py:976] (3/7) Epoch 22, batch 4300, loss[loss=0.1804, simple_loss=0.2442, pruned_loss=0.05826, over 4899.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2409, pruned_loss=0.04886, over 955948.04 frames. ], batch size: 32, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:34:15,443 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124628.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:34:17,829 INFO [finetune.py:976] (3/7) Epoch 22, batch 4350, loss[loss=0.1913, simple_loss=0.2654, pruned_loss=0.05862, over 4933.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2381, pruned_loss=0.04781, over 957820.90 frames. ], batch size: 38, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:34:22,053 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124638.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:34:38,878 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124663.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:34:39,364 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.058e+02 1.483e+02 1.838e+02 2.190e+02 4.498e+02, threshold=3.677e+02, percent-clipped=2.0 +2023-04-27 17:34:51,253 INFO [finetune.py:976] (3/7) Epoch 22, batch 4400, loss[loss=0.1554, simple_loss=0.2336, pruned_loss=0.03856, over 4805.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2412, pruned_loss=0.04874, over 959694.10 frames. ], batch size: 41, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:34:58,043 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:02,327 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124699.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:11,389 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124704.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 17:35:19,395 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9769, 2.3918, 1.9957, 2.2997, 1.7524, 1.9608, 1.9506, 1.5056], + device='cuda:3'), covar=tensor([0.1606, 0.0979, 0.0748, 0.0932, 0.3104, 0.0970, 0.1700, 0.2182], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0298, 0.0216, 0.0274, 0.0314, 0.0255, 0.0248, 0.0261], + device='cuda:3'), out_proj_covar=tensor([1.1316e-04, 1.1794e-04, 8.5127e-05, 1.0814e-04, 1.2686e-04, 1.0072e-04, + 9.9973e-05, 1.0305e-04], device='cuda:3') +2023-04-27 17:35:33,055 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:42,963 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124724.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:53,109 INFO [finetune.py:976] (3/7) Epoch 22, batch 4450, loss[loss=0.2023, simple_loss=0.2862, pruned_loss=0.05924, over 4914.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2446, pruned_loss=0.04958, over 957128.61 frames. ], batch size: 37, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:36:03,832 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124740.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:36:30,399 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.602e+02 1.930e+02 2.289e+02 4.578e+02, threshold=3.860e+02, percent-clipped=3.0 +2023-04-27 17:36:38,224 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:37:00,152 INFO [finetune.py:976] (3/7) Epoch 22, batch 4500, loss[loss=0.1775, simple_loss=0.2689, pruned_loss=0.04304, over 4858.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2458, pruned_loss=0.04983, over 955698.79 frames. ], batch size: 31, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:37:03,926 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124788.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:37:24,990 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7549, 1.3910, 1.4664, 1.3836, 1.8671, 1.4910, 1.2835, 1.3953], + device='cuda:3'), covar=tensor([0.1516, 0.1403, 0.1914, 0.1376, 0.0877, 0.1734, 0.2012, 0.2486], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0312, 0.0353, 0.0290, 0.0327, 0.0308, 0.0301, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.4362e-05, 6.4543e-05, 7.4316e-05, 5.8362e-05, 6.7407e-05, 6.4603e-05, + 6.2894e-05, 7.9452e-05], device='cuda:3') +2023-04-27 17:37:47,335 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:38:07,812 INFO [finetune.py:976] (3/7) Epoch 22, batch 4550, loss[loss=0.225, simple_loss=0.2908, pruned_loss=0.07958, over 4915.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2475, pruned_loss=0.05058, over 956934.94 frames. ], batch size: 38, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:38:28,879 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9313, 2.1402, 1.1961, 1.6656, 2.0800, 1.7606, 1.6884, 1.8534], + device='cuda:3'), covar=tensor([0.0454, 0.0335, 0.0325, 0.0503, 0.0255, 0.0491, 0.0503, 0.0534], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 17:38:49,605 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.528e+02 1.862e+02 2.235e+02 3.504e+02, threshold=3.725e+02, percent-clipped=0.0 +2023-04-27 17:39:00,623 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-27 17:39:13,967 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124881.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:39:14,477 INFO [finetune.py:976] (3/7) Epoch 22, batch 4600, loss[loss=0.1501, simple_loss=0.2198, pruned_loss=0.04017, over 4781.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2463, pruned_loss=0.04977, over 956477.78 frames. ], batch size: 29, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:39:58,736 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5421, 2.0069, 1.9376, 2.1014, 1.9213, 2.0714, 1.9937, 1.9835], + device='cuda:3'), covar=tensor([0.3487, 0.4644, 0.4298, 0.3583, 0.5034, 0.6072, 0.4992, 0.4515], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0373, 0.0324, 0.0339, 0.0348, 0.0393, 0.0357, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:40:00,428 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124928.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:40:03,270 INFO [finetune.py:976] (3/7) Epoch 22, batch 4650, loss[loss=0.1394, simple_loss=0.2127, pruned_loss=0.03305, over 4903.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2428, pruned_loss=0.04855, over 955921.58 frames. ], batch size: 35, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:40:23,405 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.522e+01 1.501e+02 1.822e+02 2.156e+02 6.565e+02, threshold=3.645e+02, percent-clipped=1.0 +2023-04-27 17:40:31,672 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124976.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:40:35,252 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4352, 1.7826, 1.8610, 1.9857, 1.8308, 1.9192, 1.9185, 1.8937], + device='cuda:3'), covar=tensor([0.3809, 0.5085, 0.4192, 0.3889, 0.5374, 0.6754, 0.5273, 0.4743], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0374, 0.0324, 0.0340, 0.0348, 0.0394, 0.0358, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:40:35,292 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.13 vs. limit=5.0 +2023-04-27 17:40:36,852 INFO [finetune.py:976] (3/7) Epoch 22, batch 4700, loss[loss=0.1913, simple_loss=0.2572, pruned_loss=0.0627, over 4831.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2397, pruned_loss=0.04771, over 956984.43 frames. ], batch size: 39, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:40:44,818 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124994.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:40:51,170 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125004.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 17:41:00,877 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125019.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:41:10,030 INFO [finetune.py:976] (3/7) Epoch 22, batch 4750, loss[loss=0.1459, simple_loss=0.2232, pruned_loss=0.03436, over 4849.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2377, pruned_loss=0.04734, over 957111.08 frames. ], batch size: 44, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:41:23,893 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125052.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 17:41:31,619 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.684e+01 1.607e+02 1.875e+02 2.293e+02 4.178e+02, threshold=3.749e+02, percent-clipped=1.0 +2023-04-27 17:41:43,190 INFO [finetune.py:976] (3/7) Epoch 22, batch 4800, loss[loss=0.1704, simple_loss=0.2255, pruned_loss=0.05764, over 4108.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2414, pruned_loss=0.04889, over 956024.21 frames. ], batch size: 17, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:41:48,387 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:42:14,791 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125128.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:42:17,145 INFO [finetune.py:976] (3/7) Epoch 22, batch 4850, loss[loss=0.1768, simple_loss=0.2646, pruned_loss=0.04448, over 4815.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2436, pruned_loss=0.04918, over 956550.67 frames. ], batch size: 51, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:42:20,148 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:42:32,551 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4354, 0.9740, 0.4459, 1.1433, 1.0815, 1.3178, 1.2326, 1.2492], + device='cuda:3'), covar=tensor([0.0509, 0.0402, 0.0390, 0.0533, 0.0300, 0.0502, 0.0471, 0.0546], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 17:42:44,288 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.525e+02 1.734e+02 2.236e+02 4.357e+02, threshold=3.469e+02, percent-clipped=1.0 +2023-04-27 17:42:54,443 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-27 17:43:02,117 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5418, 1.7788, 1.4567, 1.2199, 1.1913, 1.1373, 1.4533, 1.1090], + device='cuda:3'), covar=tensor([0.1650, 0.1287, 0.1460, 0.1657, 0.2373, 0.2033, 0.1035, 0.2099], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0212, 0.0169, 0.0204, 0.0201, 0.0186, 0.0156, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 17:43:03,304 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:43:13,298 INFO [finetune.py:976] (3/7) Epoch 22, batch 4900, loss[loss=0.1308, simple_loss=0.2142, pruned_loss=0.0237, over 4830.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2457, pruned_loss=0.05013, over 957728.85 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:43:18,234 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125189.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:44:14,012 INFO [finetune.py:976] (3/7) Epoch 22, batch 4950, loss[loss=0.1831, simple_loss=0.2581, pruned_loss=0.05405, over 4762.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2464, pruned_loss=0.05007, over 957760.39 frames. ], batch size: 28, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:44:39,158 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1743, 2.6098, 2.3822, 2.5879, 1.9742, 2.3318, 2.5139, 1.9475], + device='cuda:3'), covar=tensor([0.1562, 0.0741, 0.0604, 0.0851, 0.2874, 0.0850, 0.1367, 0.1971], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0298, 0.0215, 0.0273, 0.0313, 0.0254, 0.0247, 0.0260], + device='cuda:3'), out_proj_covar=tensor([1.1286e-04, 1.1778e-04, 8.5041e-05, 1.0787e-04, 1.2658e-04, 1.0051e-04, + 9.9599e-05, 1.0277e-04], device='cuda:3') +2023-04-27 17:44:50,832 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.480e+02 1.752e+02 2.103e+02 4.760e+02, threshold=3.505e+02, percent-clipped=3.0 +2023-04-27 17:45:13,456 INFO [finetune.py:976] (3/7) Epoch 22, batch 5000, loss[loss=0.1214, simple_loss=0.1958, pruned_loss=0.02352, over 4740.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2447, pruned_loss=0.04959, over 958127.66 frames. ], batch size: 27, lr: 3.13e-03, grad_scale: 16.0 +2023-04-27 17:45:32,793 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:45:34,651 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4985, 1.7263, 1.8869, 2.0221, 1.8403, 1.9196, 1.9765, 1.9285], + device='cuda:3'), covar=tensor([0.3846, 0.5470, 0.4101, 0.4037, 0.5118, 0.6668, 0.4957, 0.4654], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0373, 0.0325, 0.0340, 0.0347, 0.0393, 0.0357, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:46:05,808 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:16,483 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9983, 1.6007, 1.6071, 1.8264, 2.1259, 1.7673, 1.5364, 1.4864], + device='cuda:3'), covar=tensor([0.1492, 0.1349, 0.1849, 0.1044, 0.0824, 0.1383, 0.1772, 0.2271], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0307, 0.0347, 0.0284, 0.0322, 0.0303, 0.0296, 0.0369], + device='cuda:3'), out_proj_covar=tensor([6.3432e-05, 6.3447e-05, 7.3040e-05, 5.7129e-05, 6.6391e-05, 6.3534e-05, + 6.1876e-05, 7.8200e-05], device='cuda:3') +2023-04-27 17:46:18,222 INFO [finetune.py:976] (3/7) Epoch 22, batch 5050, loss[loss=0.1763, simple_loss=0.248, pruned_loss=0.05231, over 4832.00 frames. ], tot_loss[loss=0.169, simple_loss=0.241, pruned_loss=0.0485, over 956282.17 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 16.0 +2023-04-27 17:46:32,850 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125342.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:32,954 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 17:46:43,445 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:48,788 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.666e+02 1.966e+02 2.426e+02 4.294e+02, threshold=3.932e+02, percent-clipped=5.0 +2023-04-27 17:46:50,091 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125367.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:55,693 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125376.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:59,781 INFO [finetune.py:976] (3/7) Epoch 22, batch 5100, loss[loss=0.1425, simple_loss=0.2059, pruned_loss=0.03959, over 4730.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2385, pruned_loss=0.04755, over 957257.88 frames. ], batch size: 23, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:47:07,694 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0952, 2.4529, 2.1602, 2.3877, 1.7463, 2.0764, 2.0870, 1.5736], + device='cuda:3'), covar=tensor([0.1863, 0.1112, 0.0779, 0.1073, 0.3311, 0.1213, 0.1705, 0.2518], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0299, 0.0216, 0.0274, 0.0314, 0.0255, 0.0248, 0.0261], + device='cuda:3'), out_proj_covar=tensor([1.1326e-04, 1.1819e-04, 8.5264e-05, 1.0830e-04, 1.2695e-04, 1.0091e-04, + 1.0003e-04, 1.0305e-04], device='cuda:3') +2023-04-27 17:47:24,363 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125417.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 17:47:26,766 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125421.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:47:33,408 INFO [finetune.py:976] (3/7) Epoch 22, batch 5150, loss[loss=0.2192, simple_loss=0.28, pruned_loss=0.07922, over 4932.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2386, pruned_loss=0.04813, over 954987.35 frames. ], batch size: 38, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:47:37,082 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125437.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:02,910 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.720e+01 1.588e+02 1.846e+02 2.174e+02 3.765e+02, threshold=3.692e+02, percent-clipped=0.0 +2023-04-27 17:48:09,685 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125476.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:13,159 INFO [finetune.py:976] (3/7) Epoch 22, batch 5200, loss[loss=0.204, simple_loss=0.2776, pruned_loss=0.06522, over 4901.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2424, pruned_loss=0.04969, over 954349.91 frames. ], batch size: 37, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:48:13,278 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125482.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:14,969 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125484.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:42,435 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125524.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:47,213 INFO [finetune.py:976] (3/7) Epoch 22, batch 5250, loss[loss=0.178, simple_loss=0.2599, pruned_loss=0.04807, over 4821.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2461, pruned_loss=0.05045, over 954282.01 frames. ], batch size: 39, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:48:50,388 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3868, 1.9364, 2.3139, 2.6405, 2.6443, 2.0284, 2.0454, 2.3425], + device='cuda:3'), covar=tensor([0.0749, 0.1069, 0.0622, 0.0541, 0.0520, 0.0862, 0.0651, 0.0545], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0201, 0.0183, 0.0176, 0.0177, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 17:49:08,675 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4245, 2.9003, 1.1109, 1.6137, 2.3861, 1.5145, 4.0358, 1.8847], + device='cuda:3'), covar=tensor([0.0633, 0.0724, 0.0842, 0.1226, 0.0472, 0.0904, 0.0145, 0.0589], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0048, 0.0046, 0.0050, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 17:49:08,691 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7515, 2.0287, 0.9638, 1.5038, 2.0825, 1.5461, 1.5395, 1.6659], + device='cuda:3'), covar=tensor([0.0479, 0.0332, 0.0296, 0.0518, 0.0241, 0.0500, 0.0462, 0.0539], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 17:49:09,789 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.600e+02 1.860e+02 2.252e+02 4.468e+02, threshold=3.721e+02, percent-clipped=2.0 +2023-04-27 17:49:20,593 INFO [finetune.py:976] (3/7) Epoch 22, batch 5300, loss[loss=0.1717, simple_loss=0.2361, pruned_loss=0.05361, over 4762.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2466, pruned_loss=0.05061, over 955631.49 frames. ], batch size: 28, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:49:24,920 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3508, 1.2255, 1.6123, 1.5259, 1.2595, 1.1512, 1.2479, 0.7709], + device='cuda:3'), covar=tensor([0.0512, 0.0700, 0.0358, 0.0693, 0.0846, 0.1051, 0.0548, 0.0606], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0095, 0.0073, 0.0065], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 17:49:34,623 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125603.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:49:54,073 INFO [finetune.py:976] (3/7) Epoch 22, batch 5350, loss[loss=0.1537, simple_loss=0.2284, pruned_loss=0.03949, over 4904.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2473, pruned_loss=0.05058, over 957477.28 frames. ], batch size: 36, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:50:15,140 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125664.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:50:16,104 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.320e+01 1.581e+02 1.868e+02 2.376e+02 4.466e+02, threshold=3.736e+02, percent-clipped=2.0 +2023-04-27 17:50:38,577 INFO [finetune.py:976] (3/7) Epoch 22, batch 5400, loss[loss=0.1521, simple_loss=0.2168, pruned_loss=0.04371, over 4907.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2435, pruned_loss=0.04955, over 955549.03 frames. ], batch size: 36, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:51:11,192 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.09 vs. limit=2.0 +2023-04-27 17:51:19,931 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125712.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 17:51:40,617 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125725.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:51:45,247 INFO [finetune.py:976] (3/7) Epoch 22, batch 5450, loss[loss=0.1443, simple_loss=0.2112, pruned_loss=0.03868, over 4943.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2403, pruned_loss=0.04869, over 954597.90 frames. ], batch size: 38, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:51:45,327 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125732.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:52:24,627 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-27 17:52:28,291 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.173e+01 1.432e+02 1.653e+02 1.978e+02 3.483e+02, threshold=3.306e+02, percent-clipped=0.0 +2023-04-27 17:52:47,950 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125777.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:52:51,411 INFO [finetune.py:976] (3/7) Epoch 22, batch 5500, loss[loss=0.1743, simple_loss=0.2367, pruned_loss=0.05597, over 4906.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2369, pruned_loss=0.04758, over 953101.04 frames. ], batch size: 32, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:52:58,430 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:52:59,639 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125786.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:53:43,240 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8829, 1.2097, 3.2754, 3.0447, 2.8945, 3.2142, 3.1992, 2.8758], + device='cuda:3'), covar=tensor([0.7651, 0.5510, 0.1665, 0.2482, 0.1551, 0.2115, 0.1310, 0.1844], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0305, 0.0405, 0.0406, 0.0347, 0.0409, 0.0313, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 17:53:54,974 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1555, 2.5798, 1.0842, 1.4408, 2.0456, 1.3777, 3.3277, 1.8071], + device='cuda:3'), covar=tensor([0.0699, 0.0587, 0.0779, 0.1265, 0.0492, 0.0972, 0.0226, 0.0603], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 17:53:57,319 INFO [finetune.py:976] (3/7) Epoch 22, batch 5550, loss[loss=0.1855, simple_loss=0.261, pruned_loss=0.05498, over 4857.00 frames. ], tot_loss[loss=0.168, simple_loss=0.239, pruned_loss=0.0485, over 953450.08 frames. ], batch size: 31, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:54:03,351 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125832.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:54:41,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.577e+02 1.939e+02 2.274e+02 3.783e+02, threshold=3.878e+02, percent-clipped=3.0 +2023-04-27 17:54:41,919 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-04-27 17:55:02,079 INFO [finetune.py:976] (3/7) Epoch 22, batch 5600, loss[loss=0.1611, simple_loss=0.2453, pruned_loss=0.03844, over 4783.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2424, pruned_loss=0.0492, over 953704.79 frames. ], batch size: 29, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:56:05,133 INFO [finetune.py:976] (3/7) Epoch 22, batch 5650, loss[loss=0.1838, simple_loss=0.2644, pruned_loss=0.05163, over 4831.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2453, pruned_loss=0.04974, over 954240.20 frames. ], batch size: 47, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:56:31,894 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125959.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:56:35,416 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.900e+01 1.470e+02 1.804e+02 2.167e+02 3.376e+02, threshold=3.608e+02, percent-clipped=0.0 +2023-04-27 17:56:45,404 INFO [finetune.py:976] (3/7) Epoch 22, batch 5700, loss[loss=0.1445, simple_loss=0.2084, pruned_loss=0.04032, over 4207.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2408, pruned_loss=0.04865, over 936322.20 frames. ], batch size: 18, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:56:54,941 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:57:25,280 INFO [finetune.py:976] (3/7) Epoch 23, batch 0, loss[loss=0.2436, simple_loss=0.3017, pruned_loss=0.09279, over 4093.00 frames. ], tot_loss[loss=0.2436, simple_loss=0.3017, pruned_loss=0.09279, over 4093.00 frames. ], batch size: 66, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:57:25,280 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 17:57:31,347 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4776, 3.0260, 0.9237, 1.8438, 1.9131, 2.1852, 1.8896, 1.0319], + device='cuda:3'), covar=tensor([0.1313, 0.1027, 0.1912, 0.1132, 0.0969, 0.0910, 0.1450, 0.1711], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0240, 0.0138, 0.0120, 0.0132, 0.0151, 0.0118, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 17:57:32,382 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8122, 1.6176, 1.8122, 2.1434, 2.1088, 1.6926, 1.5236, 1.9113], + device='cuda:3'), covar=tensor([0.0763, 0.1123, 0.0737, 0.0589, 0.0572, 0.0822, 0.0728, 0.0544], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0202, 0.0184, 0.0175, 0.0177, 0.0181, 0.0152, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 17:57:41,120 INFO [finetune.py:1010] (3/7) Epoch 23, validation: loss=0.1552, simple_loss=0.2246, pruned_loss=0.04292, over 2265189.00 frames. +2023-04-27 17:57:41,120 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 17:57:48,907 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126012.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 17:58:05,554 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:13,331 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3106, 1.9049, 2.1462, 2.3549, 2.1625, 1.7983, 1.3326, 1.8647], + device='cuda:3'), covar=tensor([0.3358, 0.3114, 0.1822, 0.2084, 0.2774, 0.2727, 0.3902, 0.2021], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0244, 0.0226, 0.0313, 0.0220, 0.0233, 0.0227, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 17:58:20,967 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9896, 2.2577, 2.1842, 2.3085, 2.0571, 2.1971, 2.1971, 2.2020], + device='cuda:3'), covar=tensor([0.3292, 0.5980, 0.5249, 0.4508, 0.5448, 0.6438, 0.6724, 0.5917], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0371, 0.0322, 0.0337, 0.0345, 0.0392, 0.0354, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 17:58:23,638 INFO [finetune.py:976] (3/7) Epoch 23, batch 50, loss[loss=0.1536, simple_loss=0.2317, pruned_loss=0.03775, over 4825.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.249, pruned_loss=0.05187, over 216711.54 frames. ], batch size: 33, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:58:24,214 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126059.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:25,268 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126060.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:28,244 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.502e+02 1.744e+02 2.087e+02 3.495e+02, threshold=3.488e+02, percent-clipped=0.0 +2023-04-27 17:58:41,184 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126077.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:42,973 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:48,845 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126081.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:48,924 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3320, 1.7921, 2.1642, 2.6977, 2.1525, 1.7278, 1.5297, 2.0216], + device='cuda:3'), covar=tensor([0.3079, 0.3184, 0.1688, 0.2195, 0.2667, 0.2666, 0.3953, 0.1896], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0243, 0.0226, 0.0313, 0.0220, 0.0232, 0.0227, 0.0183], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 17:59:21,045 INFO [finetune.py:976] (3/7) Epoch 23, batch 100, loss[loss=0.174, simple_loss=0.2369, pruned_loss=0.0555, over 4822.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2429, pruned_loss=0.04989, over 382491.36 frames. ], batch size: 33, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:59:42,929 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126125.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:00:28,539 INFO [finetune.py:976] (3/7) Epoch 23, batch 150, loss[loss=0.1581, simple_loss=0.2303, pruned_loss=0.04293, over 4905.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2376, pruned_loss=0.04907, over 508054.80 frames. ], batch size: 36, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 18:00:38,440 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.878e+01 1.494e+02 1.914e+02 2.300e+02 4.167e+02, threshold=3.828e+02, percent-clipped=5.0 +2023-04-27 18:01:34,971 INFO [finetune.py:976] (3/7) Epoch 23, batch 200, loss[loss=0.1735, simple_loss=0.2474, pruned_loss=0.04978, over 4908.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2355, pruned_loss=0.04749, over 609193.80 frames. ], batch size: 35, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 18:02:19,986 INFO [finetune.py:976] (3/7) Epoch 23, batch 250, loss[loss=0.2328, simple_loss=0.3061, pruned_loss=0.07981, over 4806.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2411, pruned_loss=0.04981, over 686773.65 frames. ], batch size: 51, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:02:20,609 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126259.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:02:24,185 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.438e+01 1.622e+02 1.975e+02 2.331e+02 7.246e+02, threshold=3.950e+02, percent-clipped=3.0 +2023-04-27 18:02:44,418 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-04-27 18:02:51,716 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126307.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:02:53,380 INFO [finetune.py:976] (3/7) Epoch 23, batch 300, loss[loss=0.1684, simple_loss=0.2545, pruned_loss=0.04111, over 4740.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2446, pruned_loss=0.05004, over 745488.85 frames. ], batch size: 54, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:02:55,357 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1162, 1.8346, 2.1095, 2.3648, 2.3619, 2.0020, 1.9133, 2.0732], + device='cuda:3'), covar=tensor([0.0825, 0.1112, 0.0698, 0.0703, 0.0678, 0.0755, 0.0696, 0.0620], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0201, 0.0183, 0.0175, 0.0177, 0.0180, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:03:10,177 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8980, 2.8869, 2.2158, 3.2697, 2.8921, 2.8471, 1.0406, 2.7876], + device='cuda:3'), covar=tensor([0.2094, 0.1584, 0.3485, 0.3020, 0.3216, 0.2213, 0.6241, 0.2948], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0217, 0.0253, 0.0305, 0.0297, 0.0246, 0.0275, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:03:22,883 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126354.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:03:26,333 INFO [finetune.py:976] (3/7) Epoch 23, batch 350, loss[loss=0.185, simple_loss=0.2668, pruned_loss=0.05163, over 4827.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2471, pruned_loss=0.0508, over 791067.77 frames. ], batch size: 49, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:03:28,672 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4175, 1.6748, 1.6248, 1.9634, 1.9740, 1.9576, 1.6916, 4.2840], + device='cuda:3'), covar=tensor([0.0541, 0.0805, 0.0743, 0.1158, 0.0584, 0.0585, 0.0698, 0.0129], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0037, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 18:03:30,385 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.435e+01 1.518e+02 1.849e+02 2.150e+02 3.695e+02, threshold=3.697e+02, percent-clipped=0.0 +2023-04-27 18:03:33,131 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 18:03:39,106 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1965, 1.7892, 2.1085, 2.3332, 2.3149, 1.9603, 1.8268, 2.1058], + device='cuda:3'), covar=tensor([0.0725, 0.1064, 0.0605, 0.0599, 0.0601, 0.0830, 0.0680, 0.0525], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0203, 0.0184, 0.0176, 0.0177, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:03:41,542 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:03:42,137 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126381.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:03:51,934 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1841, 1.3431, 1.5727, 1.7188, 1.6128, 1.7402, 1.6010, 1.5888], + device='cuda:3'), covar=tensor([0.3790, 0.4873, 0.3937, 0.4148, 0.5291, 0.6880, 0.4718, 0.4734], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0324, 0.0339, 0.0347, 0.0394, 0.0357, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:03:59,740 INFO [finetune.py:976] (3/7) Epoch 23, batch 400, loss[loss=0.1432, simple_loss=0.2181, pruned_loss=0.0342, over 4781.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2472, pruned_loss=0.05044, over 828362.02 frames. ], batch size: 26, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:04:30,796 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126429.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:04:37,950 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9075, 1.0703, 1.5186, 1.6558, 1.5997, 1.7000, 1.5328, 1.5630], + device='cuda:3'), covar=tensor([0.3954, 0.5098, 0.4180, 0.3986, 0.5008, 0.6580, 0.4560, 0.4288], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0376, 0.0325, 0.0341, 0.0349, 0.0396, 0.0358, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:04:42,870 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126441.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:04:45,932 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5739, 1.5147, 1.9993, 2.0109, 1.4541, 1.3116, 1.6133, 1.0443], + device='cuda:3'), covar=tensor([0.0536, 0.0643, 0.0330, 0.0543, 0.0747, 0.1149, 0.0576, 0.0568], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0066, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 18:04:48,394 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126450.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:04:54,266 INFO [finetune.py:976] (3/7) Epoch 23, batch 450, loss[loss=0.2525, simple_loss=0.3043, pruned_loss=0.1003, over 4825.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2451, pruned_loss=0.04929, over 858755.70 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:04:58,394 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.511e+02 1.837e+02 2.263e+02 5.457e+02, threshold=3.673e+02, percent-clipped=4.0 +2023-04-27 18:05:03,927 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8691, 2.5358, 1.9790, 1.9573, 1.3749, 1.4204, 2.0205, 1.3595], + device='cuda:3'), covar=tensor([0.1671, 0.1403, 0.1416, 0.1685, 0.2422, 0.1993, 0.0979, 0.2068], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0209, 0.0168, 0.0202, 0.0198, 0.0184, 0.0154, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 18:05:21,888 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-27 18:05:27,628 INFO [finetune.py:976] (3/7) Epoch 23, batch 500, loss[loss=0.1461, simple_loss=0.2206, pruned_loss=0.03579, over 4788.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2428, pruned_loss=0.0488, over 882903.66 frames. ], batch size: 28, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:05:29,450 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126511.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:05:56,783 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5963, 1.4905, 1.9066, 1.8909, 1.4449, 1.3589, 1.5812, 1.0403], + device='cuda:3'), covar=tensor([0.0491, 0.0642, 0.0401, 0.0608, 0.0734, 0.1119, 0.0590, 0.0591], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0066, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 18:06:06,863 INFO [finetune.py:976] (3/7) Epoch 23, batch 550, loss[loss=0.1399, simple_loss=0.2083, pruned_loss=0.03576, over 4937.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2385, pruned_loss=0.0472, over 899358.26 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:06:16,306 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.502e+02 1.815e+02 2.160e+02 5.481e+02, threshold=3.630e+02, percent-clipped=1.0 +2023-04-27 18:06:27,672 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 18:06:47,604 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 18:07:09,784 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.59 vs. limit=5.0 +2023-04-27 18:07:12,072 INFO [finetune.py:976] (3/7) Epoch 23, batch 600, loss[loss=0.1838, simple_loss=0.2624, pruned_loss=0.05265, over 4826.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2401, pruned_loss=0.04841, over 911653.03 frames. ], batch size: 40, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:08:14,787 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126654.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:08:17,791 INFO [finetune.py:976] (3/7) Epoch 23, batch 650, loss[loss=0.1342, simple_loss=0.1968, pruned_loss=0.03584, over 4297.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2416, pruned_loss=0.04825, over 921568.68 frames. ], batch size: 18, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:08:26,640 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.615e+02 1.965e+02 2.363e+02 5.710e+02, threshold=3.929e+02, percent-clipped=5.0 +2023-04-27 18:09:12,204 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2068, 1.8232, 1.9725, 2.3094, 2.3826, 1.8403, 1.8033, 2.0255], + device='cuda:3'), covar=tensor([0.0697, 0.0947, 0.0626, 0.0568, 0.0538, 0.0862, 0.0686, 0.0571], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0201, 0.0184, 0.0175, 0.0176, 0.0181, 0.0152, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:09:20,359 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126702.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:09:24,630 INFO [finetune.py:976] (3/7) Epoch 23, batch 700, loss[loss=0.1694, simple_loss=0.2356, pruned_loss=0.0516, over 4759.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2428, pruned_loss=0.04851, over 929326.45 frames. ], batch size: 27, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:10:03,669 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:10:13,085 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1199, 1.7606, 2.1850, 2.5724, 2.2264, 2.0481, 2.1038, 2.0213], + device='cuda:3'), covar=tensor([0.4450, 0.6093, 0.5987, 0.5654, 0.5418, 0.7670, 0.7627, 0.8283], + device='cuda:3'), in_proj_covar=tensor([0.0434, 0.0417, 0.0510, 0.0508, 0.0463, 0.0494, 0.0500, 0.0510], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:10:19,058 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5578, 3.5442, 0.9343, 1.8307, 1.8590, 2.3932, 1.8372, 1.0214], + device='cuda:3'), covar=tensor([0.1359, 0.0955, 0.2006, 0.1223, 0.1124, 0.1063, 0.1555, 0.2119], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0237, 0.0136, 0.0118, 0.0130, 0.0149, 0.0115, 0.0117], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 18:10:19,571 INFO [finetune.py:976] (3/7) Epoch 23, batch 750, loss[loss=0.1665, simple_loss=0.2483, pruned_loss=0.04238, over 4878.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2435, pruned_loss=0.04872, over 934593.92 frames. ], batch size: 35, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:10:20,343 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9912, 2.2809, 2.2153, 2.3018, 2.0843, 2.2339, 2.3076, 2.1560], + device='cuda:3'), covar=tensor([0.3713, 0.5954, 0.4978, 0.5022, 0.5940, 0.7117, 0.6051, 0.5699], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0374, 0.0324, 0.0339, 0.0346, 0.0394, 0.0356, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:10:23,177 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.166e+01 1.471e+02 1.836e+02 2.076e+02 2.754e+02, threshold=3.672e+02, percent-clipped=0.0 +2023-04-27 18:10:51,753 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126806.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:10:53,513 INFO [finetune.py:976] (3/7) Epoch 23, batch 800, loss[loss=0.1388, simple_loss=0.2138, pruned_loss=0.03189, over 4857.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2423, pruned_loss=0.04807, over 939912.96 frames. ], batch size: 44, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:10:56,111 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5250, 2.4528, 1.9680, 2.2116, 2.5055, 2.0536, 3.1395, 1.8334], + device='cuda:3'), covar=tensor([0.3408, 0.1868, 0.4087, 0.3019, 0.1717, 0.2320, 0.1717, 0.4024], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0345, 0.0421, 0.0347, 0.0375, 0.0369, 0.0366, 0.0415], + device='cuda:3'), out_proj_covar=tensor([9.8944e-05, 1.0312e-04, 1.2755e-04, 1.0425e-04, 1.1138e-04, 1.0994e-04, + 1.0773e-04, 1.2505e-04], device='cuda:3') +2023-04-27 18:10:57,926 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126816.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:10:58,574 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5375, 2.6512, 2.2194, 2.3649, 2.7533, 2.3137, 3.5364, 2.0069], + device='cuda:3'), covar=tensor([0.3512, 0.2178, 0.3789, 0.3159, 0.1743, 0.2303, 0.1603, 0.4101], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0345, 0.0420, 0.0347, 0.0374, 0.0369, 0.0366, 0.0415], + device='cuda:3'), out_proj_covar=tensor([9.8902e-05, 1.0309e-04, 1.2751e-04, 1.0425e-04, 1.1136e-04, 1.0992e-04, + 1.0771e-04, 1.2503e-04], device='cuda:3') +2023-04-27 18:11:27,333 INFO [finetune.py:976] (3/7) Epoch 23, batch 850, loss[loss=0.2017, simple_loss=0.2565, pruned_loss=0.07342, over 4820.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2404, pruned_loss=0.04738, over 942268.31 frames. ], batch size: 40, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:11:30,954 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.555e+01 1.428e+02 1.632e+02 2.009e+02 4.490e+02, threshold=3.264e+02, percent-clipped=1.0 +2023-04-27 18:11:44,273 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126877.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:11:57,519 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 18:11:58,245 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-27 18:12:28,955 INFO [finetune.py:976] (3/7) Epoch 23, batch 900, loss[loss=0.2021, simple_loss=0.2604, pruned_loss=0.07197, over 4827.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2398, pruned_loss=0.04801, over 942571.97 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:13:18,541 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2356, 2.2678, 2.0425, 2.0338, 2.5883, 2.0767, 3.0021, 1.8331], + device='cuda:3'), covar=tensor([0.3571, 0.1872, 0.4103, 0.2681, 0.1291, 0.2379, 0.1286, 0.4229], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0347, 0.0425, 0.0349, 0.0377, 0.0372, 0.0368, 0.0418], + device='cuda:3'), out_proj_covar=tensor([9.9554e-05, 1.0393e-04, 1.2882e-04, 1.0500e-04, 1.1208e-04, 1.1079e-04, + 1.0835e-04, 1.2598e-04], device='cuda:3') +2023-04-27 18:13:36,993 INFO [finetune.py:976] (3/7) Epoch 23, batch 950, loss[loss=0.1785, simple_loss=0.2504, pruned_loss=0.05325, over 4816.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2383, pruned_loss=0.04807, over 946140.36 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:13:40,660 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.507e+02 1.824e+02 2.265e+02 3.979e+02, threshold=3.647e+02, percent-clipped=2.0 +2023-04-27 18:13:42,593 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0143, 2.3933, 2.0331, 2.3427, 1.6206, 2.0089, 2.1115, 1.5452], + device='cuda:3'), covar=tensor([0.1955, 0.1135, 0.0805, 0.1226, 0.3358, 0.1142, 0.1831, 0.2597], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0299, 0.0216, 0.0276, 0.0313, 0.0256, 0.0249, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1365e-04, 1.1849e-04, 8.5100e-05, 1.0909e-04, 1.2660e-04, 1.0131e-04, + 1.0052e-04, 1.0355e-04], device='cuda:3') +2023-04-27 18:14:11,162 INFO [finetune.py:976] (3/7) Epoch 23, batch 1000, loss[loss=0.2142, simple_loss=0.2838, pruned_loss=0.07231, over 4160.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.242, pruned_loss=0.04955, over 947697.38 frames. ], batch size: 65, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:14:19,160 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7961, 1.4990, 1.6865, 2.0770, 2.1277, 1.5339, 1.3378, 1.8169], + device='cuda:3'), covar=tensor([0.0762, 0.1190, 0.0760, 0.0565, 0.0576, 0.0851, 0.0750, 0.0575], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0203, 0.0185, 0.0176, 0.0177, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:14:28,622 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:14:45,070 INFO [finetune.py:976] (3/7) Epoch 23, batch 1050, loss[loss=0.2088, simple_loss=0.2856, pruned_loss=0.06594, over 4903.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2455, pruned_loss=0.05006, over 950081.25 frames. ], batch size: 43, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:14:48,722 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.212e+02 1.574e+02 1.858e+02 2.309e+02 5.197e+02, threshold=3.716e+02, percent-clipped=5.0 +2023-04-27 18:15:00,298 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127084.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:15:04,132 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-27 18:15:16,219 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127106.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:15:18,458 INFO [finetune.py:976] (3/7) Epoch 23, batch 1100, loss[loss=0.1784, simple_loss=0.2499, pruned_loss=0.05343, over 4855.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.246, pruned_loss=0.05007, over 952928.61 frames. ], batch size: 31, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:15:20,821 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4285, 3.3887, 2.4866, 3.9061, 3.4456, 3.4217, 1.3925, 3.4346], + device='cuda:3'), covar=tensor([0.1823, 0.1360, 0.3178, 0.2120, 0.2640, 0.1816, 0.6085, 0.2452], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0217, 0.0251, 0.0304, 0.0295, 0.0244, 0.0273, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:15:38,379 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1623, 2.6383, 2.1511, 2.1828, 1.5458, 1.5719, 2.1947, 1.4805], + device='cuda:3'), covar=tensor([0.1583, 0.1518, 0.1339, 0.1657, 0.2228, 0.1944, 0.0999, 0.1966], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0203, 0.0199, 0.0185, 0.0154, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 18:15:47,875 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127154.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:15:51,853 INFO [finetune.py:976] (3/7) Epoch 23, batch 1150, loss[loss=0.2256, simple_loss=0.2865, pruned_loss=0.08239, over 4370.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2459, pruned_loss=0.05006, over 950684.93 frames. ], batch size: 66, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:15:56,436 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.449e+02 1.782e+02 2.216e+02 4.823e+02, threshold=3.563e+02, percent-clipped=1.0 +2023-04-27 18:16:00,785 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127172.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:16:12,277 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9776, 2.3068, 1.9480, 2.2433, 1.6951, 1.9844, 1.9297, 1.4348], + device='cuda:3'), covar=tensor([0.1867, 0.1267, 0.0823, 0.1131, 0.3330, 0.1250, 0.1897, 0.2872], + device='cuda:3'), in_proj_covar=tensor([0.0281, 0.0297, 0.0213, 0.0274, 0.0310, 0.0254, 0.0246, 0.0259], + device='cuda:3'), out_proj_covar=tensor([1.1243e-04, 1.1746e-04, 8.4056e-05, 1.0796e-04, 1.2531e-04, 1.0024e-04, + 9.9372e-05, 1.0244e-04], device='cuda:3') +2023-04-27 18:16:25,332 INFO [finetune.py:976] (3/7) Epoch 23, batch 1200, loss[loss=0.1932, simple_loss=0.253, pruned_loss=0.06672, over 4893.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2455, pruned_loss=0.05058, over 952581.43 frames. ], batch size: 43, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:16:40,589 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.38 vs. limit=5.0 +2023-04-27 18:16:49,062 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4160, 1.9123, 2.3149, 2.8470, 2.2732, 1.9085, 1.8060, 2.1709], + device='cuda:3'), covar=tensor([0.3237, 0.3252, 0.1692, 0.2541, 0.2795, 0.2785, 0.3963, 0.2184], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0244, 0.0227, 0.0315, 0.0220, 0.0233, 0.0227, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 18:16:52,057 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127249.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:17:03,697 INFO [finetune.py:976] (3/7) Epoch 23, batch 1250, loss[loss=0.1491, simple_loss=0.2093, pruned_loss=0.04445, over 4385.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2426, pruned_loss=0.04975, over 953453.61 frames. ], batch size: 19, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:17:14,001 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.661e+01 1.481e+02 1.752e+02 2.149e+02 4.730e+02, threshold=3.504e+02, percent-clipped=1.0 +2023-04-27 18:17:26,675 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2016, 2.7947, 2.3112, 2.7337, 1.8953, 2.3671, 2.6332, 1.8351], + device='cuda:3'), covar=tensor([0.2050, 0.1147, 0.0863, 0.1177, 0.3342, 0.1314, 0.1820, 0.2639], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0298, 0.0214, 0.0275, 0.0312, 0.0255, 0.0247, 0.0261], + device='cuda:3'), out_proj_covar=tensor([1.1295e-04, 1.1800e-04, 8.4409e-05, 1.0854e-04, 1.2595e-04, 1.0067e-04, + 9.9842e-05, 1.0289e-04], device='cuda:3') +2023-04-27 18:18:08,735 INFO [finetune.py:976] (3/7) Epoch 23, batch 1300, loss[loss=0.1748, simple_loss=0.2454, pruned_loss=0.05212, over 4904.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2397, pruned_loss=0.04849, over 952980.48 frames. ], batch size: 35, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:18:09,508 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127310.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:19:13,548 INFO [finetune.py:976] (3/7) Epoch 23, batch 1350, loss[loss=0.1693, simple_loss=0.2406, pruned_loss=0.04898, over 4835.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2403, pruned_loss=0.04885, over 953827.43 frames. ], batch size: 30, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:19:23,394 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.776e+01 1.488e+02 1.772e+02 2.253e+02 3.450e+02, threshold=3.544e+02, percent-clipped=0.0 +2023-04-27 18:20:19,976 INFO [finetune.py:976] (3/7) Epoch 23, batch 1400, loss[loss=0.2075, simple_loss=0.274, pruned_loss=0.07048, over 4891.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2427, pruned_loss=0.04937, over 953621.79 frames. ], batch size: 32, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:20:43,110 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127427.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 18:21:22,033 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3956, 1.9615, 2.1650, 2.6844, 2.7053, 2.1295, 1.9805, 2.3620], + device='cuda:3'), covar=tensor([0.0729, 0.0992, 0.0665, 0.0550, 0.0555, 0.0839, 0.0690, 0.0513], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0200, 0.0183, 0.0173, 0.0176, 0.0179, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:21:24,264 INFO [finetune.py:976] (3/7) Epoch 23, batch 1450, loss[loss=0.1949, simple_loss=0.2751, pruned_loss=0.05733, over 4813.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2433, pruned_loss=0.04902, over 953452.08 frames. ], batch size: 40, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:21:34,073 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.543e+02 1.916e+02 2.299e+02 4.494e+02, threshold=3.833e+02, percent-clipped=8.0 +2023-04-27 18:21:43,911 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127472.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:22:06,032 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127488.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 18:22:09,974 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 18:22:29,775 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-04-27 18:22:29,896 INFO [finetune.py:976] (3/7) Epoch 23, batch 1500, loss[loss=0.1928, simple_loss=0.265, pruned_loss=0.06036, over 4912.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2461, pruned_loss=0.04993, over 955674.22 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:22:37,633 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:22:43,219 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127529.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:23:19,694 INFO [finetune.py:976] (3/7) Epoch 23, batch 1550, loss[loss=0.1881, simple_loss=0.2659, pruned_loss=0.05515, over 4909.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2455, pruned_loss=0.04958, over 954003.56 frames. ], batch size: 46, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:23:23,839 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.538e+02 1.819e+02 2.091e+02 4.341e+02, threshold=3.638e+02, percent-clipped=1.0 +2023-04-27 18:23:41,772 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127590.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:23:50,932 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127605.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:23:53,256 INFO [finetune.py:976] (3/7) Epoch 23, batch 1600, loss[loss=0.1848, simple_loss=0.2528, pruned_loss=0.05844, over 4757.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2423, pruned_loss=0.04865, over 954597.31 frames. ], batch size: 27, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:24:03,040 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-27 18:24:08,544 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-04-27 18:24:26,894 INFO [finetune.py:976] (3/7) Epoch 23, batch 1650, loss[loss=0.1525, simple_loss=0.2265, pruned_loss=0.03928, over 4928.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2406, pruned_loss=0.04875, over 956442.95 frames. ], batch size: 38, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:24:28,860 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127662.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:24:31,045 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.420e+02 1.883e+02 2.201e+02 3.441e+02, threshold=3.766e+02, percent-clipped=0.0 +2023-04-27 18:25:00,769 INFO [finetune.py:976] (3/7) Epoch 23, batch 1700, loss[loss=0.1346, simple_loss=0.2102, pruned_loss=0.02948, over 4762.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2383, pruned_loss=0.04798, over 957105.07 frames. ], batch size: 27, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:25:10,323 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127723.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:25:22,255 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 18:25:34,585 INFO [finetune.py:976] (3/7) Epoch 23, batch 1750, loss[loss=0.2356, simple_loss=0.3034, pruned_loss=0.08388, over 4802.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2388, pruned_loss=0.04766, over 957450.41 frames. ], batch size: 41, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:25:38,237 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.809e+01 1.624e+02 1.849e+02 2.197e+02 5.063e+02, threshold=3.698e+02, percent-clipped=3.0 +2023-04-27 18:25:48,411 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127779.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:25:50,831 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127783.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 18:26:13,072 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9606, 1.6195, 1.4904, 1.7361, 2.1322, 1.7225, 1.4465, 1.3743], + device='cuda:3'), covar=tensor([0.1306, 0.1416, 0.1840, 0.1163, 0.0744, 0.1801, 0.2007, 0.2389], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0309, 0.0349, 0.0286, 0.0324, 0.0307, 0.0297, 0.0370], + device='cuda:3'), out_proj_covar=tensor([6.3795e-05, 6.3845e-05, 7.3441e-05, 5.7648e-05, 6.6861e-05, 6.4387e-05, + 6.1976e-05, 7.8558e-05], device='cuda:3') +2023-04-27 18:26:14,171 INFO [finetune.py:976] (3/7) Epoch 23, batch 1800, loss[loss=0.146, simple_loss=0.2156, pruned_loss=0.0382, over 4755.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2418, pruned_loss=0.04767, over 958633.33 frames. ], batch size: 26, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:26:14,936 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0669, 1.4678, 1.9251, 2.1039, 1.8915, 1.4757, 1.0133, 1.5692], + device='cuda:3'), covar=tensor([0.2924, 0.3026, 0.1546, 0.2144, 0.2393, 0.2579, 0.4276, 0.2060], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0246, 0.0229, 0.0317, 0.0221, 0.0235, 0.0228, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 18:26:57,464 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127840.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:27:09,582 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:27:21,426 INFO [finetune.py:976] (3/7) Epoch 23, batch 1850, loss[loss=0.203, simple_loss=0.28, pruned_loss=0.06294, over 4746.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2441, pruned_loss=0.04862, over 958307.23 frames. ], batch size: 54, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:27:31,374 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.632e+02 1.910e+02 2.193e+02 3.676e+02, threshold=3.820e+02, percent-clipped=0.0 +2023-04-27 18:27:39,320 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2038, 1.4534, 1.3402, 1.6626, 1.6224, 1.6529, 1.3698, 2.9653], + device='cuda:3'), covar=tensor([0.0628, 0.0799, 0.0768, 0.1166, 0.0598, 0.0556, 0.0760, 0.0183], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 18:27:40,538 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127872.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:27:49,863 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127885.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:27:58,798 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1127, 1.8105, 2.0315, 2.3869, 2.4166, 1.9417, 1.7281, 2.1216], + device='cuda:3'), covar=tensor([0.0792, 0.1064, 0.0698, 0.0620, 0.0639, 0.0860, 0.0756, 0.0593], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0201, 0.0184, 0.0174, 0.0177, 0.0181, 0.0151, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:28:04,013 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127905.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:28:06,383 INFO [finetune.py:976] (3/7) Epoch 23, batch 1900, loss[loss=0.1821, simple_loss=0.2447, pruned_loss=0.05977, over 4786.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2459, pruned_loss=0.04946, over 957850.75 frames. ], batch size: 25, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:28:07,691 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127911.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:28:22,807 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 18:28:31,633 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2328, 1.6259, 2.0052, 2.4042, 2.0448, 1.5866, 1.1533, 1.8026], + device='cuda:3'), covar=tensor([0.3018, 0.3192, 0.1689, 0.2182, 0.2615, 0.2696, 0.4122, 0.2004], + device='cuda:3'), in_proj_covar=tensor([0.0295, 0.0247, 0.0230, 0.0318, 0.0222, 0.0236, 0.0229, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 18:28:32,847 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:28:57,592 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127953.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:29:07,612 INFO [finetune.py:976] (3/7) Epoch 23, batch 1950, loss[loss=0.1595, simple_loss=0.2398, pruned_loss=0.03959, over 4259.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2443, pruned_loss=0.04874, over 957002.48 frames. ], batch size: 66, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:29:16,663 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.493e+02 1.855e+02 2.257e+02 4.055e+02, threshold=3.710e+02, percent-clipped=1.0 +2023-04-27 18:29:26,336 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-27 18:29:29,539 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 18:30:02,324 INFO [finetune.py:976] (3/7) Epoch 23, batch 2000, loss[loss=0.1377, simple_loss=0.213, pruned_loss=0.03118, over 4828.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2419, pruned_loss=0.04802, over 956733.50 frames. ], batch size: 33, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:30:02,429 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3082, 1.8164, 1.6400, 2.0074, 2.1127, 2.0603, 1.6484, 4.1860], + device='cuda:3'), covar=tensor([0.0536, 0.0781, 0.0736, 0.1132, 0.0558, 0.0542, 0.0678, 0.0117], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 18:30:07,970 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:30:35,023 INFO [finetune.py:976] (3/7) Epoch 23, batch 2050, loss[loss=0.1635, simple_loss=0.2364, pruned_loss=0.04527, over 4901.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2389, pruned_loss=0.04711, over 957652.89 frames. ], batch size: 32, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:30:39,640 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.902e+01 1.493e+02 1.754e+02 2.070e+02 4.110e+02, threshold=3.508e+02, percent-clipped=2.0 +2023-04-27 18:30:45,250 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9185, 1.2379, 3.1936, 2.9301, 2.8520, 3.1008, 3.0829, 2.7369], + device='cuda:3'), covar=tensor([0.7219, 0.5051, 0.1498, 0.2444, 0.1485, 0.2013, 0.1444, 0.2187], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0307, 0.0408, 0.0410, 0.0350, 0.0412, 0.0317, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:30:50,718 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128083.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 18:31:08,824 INFO [finetune.py:976] (3/7) Epoch 23, batch 2100, loss[loss=0.1415, simple_loss=0.2255, pruned_loss=0.0287, over 4812.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2394, pruned_loss=0.04772, over 956396.93 frames. ], batch size: 51, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:31:09,619 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 18:31:22,860 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128131.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 18:31:25,338 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128135.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:31:42,238 INFO [finetune.py:976] (3/7) Epoch 23, batch 2150, loss[loss=0.2044, simple_loss=0.279, pruned_loss=0.06488, over 4823.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2427, pruned_loss=0.04883, over 954538.43 frames. ], batch size: 39, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:31:46,839 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.634e+02 1.958e+02 2.421e+02 4.804e+02, threshold=3.916e+02, percent-clipped=1.0 +2023-04-27 18:31:59,154 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128185.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:32:13,546 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:32:15,339 INFO [finetune.py:976] (3/7) Epoch 23, batch 2200, loss[loss=0.178, simple_loss=0.2539, pruned_loss=0.05107, over 4891.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2447, pruned_loss=0.04972, over 955023.28 frames. ], batch size: 35, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:32:39,402 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128228.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:32:40,692 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7777, 2.1867, 2.0293, 2.2102, 2.0073, 2.0840, 2.1172, 2.0151], + device='cuda:3'), covar=tensor([0.3960, 0.5968, 0.5578, 0.4422, 0.5884, 0.7382, 0.6425, 0.6119], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0371, 0.0323, 0.0336, 0.0345, 0.0392, 0.0356, 0.0328], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:32:42,389 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128233.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:33:12,836 INFO [finetune.py:976] (3/7) Epoch 23, batch 2250, loss[loss=0.1941, simple_loss=0.267, pruned_loss=0.06061, over 4750.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2453, pruned_loss=0.04965, over 953637.12 frames. ], batch size: 54, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:33:22,487 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.137e+02 1.532e+02 1.818e+02 2.293e+02 4.621e+02, threshold=3.635e+02, percent-clipped=2.0 +2023-04-27 18:33:57,138 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4818, 1.3638, 0.5690, 1.2126, 1.4119, 1.3576, 1.2797, 1.3400], + device='cuda:3'), covar=tensor([0.0505, 0.0393, 0.0385, 0.0578, 0.0303, 0.0528, 0.0492, 0.0582], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 18:34:26,340 INFO [finetune.py:976] (3/7) Epoch 23, batch 2300, loss[loss=0.1881, simple_loss=0.2647, pruned_loss=0.05577, over 4194.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2453, pruned_loss=0.04942, over 953121.68 frames. ], batch size: 65, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:34:37,302 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128318.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:34:38,438 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3312, 1.2246, 1.5980, 1.5082, 1.2230, 1.1541, 1.3027, 0.7753], + device='cuda:3'), covar=tensor([0.0532, 0.0591, 0.0348, 0.0536, 0.0720, 0.1016, 0.0526, 0.0564], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 18:34:39,030 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:35:34,256 INFO [finetune.py:976] (3/7) Epoch 23, batch 2350, loss[loss=0.1908, simple_loss=0.2507, pruned_loss=0.06542, over 4925.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2433, pruned_loss=0.0489, over 952754.90 frames. ], batch size: 38, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:35:37,972 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.560e+02 1.785e+02 2.201e+02 3.854e+02, threshold=3.569e+02, percent-clipped=2.0 +2023-04-27 18:35:38,685 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128366.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:36:01,144 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128381.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:36:31,442 INFO [finetune.py:976] (3/7) Epoch 23, batch 2400, loss[loss=0.1645, simple_loss=0.2347, pruned_loss=0.04715, over 4921.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2404, pruned_loss=0.0483, over 954396.97 frames. ], batch size: 37, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:36:34,503 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128413.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:37:01,411 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128435.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:37:29,231 INFO [finetune.py:976] (3/7) Epoch 23, batch 2450, loss[loss=0.1955, simple_loss=0.2644, pruned_loss=0.0633, over 4832.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2381, pruned_loss=0.04772, over 956077.62 frames. ], batch size: 40, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:37:35,176 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.697e+01 1.581e+02 1.977e+02 2.297e+02 4.168e+02, threshold=3.955e+02, percent-clipped=2.0 +2023-04-27 18:37:46,131 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7403, 1.6479, 2.0025, 2.1216, 1.5847, 1.3398, 1.7104, 0.9640], + device='cuda:3'), covar=tensor([0.0731, 0.0598, 0.0533, 0.0728, 0.0822, 0.1067, 0.0698, 0.0758], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 18:37:46,710 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128474.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:37:50,237 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5177, 3.7912, 0.8110, 2.1195, 2.0398, 2.4967, 2.2378, 0.9775], + device='cuda:3'), covar=tensor([0.1394, 0.0873, 0.2070, 0.1166, 0.1046, 0.1142, 0.1358, 0.2139], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0239, 0.0137, 0.0120, 0.0132, 0.0151, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 18:37:56,947 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:38:27,878 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128506.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:38:29,613 INFO [finetune.py:976] (3/7) Epoch 23, batch 2500, loss[loss=0.1377, simple_loss=0.2098, pruned_loss=0.03276, over 4689.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2389, pruned_loss=0.04749, over 956563.53 frames. ], batch size: 23, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:38:53,777 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128528.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:39:04,007 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9297, 1.1600, 1.5431, 1.6403, 1.6045, 1.6583, 1.5622, 1.5426], + device='cuda:3'), covar=tensor([0.3797, 0.4716, 0.3875, 0.3787, 0.4833, 0.6629, 0.4141, 0.3942], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0325, 0.0339, 0.0348, 0.0396, 0.0357, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:39:12,179 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1400, 2.6471, 0.9649, 1.5253, 1.9089, 1.2805, 3.4710, 1.8005], + device='cuda:3'), covar=tensor([0.0696, 0.0772, 0.0926, 0.1224, 0.0539, 0.1044, 0.0224, 0.0619], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 18:39:29,770 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128554.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:39:32,696 INFO [finetune.py:976] (3/7) Epoch 23, batch 2550, loss[loss=0.2047, simple_loss=0.2907, pruned_loss=0.05933, over 4179.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2433, pruned_loss=0.049, over 958367.81 frames. ], batch size: 65, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:39:42,963 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.730e+02 1.875e+02 2.334e+02 4.772e+02, threshold=3.751e+02, percent-clipped=1.0 +2023-04-27 18:39:55,799 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128576.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:40:24,870 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5247, 1.3111, 0.5235, 1.2907, 1.3413, 1.3754, 1.3393, 1.3865], + device='cuda:3'), covar=tensor([0.0586, 0.0351, 0.0366, 0.0600, 0.0293, 0.0633, 0.0608, 0.0625], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 18:40:27,907 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5168, 3.1932, 0.9314, 1.8248, 1.8021, 2.4179, 1.8261, 1.0538], + device='cuda:3'), covar=tensor([0.1275, 0.0959, 0.1980, 0.1150, 0.1041, 0.0920, 0.1374, 0.2013], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0239, 0.0137, 0.0119, 0.0131, 0.0151, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 18:40:40,367 INFO [finetune.py:976] (3/7) Epoch 23, batch 2600, loss[loss=0.1949, simple_loss=0.2621, pruned_loss=0.06386, over 4742.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2441, pruned_loss=0.04907, over 957623.54 frames. ], batch size: 59, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:41:08,651 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2156, 2.8962, 2.1904, 2.1609, 1.5827, 1.5714, 2.2673, 1.5580], + device='cuda:3'), covar=tensor([0.1644, 0.1391, 0.1380, 0.1767, 0.2241, 0.1980, 0.0995, 0.2027], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0212, 0.0169, 0.0206, 0.0200, 0.0187, 0.0156, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 18:41:20,230 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1983, 1.8481, 2.3848, 2.6538, 2.3069, 2.2332, 2.3012, 2.1898], + device='cuda:3'), covar=tensor([0.4403, 0.6223, 0.5697, 0.5324, 0.5804, 0.7039, 0.7565, 0.7255], + device='cuda:3'), in_proj_covar=tensor([0.0433, 0.0419, 0.0511, 0.0506, 0.0464, 0.0494, 0.0499, 0.0510], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:41:44,294 INFO [finetune.py:976] (3/7) Epoch 23, batch 2650, loss[loss=0.1616, simple_loss=0.2377, pruned_loss=0.04272, over 4800.00 frames. ], tot_loss[loss=0.171, simple_loss=0.244, pruned_loss=0.04896, over 954292.84 frames. ], batch size: 51, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:41:53,484 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.191e+02 1.651e+02 1.881e+02 2.286e+02 5.160e+02, threshold=3.761e+02, percent-clipped=1.0 +2023-04-27 18:42:11,801 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128676.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:42:55,228 INFO [finetune.py:976] (3/7) Epoch 23, batch 2700, loss[loss=0.144, simple_loss=0.227, pruned_loss=0.03057, over 4758.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2437, pruned_loss=0.04816, over 955858.65 frames. ], batch size: 26, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:43:39,666 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-27 18:44:00,969 INFO [finetune.py:976] (3/7) Epoch 23, batch 2750, loss[loss=0.1268, simple_loss=0.1851, pruned_loss=0.03425, over 4298.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2421, pruned_loss=0.04838, over 955461.29 frames. ], batch size: 18, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:44:04,649 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.255e+01 1.547e+02 1.823e+02 2.138e+02 3.437e+02, threshold=3.646e+02, percent-clipped=0.0 +2023-04-27 18:44:12,832 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128769.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:44:32,567 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128781.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:45:09,448 INFO [finetune.py:976] (3/7) Epoch 23, batch 2800, loss[loss=0.1617, simple_loss=0.2427, pruned_loss=0.0404, over 4908.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2392, pruned_loss=0.04717, over 957713.08 frames. ], batch size: 46, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:45:11,412 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8305, 1.4313, 1.4245, 1.6200, 2.0017, 1.6161, 1.4542, 1.3494], + device='cuda:3'), covar=tensor([0.1398, 0.1559, 0.1578, 0.1258, 0.0722, 0.1641, 0.2030, 0.2632], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0308, 0.0350, 0.0285, 0.0323, 0.0306, 0.0297, 0.0369], + device='cuda:3'), out_proj_covar=tensor([6.3854e-05, 6.3694e-05, 7.3717e-05, 5.7212e-05, 6.6628e-05, 6.4259e-05, + 6.1903e-05, 7.8404e-05], device='cuda:3') +2023-04-27 18:45:55,102 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128842.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:46:17,709 INFO [finetune.py:976] (3/7) Epoch 23, batch 2850, loss[loss=0.1642, simple_loss=0.2404, pruned_loss=0.04402, over 4743.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2371, pruned_loss=0.0467, over 956480.01 frames. ], batch size: 54, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:46:22,645 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.467e+02 1.826e+02 2.202e+02 4.194e+02, threshold=3.652e+02, percent-clipped=1.0 +2023-04-27 18:46:50,199 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128890.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 18:47:11,864 INFO [finetune.py:976] (3/7) Epoch 23, batch 2900, loss[loss=0.1714, simple_loss=0.2465, pruned_loss=0.04814, over 4893.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2398, pruned_loss=0.04724, over 956819.36 frames. ], batch size: 43, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:47:17,542 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 18:47:56,524 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128951.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 18:48:04,747 INFO [finetune.py:976] (3/7) Epoch 23, batch 2950, loss[loss=0.2034, simple_loss=0.2879, pruned_loss=0.05949, over 4803.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.243, pruned_loss=0.04792, over 957372.19 frames. ], batch size: 41, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:48:10,254 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.058e+02 1.560e+02 1.806e+02 2.158e+02 5.289e+02, threshold=3.612e+02, percent-clipped=5.0 +2023-04-27 18:48:22,803 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128976.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:48:54,312 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0035, 2.3224, 2.0876, 2.2825, 1.8193, 2.0709, 1.9693, 1.7066], + device='cuda:3'), covar=tensor([0.1394, 0.1160, 0.0799, 0.0960, 0.2622, 0.1041, 0.1408, 0.1931], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0299, 0.0213, 0.0274, 0.0314, 0.0255, 0.0247, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1329e-04, 1.1842e-04, 8.4084e-05, 1.0815e-04, 1.2667e-04, 1.0087e-04, + 9.9863e-05, 1.0347e-04], device='cuda:3') +2023-04-27 18:49:01,105 INFO [finetune.py:976] (3/7) Epoch 23, batch 3000, loss[loss=0.2029, simple_loss=0.2745, pruned_loss=0.06567, over 4011.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2446, pruned_loss=0.04877, over 955006.93 frames. ], batch size: 65, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:49:01,105 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 18:49:07,550 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3452, 3.4220, 2.5724, 3.8358, 3.4906, 3.3662, 1.4344, 3.4253], + device='cuda:3'), covar=tensor([0.1580, 0.1397, 0.3136, 0.2210, 0.2610, 0.1704, 0.5442, 0.2204], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0220, 0.0254, 0.0307, 0.0298, 0.0247, 0.0275, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 18:49:07,809 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5075, 1.7225, 1.6247, 1.9594, 1.8754, 1.8818, 1.6040, 3.1213], + device='cuda:3'), covar=tensor([0.0517, 0.0725, 0.0670, 0.1057, 0.0527, 0.0389, 0.0672, 0.0197], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 18:49:09,848 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8377, 2.1473, 1.8124, 1.5431, 1.4133, 1.4194, 1.7848, 1.3602], + device='cuda:3'), covar=tensor([0.1649, 0.1371, 0.1447, 0.1758, 0.2303, 0.1833, 0.1021, 0.2049], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0210, 0.0168, 0.0205, 0.0199, 0.0186, 0.0155, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 18:49:17,637 INFO [finetune.py:1010] (3/7) Epoch 23, validation: loss=0.1527, simple_loss=0.2222, pruned_loss=0.04158, over 2265189.00 frames. +2023-04-27 18:49:17,638 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 18:49:23,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6784, 1.0555, 1.6915, 2.1362, 1.7362, 1.6521, 1.7386, 1.6461], + device='cuda:3'), covar=tensor([0.4179, 0.6130, 0.5854, 0.5374, 0.5310, 0.7002, 0.6526, 0.7950], + device='cuda:3'), in_proj_covar=tensor([0.0435, 0.0420, 0.0512, 0.0509, 0.0465, 0.0497, 0.0500, 0.0511], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:49:28,592 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6286, 1.8788, 1.8854, 2.2743, 2.2063, 2.3264, 1.7635, 4.7105], + device='cuda:3'), covar=tensor([0.0510, 0.0783, 0.0746, 0.1140, 0.0574, 0.0424, 0.0703, 0.0095], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 18:49:31,982 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129024.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:50:01,463 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-27 18:50:16,853 INFO [finetune.py:976] (3/7) Epoch 23, batch 3050, loss[loss=0.1678, simple_loss=0.2439, pruned_loss=0.04589, over 4928.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2443, pruned_loss=0.04857, over 952806.41 frames. ], batch size: 42, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:50:25,152 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.682e+02 1.957e+02 2.360e+02 4.958e+02, threshold=3.913e+02, percent-clipped=5.0 +2023-04-27 18:50:32,879 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 18:50:34,488 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129069.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:50:36,841 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8592, 2.1312, 1.4838, 1.6823, 2.3128, 1.7143, 1.7012, 1.8270], + device='cuda:3'), covar=tensor([0.0425, 0.0272, 0.0288, 0.0461, 0.0258, 0.0421, 0.0416, 0.0463], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 18:51:30,731 INFO [finetune.py:976] (3/7) Epoch 23, batch 3100, loss[loss=0.145, simple_loss=0.2175, pruned_loss=0.03621, over 4821.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2431, pruned_loss=0.04834, over 954132.66 frames. ], batch size: 41, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:51:41,684 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129117.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:52:07,896 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129137.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:52:09,169 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8928, 1.6867, 1.7919, 2.3029, 2.2593, 1.7773, 1.4174, 2.0794], + device='cuda:3'), covar=tensor([0.0816, 0.1150, 0.0802, 0.0567, 0.0614, 0.0918, 0.0830, 0.0538], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0202, 0.0184, 0.0172, 0.0177, 0.0180, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 18:52:39,821 INFO [finetune.py:976] (3/7) Epoch 23, batch 3150, loss[loss=0.1317, simple_loss=0.2071, pruned_loss=0.02817, over 4797.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2393, pruned_loss=0.04736, over 953551.57 frames. ], batch size: 29, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:52:50,085 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.492e+02 1.784e+02 2.330e+02 3.707e+02, threshold=3.568e+02, percent-clipped=0.0 +2023-04-27 18:53:46,919 INFO [finetune.py:976] (3/7) Epoch 23, batch 3200, loss[loss=0.1507, simple_loss=0.2247, pruned_loss=0.03841, over 4908.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.237, pruned_loss=0.04726, over 954359.12 frames. ], batch size: 37, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:54:35,399 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7670, 1.7131, 0.8075, 1.5064, 1.7895, 1.6052, 1.5356, 1.6438], + device='cuda:3'), covar=tensor([0.0485, 0.0361, 0.0351, 0.0517, 0.0272, 0.0484, 0.0492, 0.0561], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 18:54:42,093 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129246.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 18:54:55,918 INFO [finetune.py:976] (3/7) Epoch 23, batch 3250, loss[loss=0.1899, simple_loss=0.263, pruned_loss=0.05841, over 4818.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2383, pruned_loss=0.0477, over 954338.53 frames. ], batch size: 33, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:55:06,949 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.754e+01 1.533e+02 1.786e+02 2.239e+02 3.821e+02, threshold=3.572e+02, percent-clipped=2.0 +2023-04-27 18:55:49,000 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.15 vs. limit=5.0 +2023-04-27 18:55:51,329 INFO [finetune.py:976] (3/7) Epoch 23, batch 3300, loss[loss=0.163, simple_loss=0.2586, pruned_loss=0.03369, over 4828.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2423, pruned_loss=0.04868, over 954428.68 frames. ], batch size: 39, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:56:51,618 INFO [finetune.py:976] (3/7) Epoch 23, batch 3350, loss[loss=0.1869, simple_loss=0.2629, pruned_loss=0.05545, over 4825.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2446, pruned_loss=0.04923, over 955366.64 frames. ], batch size: 30, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:57:02,541 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.708e+02 1.958e+02 2.263e+02 4.173e+02, threshold=3.917e+02, percent-clipped=1.0 +2023-04-27 18:57:55,938 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 18:57:58,597 INFO [finetune.py:976] (3/7) Epoch 23, batch 3400, loss[loss=0.1939, simple_loss=0.2565, pruned_loss=0.06568, over 4877.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2455, pruned_loss=0.05007, over 954121.12 frames. ], batch size: 31, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:58:06,879 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129413.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 18:58:40,740 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129437.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:59:06,102 INFO [finetune.py:976] (3/7) Epoch 23, batch 3450, loss[loss=0.1706, simple_loss=0.2294, pruned_loss=0.05591, over 4912.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2454, pruned_loss=0.04954, over 954850.27 frames. ], batch size: 33, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:59:15,875 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.992e+01 1.559e+02 1.867e+02 2.155e+02 3.707e+02, threshold=3.734e+02, percent-clipped=0.0 +2023-04-27 18:59:26,963 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129474.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 18:59:40,159 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129485.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:59:53,573 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3148, 1.6416, 1.7350, 1.8849, 1.7746, 1.8778, 1.8231, 1.8000], + device='cuda:3'), covar=tensor([0.3547, 0.4869, 0.4671, 0.4058, 0.5153, 0.6691, 0.4907, 0.4599], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0376, 0.0328, 0.0339, 0.0349, 0.0396, 0.0358, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:00:07,510 INFO [finetune.py:976] (3/7) Epoch 23, batch 3500, loss[loss=0.1611, simple_loss=0.2262, pruned_loss=0.04796, over 4816.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2429, pruned_loss=0.04889, over 952602.94 frames. ], batch size: 41, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:00:45,375 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-04-27 19:00:46,496 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129546.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:00:49,477 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6605, 2.8165, 2.2783, 2.4717, 2.9125, 2.4264, 3.7196, 2.1095], + device='cuda:3'), covar=tensor([0.3535, 0.2037, 0.4087, 0.3110, 0.1660, 0.2643, 0.1131, 0.3907], + device='cuda:3'), in_proj_covar=tensor([0.0334, 0.0344, 0.0421, 0.0346, 0.0372, 0.0369, 0.0363, 0.0412], + device='cuda:3'), out_proj_covar=tensor([9.8947e-05, 1.0279e-04, 1.2756e-04, 1.0394e-04, 1.1073e-04, 1.1011e-04, + 1.0664e-04, 1.2395e-04], device='cuda:3') +2023-04-27 19:00:59,652 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4938, 2.0898, 2.3157, 2.9410, 2.4872, 2.0312, 2.1121, 2.2527], + device='cuda:3'), covar=tensor([0.2322, 0.2504, 0.1420, 0.1734, 0.2359, 0.2200, 0.3260, 0.1911], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0245, 0.0229, 0.0314, 0.0220, 0.0235, 0.0228, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 19:01:00,755 INFO [finetune.py:976] (3/7) Epoch 23, batch 3550, loss[loss=0.136, simple_loss=0.1939, pruned_loss=0.03906, over 3979.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2401, pruned_loss=0.048, over 953424.56 frames. ], batch size: 17, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:01:08,390 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.192e+01 1.610e+02 1.893e+02 2.308e+02 5.470e+02, threshold=3.785e+02, percent-clipped=3.0 +2023-04-27 19:01:38,375 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129594.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:01:41,320 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4131, 3.1114, 0.9517, 1.7754, 1.8390, 2.2341, 1.8345, 1.0251], + device='cuda:3'), covar=tensor([0.1300, 0.0919, 0.1755, 0.1186, 0.0946, 0.0964, 0.1386, 0.1927], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0239, 0.0137, 0.0120, 0.0132, 0.0151, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 19:01:52,196 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 19:01:57,767 INFO [finetune.py:976] (3/7) Epoch 23, batch 3600, loss[loss=0.1303, simple_loss=0.2033, pruned_loss=0.0286, over 4745.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2377, pruned_loss=0.04744, over 954956.75 frames. ], batch size: 23, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:01:59,706 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129612.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:02:32,670 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4341, 1.8132, 1.6642, 2.2523, 2.5050, 1.9889, 1.9685, 1.7802], + device='cuda:3'), covar=tensor([0.1786, 0.1675, 0.1956, 0.1364, 0.1062, 0.1978, 0.2199, 0.2386], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0311, 0.0353, 0.0288, 0.0327, 0.0310, 0.0302, 0.0373], + device='cuda:3'), out_proj_covar=tensor([6.4887e-05, 6.4207e-05, 7.4274e-05, 5.7820e-05, 6.7219e-05, 6.5080e-05, + 6.2895e-05, 7.9287e-05], device='cuda:3') +2023-04-27 19:02:45,876 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6130, 3.5441, 2.7578, 4.1883, 3.6241, 3.5487, 1.5315, 3.6732], + device='cuda:3'), covar=tensor([0.1971, 0.1424, 0.3820, 0.1724, 0.3047, 0.1914, 0.6114, 0.2558], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0219, 0.0253, 0.0306, 0.0296, 0.0246, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:02:47,465 INFO [finetune.py:976] (3/7) Epoch 23, batch 3650, loss[loss=0.195, simple_loss=0.2612, pruned_loss=0.06445, over 4813.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2411, pruned_loss=0.04871, over 954857.88 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:02:51,892 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.776e+01 1.523e+02 1.795e+02 2.282e+02 4.473e+02, threshold=3.590e+02, percent-clipped=4.0 +2023-04-27 19:03:02,074 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129673.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:03:55,978 INFO [finetune.py:976] (3/7) Epoch 23, batch 3700, loss[loss=0.1924, simple_loss=0.2544, pruned_loss=0.06518, over 4842.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2439, pruned_loss=0.04957, over 955428.74 frames. ], batch size: 30, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:05:03,363 INFO [finetune.py:976] (3/7) Epoch 23, batch 3750, loss[loss=0.1272, simple_loss=0.1961, pruned_loss=0.02915, over 4734.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2456, pruned_loss=0.05025, over 953633.37 frames. ], batch size: 23, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:05:12,865 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.590e+02 1.958e+02 2.395e+02 5.557e+02, threshold=3.915e+02, percent-clipped=3.0 +2023-04-27 19:05:15,281 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129769.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 19:05:45,768 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129791.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:06:08,151 INFO [finetune.py:976] (3/7) Epoch 23, batch 3800, loss[loss=0.158, simple_loss=0.2389, pruned_loss=0.03851, over 4755.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2465, pruned_loss=0.05045, over 954323.05 frames. ], batch size: 27, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:06:19,774 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5990, 3.5612, 1.0188, 1.8679, 2.0634, 2.5031, 1.9358, 1.0142], + device='cuda:3'), covar=tensor([0.1330, 0.0967, 0.2006, 0.1250, 0.1012, 0.1029, 0.1634, 0.2053], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0238, 0.0136, 0.0119, 0.0131, 0.0150, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 19:06:30,200 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7339, 2.0308, 1.6913, 1.3780, 1.2982, 1.3120, 1.7573, 1.1942], + device='cuda:3'), covar=tensor([0.1728, 0.1380, 0.1475, 0.1890, 0.2350, 0.1887, 0.1005, 0.2056], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0212, 0.0170, 0.0206, 0.0200, 0.0187, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 19:06:30,809 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8914, 1.6852, 1.8040, 2.3672, 2.2882, 1.8734, 1.5608, 2.0192], + device='cuda:3'), covar=tensor([0.0900, 0.1156, 0.0859, 0.0522, 0.0629, 0.0792, 0.0767, 0.0562], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0203, 0.0185, 0.0174, 0.0178, 0.0181, 0.0151, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:07:04,479 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:07:13,928 INFO [finetune.py:976] (3/7) Epoch 23, batch 3850, loss[loss=0.1628, simple_loss=0.232, pruned_loss=0.04684, over 4932.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2446, pruned_loss=0.04939, over 955132.94 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:07:24,864 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.706e+02 1.886e+02 2.157e+02 5.437e+02, threshold=3.771e+02, percent-clipped=4.0 +2023-04-27 19:07:57,475 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.9036, 3.8121, 2.7798, 4.4719, 3.9333, 3.8145, 1.6585, 3.9060], + device='cuda:3'), covar=tensor([0.1760, 0.1153, 0.3242, 0.1671, 0.2835, 0.1947, 0.5953, 0.2251], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0218, 0.0254, 0.0306, 0.0296, 0.0246, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:08:15,874 INFO [finetune.py:976] (3/7) Epoch 23, batch 3900, loss[loss=0.1786, simple_loss=0.2599, pruned_loss=0.04868, over 4928.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2418, pruned_loss=0.0488, over 956296.38 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:09:17,305 INFO [finetune.py:976] (3/7) Epoch 23, batch 3950, loss[loss=0.1952, simple_loss=0.2685, pruned_loss=0.061, over 4940.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2381, pruned_loss=0.04743, over 956039.58 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:09:27,939 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.540e+01 1.458e+02 1.749e+02 1.978e+02 4.242e+02, threshold=3.497e+02, percent-clipped=1.0 +2023-04-27 19:09:35,336 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129968.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:09:59,710 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7725, 2.0383, 1.8748, 2.6213, 2.7918, 2.2678, 2.2874, 1.9686], + device='cuda:3'), covar=tensor([0.1806, 0.1784, 0.2194, 0.1651, 0.1198, 0.1976, 0.2539, 0.2091], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0311, 0.0353, 0.0286, 0.0326, 0.0310, 0.0301, 0.0373], + device='cuda:3'), out_proj_covar=tensor([6.4628e-05, 6.4296e-05, 7.4280e-05, 5.7476e-05, 6.7075e-05, 6.4952e-05, + 6.2756e-05, 7.9173e-05], device='cuda:3') +2023-04-27 19:10:16,018 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6938, 3.5842, 2.7349, 4.2399, 3.5921, 3.6195, 1.7291, 3.6633], + device='cuda:3'), covar=tensor([0.1613, 0.1283, 0.3261, 0.1545, 0.3566, 0.1576, 0.4956, 0.2281], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0219, 0.0253, 0.0306, 0.0297, 0.0246, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:10:16,068 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130000.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:10:22,148 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130002.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:10:33,621 INFO [finetune.py:976] (3/7) Epoch 23, batch 4000, loss[loss=0.1466, simple_loss=0.2122, pruned_loss=0.04052, over 4772.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2379, pruned_loss=0.04804, over 954965.17 frames. ], batch size: 54, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:11:11,891 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4338, 0.9882, 0.3550, 1.1078, 1.0502, 1.3092, 1.1931, 1.2071], + device='cuda:3'), covar=tensor([0.0476, 0.0405, 0.0415, 0.0566, 0.0314, 0.0486, 0.0475, 0.0559], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0049, 0.0051], + device='cuda:3') +2023-04-27 19:11:31,638 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 19:11:43,267 INFO [finetune.py:976] (3/7) Epoch 23, batch 4050, loss[loss=0.188, simple_loss=0.2671, pruned_loss=0.0544, over 4914.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2426, pruned_loss=0.05003, over 954885.25 frames. ], batch size: 37, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:11:44,633 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130061.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:11:45,890 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130063.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:11:53,610 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.688e+02 1.937e+02 2.332e+02 4.574e+02, threshold=3.874e+02, percent-clipped=1.0 +2023-04-27 19:11:56,083 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130069.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:12:18,603 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1640, 1.5794, 2.0475, 2.3843, 2.0874, 1.6205, 1.2537, 1.8242], + device='cuda:3'), covar=tensor([0.3219, 0.3064, 0.1723, 0.2120, 0.2410, 0.2673, 0.4161, 0.1883], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0245, 0.0229, 0.0313, 0.0220, 0.0234, 0.0227, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 19:12:52,036 INFO [finetune.py:976] (3/7) Epoch 23, batch 4100, loss[loss=0.1908, simple_loss=0.2733, pruned_loss=0.05415, over 4815.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2463, pruned_loss=0.05101, over 955074.74 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:13:01,825 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130117.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:13:32,161 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130147.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:13:51,937 INFO [finetune.py:976] (3/7) Epoch 23, batch 4150, loss[loss=0.1556, simple_loss=0.2254, pruned_loss=0.04289, over 4745.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2457, pruned_loss=0.05063, over 953864.57 frames. ], batch size: 26, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:14:00,257 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4507, 1.7563, 1.8362, 1.9531, 1.7974, 1.8741, 1.9356, 1.9051], + device='cuda:3'), covar=tensor([0.3520, 0.4793, 0.3929, 0.3563, 0.4902, 0.6469, 0.4523, 0.4101], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0326, 0.0339, 0.0347, 0.0395, 0.0357, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:14:02,699 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.039e+02 1.725e+02 1.972e+02 2.342e+02 5.183e+02, threshold=3.944e+02, percent-clipped=4.0 +2023-04-27 19:14:05,579 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5237, 0.6555, 1.4590, 1.9096, 1.6287, 1.4461, 1.5076, 1.4596], + device='cuda:3'), covar=tensor([0.4184, 0.6335, 0.5373, 0.5391, 0.5143, 0.7014, 0.6687, 0.7416], + device='cuda:3'), in_proj_covar=tensor([0.0435, 0.0418, 0.0512, 0.0508, 0.0463, 0.0498, 0.0501, 0.0512], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:14:48,092 INFO [finetune.py:976] (3/7) Epoch 23, batch 4200, loss[loss=0.1452, simple_loss=0.2168, pruned_loss=0.03683, over 4833.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.246, pruned_loss=0.04982, over 954954.92 frames. ], batch size: 47, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:15:51,653 INFO [finetune.py:976] (3/7) Epoch 23, batch 4250, loss[loss=0.1596, simple_loss=0.221, pruned_loss=0.04909, over 4207.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.244, pruned_loss=0.04921, over 954896.65 frames. ], batch size: 65, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:16:00,483 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 19:16:01,963 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.509e+02 1.751e+02 2.292e+02 3.450e+02, threshold=3.503e+02, percent-clipped=0.0 +2023-04-27 19:16:03,324 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130268.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:16:22,558 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6716, 1.4246, 4.4806, 4.1478, 3.9061, 4.2324, 4.2217, 3.8897], + device='cuda:3'), covar=tensor([0.7355, 0.5879, 0.1134, 0.1909, 0.1081, 0.1535, 0.1365, 0.1678], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0307, 0.0408, 0.0407, 0.0347, 0.0410, 0.0316, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:16:31,554 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6264, 1.7245, 1.5178, 1.1047, 1.2301, 1.2323, 1.4789, 1.1740], + device='cuda:3'), covar=tensor([0.1809, 0.1300, 0.1490, 0.1803, 0.2294, 0.1904, 0.1053, 0.2058], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0213, 0.0170, 0.0206, 0.0201, 0.0187, 0.0157, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 19:16:53,618 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7997, 1.3389, 1.8931, 2.2646, 1.9041, 1.7722, 1.8425, 1.7571], + device='cuda:3'), covar=tensor([0.4215, 0.6537, 0.6162, 0.5293, 0.5594, 0.8176, 0.7473, 0.8756], + device='cuda:3'), in_proj_covar=tensor([0.0434, 0.0417, 0.0511, 0.0506, 0.0462, 0.0496, 0.0499, 0.0511], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:16:55,317 INFO [finetune.py:976] (3/7) Epoch 23, batch 4300, loss[loss=0.1418, simple_loss=0.2224, pruned_loss=0.03059, over 4897.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2406, pruned_loss=0.04818, over 954924.50 frames. ], batch size: 32, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:17:05,108 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130316.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:17:58,668 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:18:05,094 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130358.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:18:05,659 INFO [finetune.py:976] (3/7) Epoch 23, batch 4350, loss[loss=0.1475, simple_loss=0.2212, pruned_loss=0.03688, over 4834.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2371, pruned_loss=0.04698, over 953848.41 frames. ], batch size: 25, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:18:10,119 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.587e+02 1.809e+02 2.203e+02 4.398e+02, threshold=3.619e+02, percent-clipped=3.0 +2023-04-27 19:18:29,067 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:19:02,657 INFO [finetune.py:976] (3/7) Epoch 23, batch 4400, loss[loss=0.1904, simple_loss=0.2602, pruned_loss=0.06033, over 4825.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2393, pruned_loss=0.04809, over 953918.82 frames. ], batch size: 40, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:19:34,787 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:19:45,173 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130438.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:19:55,344 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130445.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:19:56,560 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130447.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:20:16,251 INFO [finetune.py:976] (3/7) Epoch 23, batch 4450, loss[loss=0.1654, simple_loss=0.2296, pruned_loss=0.05061, over 4749.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2418, pruned_loss=0.04813, over 954537.59 frames. ], batch size: 26, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:20:25,829 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.347e+01 1.587e+02 1.803e+02 2.328e+02 4.838e+02, threshold=3.606e+02, percent-clipped=2.0 +2023-04-27 19:21:01,276 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130493.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:21:02,960 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:21:11,854 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130499.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:21:23,655 INFO [finetune.py:976] (3/7) Epoch 23, batch 4500, loss[loss=0.1659, simple_loss=0.2389, pruned_loss=0.04641, over 4864.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2431, pruned_loss=0.04837, over 953805.77 frames. ], batch size: 34, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:22:17,978 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130548.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:22:31,479 INFO [finetune.py:976] (3/7) Epoch 23, batch 4550, loss[loss=0.1713, simple_loss=0.2437, pruned_loss=0.04941, over 4844.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2445, pruned_loss=0.04912, over 954386.85 frames. ], batch size: 44, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:22:42,262 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130565.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:22:42,754 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.614e+02 1.836e+02 2.049e+02 5.292e+02, threshold=3.672e+02, percent-clipped=1.0 +2023-04-27 19:22:42,880 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1182, 2.4304, 2.1006, 2.4225, 1.7891, 1.9929, 2.0986, 1.6686], + device='cuda:3'), covar=tensor([0.1777, 0.1201, 0.0706, 0.0961, 0.3079, 0.1177, 0.1851, 0.2344], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0301, 0.0214, 0.0275, 0.0314, 0.0256, 0.0247, 0.0261], + device='cuda:3'), out_proj_covar=tensor([1.1377e-04, 1.1929e-04, 8.4435e-05, 1.0833e-04, 1.2693e-04, 1.0132e-04, + 9.9453e-05, 1.0319e-04], device='cuda:3') +2023-04-27 19:23:16,573 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-27 19:23:36,958 INFO [finetune.py:976] (3/7) Epoch 23, batch 4600, loss[loss=0.1378, simple_loss=0.2114, pruned_loss=0.03205, over 4753.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2436, pruned_loss=0.04808, over 955068.39 frames. ], batch size: 27, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:23:37,137 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130609.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:23:56,087 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130626.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:24:06,311 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9813, 1.5640, 2.0864, 2.4187, 2.0513, 1.9453, 1.9870, 1.9998], + device='cuda:3'), covar=tensor([0.4386, 0.6263, 0.6016, 0.5384, 0.5798, 0.7663, 0.7757, 0.7433], + device='cuda:3'), in_proj_covar=tensor([0.0436, 0.0417, 0.0511, 0.0506, 0.0463, 0.0496, 0.0500, 0.0511], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:24:37,006 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130656.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:24:38,751 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130658.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:24:39,266 INFO [finetune.py:976] (3/7) Epoch 23, batch 4650, loss[loss=0.1658, simple_loss=0.2423, pruned_loss=0.04463, over 4859.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2419, pruned_loss=0.04788, over 954673.50 frames. ], batch size: 47, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:24:48,331 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.298e+01 1.478e+02 1.669e+02 2.060e+02 3.844e+02, threshold=3.337e+02, percent-clipped=2.0 +2023-04-27 19:25:39,304 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:25:40,551 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130706.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:25:47,204 INFO [finetune.py:976] (3/7) Epoch 23, batch 4700, loss[loss=0.1702, simple_loss=0.2312, pruned_loss=0.0546, over 4708.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2389, pruned_loss=0.04726, over 955669.30 frames. ], batch size: 23, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:25:50,006 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 19:26:22,718 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130740.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:26:33,108 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9744, 1.2498, 4.8810, 4.5414, 4.2701, 4.6340, 4.3590, 4.2939], + device='cuda:3'), covar=tensor([0.7111, 0.6540, 0.1292, 0.2385, 0.1266, 0.1727, 0.1835, 0.2063], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0310, 0.0410, 0.0411, 0.0350, 0.0415, 0.0319, 0.0370], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:26:50,254 INFO [finetune.py:976] (3/7) Epoch 23, batch 4750, loss[loss=0.1374, simple_loss=0.206, pruned_loss=0.03445, over 3980.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2376, pruned_loss=0.04704, over 955348.79 frames. ], batch size: 17, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:27:00,089 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.517e+02 1.838e+02 2.137e+02 3.767e+02, threshold=3.677e+02, percent-clipped=3.0 +2023-04-27 19:27:24,061 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130788.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:27:33,678 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130794.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:27:54,319 INFO [finetune.py:976] (3/7) Epoch 23, batch 4800, loss[loss=0.2065, simple_loss=0.2782, pruned_loss=0.06738, over 4912.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2405, pruned_loss=0.04825, over 954392.90 frames. ], batch size: 36, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:27:55,060 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2777, 1.1941, 1.5237, 1.5109, 1.1650, 1.1193, 1.2482, 0.7064], + device='cuda:3'), covar=tensor([0.0558, 0.0671, 0.0393, 0.0533, 0.0771, 0.1159, 0.0531, 0.0601], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 19:28:54,603 INFO [finetune.py:976] (3/7) Epoch 23, batch 4850, loss[loss=0.1328, simple_loss=0.2139, pruned_loss=0.02585, over 4296.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2431, pruned_loss=0.04892, over 952214.97 frames. ], batch size: 65, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:29:05,299 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.564e+02 1.952e+02 2.266e+02 5.605e+02, threshold=3.905e+02, percent-clipped=3.0 +2023-04-27 19:29:40,483 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 19:29:41,861 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 19:30:00,409 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130904.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:30:02,342 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3360, 1.8800, 5.5971, 5.3305, 4.9513, 5.2925, 5.0378, 4.9150], + device='cuda:3'), covar=tensor([0.6686, 0.5647, 0.0940, 0.1499, 0.0815, 0.1063, 0.0846, 0.1621], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0306, 0.0406, 0.0406, 0.0346, 0.0409, 0.0316, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:30:03,972 INFO [finetune.py:976] (3/7) Epoch 23, batch 4900, loss[loss=0.1711, simple_loss=0.2483, pruned_loss=0.04691, over 4899.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2437, pruned_loss=0.0489, over 954072.46 frames. ], batch size: 37, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:30:23,697 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130921.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:31:11,883 INFO [finetune.py:976] (3/7) Epoch 23, batch 4950, loss[loss=0.1909, simple_loss=0.2647, pruned_loss=0.05851, over 4909.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2446, pruned_loss=0.04919, over 952171.07 frames. ], batch size: 38, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:31:21,968 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.168e+02 1.568e+02 1.814e+02 2.171e+02 6.365e+02, threshold=3.628e+02, percent-clipped=1.0 +2023-04-27 19:32:26,076 INFO [finetune.py:976] (3/7) Epoch 23, batch 5000, loss[loss=0.1623, simple_loss=0.2403, pruned_loss=0.04219, over 4932.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2431, pruned_loss=0.04843, over 951958.62 frames. ], batch size: 38, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:33:10,250 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:33:35,486 INFO [finetune.py:976] (3/7) Epoch 23, batch 5050, loss[loss=0.1279, simple_loss=0.2063, pruned_loss=0.02471, over 4784.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2404, pruned_loss=0.04791, over 952819.19 frames. ], batch size: 26, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:33:41,160 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.462e+02 1.838e+02 2.219e+02 3.974e+02, threshold=3.675e+02, percent-clipped=2.0 +2023-04-27 19:34:07,192 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 19:34:14,397 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:34:14,441 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:34:18,302 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131094.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:34:38,869 INFO [finetune.py:976] (3/7) Epoch 23, batch 5100, loss[loss=0.1334, simple_loss=0.2057, pruned_loss=0.03058, over 3559.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2372, pruned_loss=0.04671, over 952860.99 frames. ], batch size: 15, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:35:11,741 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:35:21,309 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131142.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:35:21,678 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-27 19:35:41,502 INFO [finetune.py:976] (3/7) Epoch 23, batch 5150, loss[loss=0.1503, simple_loss=0.2135, pruned_loss=0.04361, over 4565.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2377, pruned_loss=0.04699, over 954233.74 frames. ], batch size: 20, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:35:51,237 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.590e+02 1.900e+02 2.291e+02 6.669e+02, threshold=3.800e+02, percent-clipped=5.0 +2023-04-27 19:36:02,283 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131174.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:36:42,388 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131204.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:36:45,362 INFO [finetune.py:976] (3/7) Epoch 23, batch 5200, loss[loss=0.1492, simple_loss=0.2288, pruned_loss=0.03475, over 4857.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2413, pruned_loss=0.04827, over 952403.60 frames. ], batch size: 44, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:37:04,410 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131221.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:37:22,636 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131235.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 19:37:43,653 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131252.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:37:52,694 INFO [finetune.py:976] (3/7) Epoch 23, batch 5250, loss[loss=0.179, simple_loss=0.2551, pruned_loss=0.05149, over 4907.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2432, pruned_loss=0.04934, over 950997.60 frames. ], batch size: 42, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:37:57,550 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.647e+02 1.953e+02 2.357e+02 4.289e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-04-27 19:38:04,768 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131269.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:38:06,000 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8742, 1.4590, 1.4908, 1.6852, 2.0738, 1.6560, 1.3564, 1.4216], + device='cuda:3'), covar=tensor([0.1611, 0.1593, 0.2286, 0.1282, 0.0922, 0.1638, 0.2074, 0.2404], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0308, 0.0350, 0.0283, 0.0325, 0.0305, 0.0298, 0.0371], + device='cuda:3'), out_proj_covar=tensor([6.4166e-05, 6.3680e-05, 7.3732e-05, 5.6791e-05, 6.6882e-05, 6.3872e-05, + 6.2093e-05, 7.8840e-05], device='cuda:3') +2023-04-27 19:38:25,539 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 19:38:50,181 INFO [finetune.py:976] (3/7) Epoch 23, batch 5300, loss[loss=0.1203, simple_loss=0.191, pruned_loss=0.02479, over 4388.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.243, pruned_loss=0.04918, over 949991.48 frames. ], batch size: 19, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:39:15,516 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4414, 3.3877, 2.8003, 3.9959, 3.3282, 3.3920, 1.8322, 3.5377], + device='cuda:3'), covar=tensor([0.2013, 0.1510, 0.4260, 0.1662, 0.2730, 0.2062, 0.5049, 0.2352], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0219, 0.0254, 0.0306, 0.0297, 0.0248, 0.0275, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:39:20,870 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 19:39:23,992 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-04-27 19:39:53,023 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6053, 1.8928, 1.9824, 2.1032, 1.9529, 1.9791, 2.0598, 2.0500], + device='cuda:3'), covar=tensor([0.3904, 0.5377, 0.4239, 0.4206, 0.5154, 0.6789, 0.5172, 0.4677], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0327, 0.0338, 0.0348, 0.0392, 0.0356, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:39:56,069 INFO [finetune.py:976] (3/7) Epoch 23, batch 5350, loss[loss=0.1301, simple_loss=0.2013, pruned_loss=0.02944, over 4783.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2426, pruned_loss=0.04888, over 952008.29 frames. ], batch size: 25, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:40:07,008 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.812e+01 1.537e+02 1.790e+02 2.239e+02 4.084e+02, threshold=3.580e+02, percent-clipped=2.0 +2023-04-27 19:41:03,584 INFO [finetune.py:976] (3/7) Epoch 23, batch 5400, loss[loss=0.1851, simple_loss=0.2502, pruned_loss=0.06004, over 4736.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2419, pruned_loss=0.04924, over 952704.19 frames. ], batch size: 59, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:41:03,700 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131409.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:41:12,777 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131416.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:41:26,154 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2124, 1.7230, 2.1293, 2.1766, 2.0938, 1.6868, 1.1668, 1.7130], + device='cuda:3'), covar=tensor([0.3748, 0.3133, 0.1776, 0.2344, 0.2570, 0.2868, 0.4198, 0.2009], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0244, 0.0226, 0.0311, 0.0219, 0.0233, 0.0225, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 19:42:12,945 INFO [finetune.py:976] (3/7) Epoch 23, batch 5450, loss[loss=0.1742, simple_loss=0.2323, pruned_loss=0.05804, over 4898.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2401, pruned_loss=0.04887, over 955208.71 frames. ], batch size: 35, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:42:22,593 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.575e+02 1.987e+02 2.572e+02 5.206e+02, threshold=3.973e+02, percent-clipped=5.0 +2023-04-27 19:42:31,947 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131470.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:42:43,235 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131477.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:43:05,097 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131492.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:43:22,051 INFO [finetune.py:976] (3/7) Epoch 23, batch 5500, loss[loss=0.135, simple_loss=0.2079, pruned_loss=0.03101, over 4750.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2366, pruned_loss=0.04749, over 954462.20 frames. ], batch size: 54, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:43:52,912 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131530.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:44:16,311 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131553.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:44:21,723 INFO [finetune.py:976] (3/7) Epoch 23, batch 5550, loss[loss=0.1916, simple_loss=0.2754, pruned_loss=0.0539, over 4922.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2373, pruned_loss=0.04753, over 956602.50 frames. ], batch size: 42, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:44:32,379 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.039e+02 1.589e+02 1.849e+02 2.302e+02 7.246e+02, threshold=3.698e+02, percent-clipped=1.0 +2023-04-27 19:44:53,148 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9178, 1.2926, 4.9049, 4.5804, 4.2641, 4.7204, 4.3400, 4.3235], + device='cuda:3'), covar=tensor([0.7364, 0.6359, 0.1013, 0.1829, 0.1103, 0.1380, 0.1971, 0.1859], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0307, 0.0407, 0.0408, 0.0347, 0.0410, 0.0316, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:45:01,796 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-04-27 19:45:12,901 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-27 19:45:32,465 INFO [finetune.py:976] (3/7) Epoch 23, batch 5600, loss[loss=0.1721, simple_loss=0.2545, pruned_loss=0.04483, over 4871.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2392, pruned_loss=0.04736, over 954623.81 frames. ], batch size: 34, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:45:34,374 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5173, 1.2732, 1.6572, 1.7493, 1.3959, 1.0993, 1.2876, 0.7023], + device='cuda:3'), covar=tensor([0.0517, 0.0577, 0.0362, 0.0462, 0.0654, 0.1439, 0.0562, 0.0840], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0067, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 19:46:28,599 INFO [finetune.py:976] (3/7) Epoch 23, batch 5650, loss[loss=0.1919, simple_loss=0.2754, pruned_loss=0.05417, over 4927.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2437, pruned_loss=0.04874, over 953231.30 frames. ], batch size: 42, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:46:38,007 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.000e+01 1.549e+02 1.818e+02 2.173e+02 4.178e+02, threshold=3.635e+02, percent-clipped=2.0 +2023-04-27 19:46:38,124 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5227, 2.6520, 2.1100, 2.3070, 2.6959, 2.2493, 3.5927, 2.1133], + device='cuda:3'), covar=tensor([0.3905, 0.2368, 0.4794, 0.3505, 0.1938, 0.2790, 0.1208, 0.4156], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0354, 0.0431, 0.0355, 0.0382, 0.0379, 0.0373, 0.0426], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:46:49,427 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-27 19:47:29,412 INFO [finetune.py:976] (3/7) Epoch 23, batch 5700, loss[loss=0.1495, simple_loss=0.2098, pruned_loss=0.04457, over 4244.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2419, pruned_loss=0.04843, over 938464.92 frames. ], batch size: 18, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:48:21,856 INFO [finetune.py:976] (3/7) Epoch 24, batch 0, loss[loss=0.1446, simple_loss=0.2168, pruned_loss=0.03627, over 4744.00 frames. ], tot_loss[loss=0.1446, simple_loss=0.2168, pruned_loss=0.03627, over 4744.00 frames. ], batch size: 23, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:48:21,856 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 19:48:25,052 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3562, 1.5407, 1.9285, 2.0542, 1.9830, 2.0690, 1.9780, 2.0219], + device='cuda:3'), covar=tensor([0.3621, 0.5449, 0.4488, 0.4584, 0.5337, 0.7019, 0.5214, 0.4130], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0375, 0.0328, 0.0339, 0.0349, 0.0394, 0.0357, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:48:32,040 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3527, 1.5445, 1.9283, 2.0275, 1.9780, 2.1276, 1.9624, 2.0142], + device='cuda:3'), covar=tensor([0.3442, 0.4892, 0.4127, 0.4234, 0.5030, 0.6156, 0.4552, 0.4176], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0375, 0.0328, 0.0339, 0.0349, 0.0394, 0.0357, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:48:37,427 INFO [finetune.py:1010] (3/7) Epoch 24, validation: loss=0.1552, simple_loss=0.2243, pruned_loss=0.04308, over 2265189.00 frames. +2023-04-27 19:48:37,427 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 19:48:45,627 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2103, 1.4941, 1.3465, 1.7134, 1.6714, 1.8574, 1.3647, 3.2570], + device='cuda:3'), covar=tensor([0.0614, 0.0793, 0.0802, 0.1167, 0.0606, 0.0515, 0.0737, 0.0149], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 19:49:12,804 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:49:13,312 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.101e+01 1.500e+02 1.790e+02 2.328e+02 5.462e+02, threshold=3.579e+02, percent-clipped=4.0 +2023-04-27 19:49:21,988 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131772.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:49:37,728 INFO [finetune.py:976] (3/7) Epoch 24, batch 50, loss[loss=0.1807, simple_loss=0.2597, pruned_loss=0.05083, over 4878.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2437, pruned_loss=0.04742, over 217604.23 frames. ], batch size: 32, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:50:39,504 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131830.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:50:48,616 INFO [finetune.py:976] (3/7) Epoch 24, batch 100, loss[loss=0.1818, simple_loss=0.2424, pruned_loss=0.06055, over 4915.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2394, pruned_loss=0.04833, over 381517.53 frames. ], batch size: 46, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:50:56,376 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:50:58,855 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1812, 1.9600, 2.2343, 2.5071, 2.5227, 2.1603, 1.8463, 2.2890], + device='cuda:3'), covar=tensor([0.0717, 0.1125, 0.0657, 0.0576, 0.0531, 0.0789, 0.0708, 0.0552], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0201, 0.0182, 0.0171, 0.0174, 0.0177, 0.0147, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:51:01,829 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131857.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:51:07,126 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.089e+01 1.524e+02 1.795e+02 2.090e+02 4.072e+02, threshold=3.590e+02, percent-clipped=1.0 +2023-04-27 19:51:14,669 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8125, 1.8001, 2.2390, 2.3261, 1.7172, 1.4656, 1.8890, 0.9674], + device='cuda:3'), covar=tensor([0.0647, 0.0613, 0.0372, 0.0742, 0.0753, 0.0998, 0.0633, 0.0712], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0067, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 19:51:23,453 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131878.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:51:35,581 INFO [finetune.py:976] (3/7) Epoch 24, batch 150, loss[loss=0.1239, simple_loss=0.1954, pruned_loss=0.02624, over 4782.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2328, pruned_loss=0.04522, over 508967.79 frames. ], batch size: 28, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:52:18,771 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131918.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:52:44,204 INFO [finetune.py:976] (3/7) Epoch 24, batch 200, loss[loss=0.1734, simple_loss=0.2372, pruned_loss=0.05483, over 4825.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2326, pruned_loss=0.04607, over 608882.46 frames. ], batch size: 33, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:53:12,911 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.534e+02 1.797e+02 2.198e+02 4.905e+02, threshold=3.594e+02, percent-clipped=2.0 +2023-04-27 19:53:27,656 INFO [finetune.py:976] (3/7) Epoch 24, batch 250, loss[loss=0.2175, simple_loss=0.2717, pruned_loss=0.08168, over 4832.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2386, pruned_loss=0.04888, over 683365.71 frames. ], batch size: 30, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:53:57,465 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132031.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:54:01,446 INFO [finetune.py:976] (3/7) Epoch 24, batch 300, loss[loss=0.1834, simple_loss=0.2631, pruned_loss=0.05181, over 4773.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2431, pruned_loss=0.04989, over 743265.33 frames. ], batch size: 28, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:54:17,606 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-04-27 19:54:24,640 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132054.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:54:36,455 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132065.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:54:36,978 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.704e+02 1.925e+02 2.353e+02 6.924e+02, threshold=3.849e+02, percent-clipped=2.0 +2023-04-27 19:54:46,358 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132072.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:54:51,131 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7212, 1.6748, 0.7817, 1.4646, 1.7789, 1.5608, 1.5002, 1.5521], + device='cuda:3'), covar=tensor([0.0480, 0.0375, 0.0332, 0.0541, 0.0268, 0.0524, 0.0498, 0.0570], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:3') +2023-04-27 19:55:00,446 INFO [finetune.py:976] (3/7) Epoch 24, batch 350, loss[loss=0.1252, simple_loss=0.1964, pruned_loss=0.02698, over 4711.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.245, pruned_loss=0.05004, over 789187.88 frames. ], batch size: 23, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:55:10,377 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132092.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:55:40,637 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132113.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:55:41,938 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132115.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:55:44,962 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132120.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:55:45,003 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5190, 2.9758, 1.1494, 1.7398, 2.4457, 1.4941, 4.1490, 1.8725], + device='cuda:3'), covar=tensor([0.0574, 0.0917, 0.0847, 0.1244, 0.0478, 0.0997, 0.0221, 0.0662], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0050, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 19:56:06,851 INFO [finetune.py:976] (3/7) Epoch 24, batch 400, loss[loss=0.1932, simple_loss=0.2545, pruned_loss=0.06591, over 4879.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2465, pruned_loss=0.05054, over 823586.16 frames. ], batch size: 32, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:56:25,728 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132148.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:56:49,860 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.575e+02 1.808e+02 2.256e+02 3.614e+02, threshold=3.616e+02, percent-clipped=0.0 +2023-04-27 19:57:08,719 INFO [finetune.py:976] (3/7) Epoch 24, batch 450, loss[loss=0.1368, simple_loss=0.2083, pruned_loss=0.0326, over 4854.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.245, pruned_loss=0.04982, over 853633.68 frames. ], batch size: 44, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:57:12,343 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.6514, 1.4171, 1.3167, 1.0971, 1.3124, 1.2201, 1.5352, 1.2447], + device='cuda:3'), covar=tensor([0.2737, 0.1517, 0.3793, 0.2399, 0.1457, 0.1814, 0.1663, 0.3819], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0353, 0.0428, 0.0354, 0.0380, 0.0378, 0.0370, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:57:15,197 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132196.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:57:27,640 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132213.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:57:31,906 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4801, 1.3054, 1.6099, 1.6642, 1.3491, 1.3052, 1.3070, 0.8894], + device='cuda:3'), covar=tensor([0.0446, 0.0612, 0.0378, 0.0538, 0.0632, 0.1019, 0.0450, 0.0515], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0067, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 19:57:42,208 INFO [finetune.py:976] (3/7) Epoch 24, batch 500, loss[loss=0.169, simple_loss=0.2345, pruned_loss=0.05172, over 4221.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2429, pruned_loss=0.04929, over 875376.41 frames. ], batch size: 18, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:58:03,238 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.544e+02 1.768e+02 2.170e+02 4.737e+02, threshold=3.537e+02, percent-clipped=2.0 +2023-04-27 19:58:06,505 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4388, 1.8293, 1.8792, 1.9860, 1.8997, 2.0188, 1.9002, 1.9201], + device='cuda:3'), covar=tensor([0.3488, 0.4318, 0.3610, 0.3632, 0.4549, 0.5933, 0.4650, 0.4032], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0327, 0.0339, 0.0349, 0.0394, 0.0357, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 19:58:16,037 INFO [finetune.py:976] (3/7) Epoch 24, batch 550, loss[loss=0.1407, simple_loss=0.2076, pruned_loss=0.03684, over 4741.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2396, pruned_loss=0.0482, over 891030.39 frames. ], batch size: 23, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:58:49,839 INFO [finetune.py:976] (3/7) Epoch 24, batch 600, loss[loss=0.2023, simple_loss=0.2578, pruned_loss=0.07342, over 4761.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2408, pruned_loss=0.04861, over 903321.70 frames. ], batch size: 27, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:59:10,241 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.689e+02 1.942e+02 2.477e+02 4.504e+02, threshold=3.885e+02, percent-clipped=3.0 +2023-04-27 19:59:22,996 INFO [finetune.py:976] (3/7) Epoch 24, batch 650, loss[loss=0.1985, simple_loss=0.2818, pruned_loss=0.05762, over 4897.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2434, pruned_loss=0.04897, over 916068.78 frames. ], batch size: 43, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:59:23,064 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132387.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:59:39,629 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132410.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:59:46,889 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8464, 1.3626, 1.9169, 2.2834, 1.9351, 1.8016, 1.8762, 1.7819], + device='cuda:3'), covar=tensor([0.4621, 0.7193, 0.6513, 0.5574, 0.5769, 0.7868, 0.8566, 0.9722], + device='cuda:3'), in_proj_covar=tensor([0.0434, 0.0416, 0.0511, 0.0505, 0.0462, 0.0494, 0.0499, 0.0510], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 19:59:58,074 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6592, 1.4706, 1.7614, 1.8855, 1.4765, 1.4076, 1.4879, 0.8838], + device='cuda:3'), covar=tensor([0.0469, 0.0631, 0.0398, 0.0504, 0.0682, 0.1104, 0.0519, 0.0615], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:00:08,097 INFO [finetune.py:976] (3/7) Epoch 24, batch 700, loss[loss=0.1764, simple_loss=0.2478, pruned_loss=0.05246, over 4849.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2448, pruned_loss=0.04959, over 925156.63 frames. ], batch size: 44, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 20:00:50,349 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.633e+02 1.958e+02 2.360e+02 4.192e+02, threshold=3.915e+02, percent-clipped=4.0 +2023-04-27 20:01:15,825 INFO [finetune.py:976] (3/7) Epoch 24, batch 750, loss[loss=0.1523, simple_loss=0.2318, pruned_loss=0.03638, over 4747.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.246, pruned_loss=0.04981, over 931977.18 frames. ], batch size: 26, lr: 3.06e-03, grad_scale: 32.0 +2023-04-27 20:01:48,573 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132513.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:02:21,498 INFO [finetune.py:976] (3/7) Epoch 24, batch 800, loss[loss=0.2059, simple_loss=0.2648, pruned_loss=0.07355, over 4874.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2439, pruned_loss=0.04887, over 936357.24 frames. ], batch size: 35, lr: 3.06e-03, grad_scale: 32.0 +2023-04-27 20:02:40,449 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1936, 1.3837, 1.2621, 1.6251, 1.5282, 1.5869, 1.3546, 2.5181], + device='cuda:3'), covar=tensor([0.0586, 0.0849, 0.0831, 0.1211, 0.0656, 0.0499, 0.0723, 0.0214], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 20:02:53,594 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132561.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:03:01,810 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.951e+01 1.478e+02 1.835e+02 2.240e+02 6.510e+02, threshold=3.671e+02, percent-clipped=1.0 +2023-04-27 20:03:23,026 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5912, 3.6290, 0.9602, 1.8551, 2.1295, 2.6316, 2.0513, 0.9819], + device='cuda:3'), covar=tensor([0.1335, 0.0874, 0.2101, 0.1294, 0.1063, 0.0947, 0.1527, 0.2004], + device='cuda:3'), in_proj_covar=tensor([0.0115, 0.0236, 0.0135, 0.0119, 0.0131, 0.0150, 0.0115, 0.0117], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:03:27,812 INFO [finetune.py:976] (3/7) Epoch 24, batch 850, loss[loss=0.1642, simple_loss=0.2414, pruned_loss=0.04353, over 4890.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2423, pruned_loss=0.04823, over 939652.69 frames. ], batch size: 32, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:04:08,614 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9103, 1.9748, 1.8094, 1.6772, 2.0952, 1.5928, 2.5850, 1.6089], + device='cuda:3'), covar=tensor([0.3859, 0.1812, 0.4833, 0.3128, 0.1675, 0.2809, 0.1395, 0.4710], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0351, 0.0424, 0.0351, 0.0377, 0.0375, 0.0367, 0.0420], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 20:04:34,504 INFO [finetune.py:976] (3/7) Epoch 24, batch 900, loss[loss=0.1748, simple_loss=0.253, pruned_loss=0.04836, over 4750.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2391, pruned_loss=0.04714, over 941693.73 frames. ], batch size: 23, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:05:01,491 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7927, 1.8753, 1.7976, 1.5495, 1.9180, 1.6107, 2.3509, 1.5629], + device='cuda:3'), covar=tensor([0.3364, 0.1633, 0.4208, 0.2530, 0.1572, 0.2222, 0.1416, 0.4412], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0350, 0.0424, 0.0351, 0.0377, 0.0374, 0.0367, 0.0419], + device='cuda:3'), out_proj_covar=tensor([9.9948e-05, 1.0460e-04, 1.2852e-04, 1.0544e-04, 1.1195e-04, 1.1158e-04, + 1.0777e-04, 1.2605e-04], device='cuda:3') +2023-04-27 20:05:02,097 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132660.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:05:12,904 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.989e+01 1.487e+02 1.792e+02 2.042e+02 3.425e+02, threshold=3.585e+02, percent-clipped=0.0 +2023-04-27 20:05:44,311 INFO [finetune.py:976] (3/7) Epoch 24, batch 950, loss[loss=0.1998, simple_loss=0.272, pruned_loss=0.06382, over 4825.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2384, pruned_loss=0.04745, over 942104.22 frames. ], batch size: 39, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:05:44,410 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132687.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:06:08,277 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132710.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 20:06:22,001 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 20:06:27,931 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132721.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:06:29,153 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2017, 2.8278, 2.1827, 2.6536, 1.7693, 2.4040, 2.4653, 1.9382], + device='cuda:3'), covar=tensor([0.2266, 0.1214, 0.1030, 0.1283, 0.3747, 0.1161, 0.1959, 0.2662], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0303, 0.0216, 0.0275, 0.0316, 0.0256, 0.0248, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1478e-04, 1.1992e-04, 8.5110e-05, 1.0858e-04, 1.2732e-04, 1.0104e-04, + 1.0026e-04, 1.0445e-04], device='cuda:3') +2023-04-27 20:06:43,787 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132735.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:06:49,488 INFO [finetune.py:976] (3/7) Epoch 24, batch 1000, loss[loss=0.1798, simple_loss=0.2464, pruned_loss=0.05662, over 4930.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2402, pruned_loss=0.04822, over 946100.80 frames. ], batch size: 37, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:06:55,147 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132746.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:07:02,346 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132758.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:07:07,765 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.549e+02 1.832e+02 2.153e+02 4.117e+02, threshold=3.664e+02, percent-clipped=1.0 +2023-04-27 20:07:22,249 INFO [finetune.py:976] (3/7) Epoch 24, batch 1050, loss[loss=0.1859, simple_loss=0.2533, pruned_loss=0.05924, over 4741.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2419, pruned_loss=0.04836, over 948674.89 frames. ], batch size: 59, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:07:35,145 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132807.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:07:38,789 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4344, 1.7347, 1.7874, 1.9097, 1.7547, 1.8590, 1.8809, 1.8220], + device='cuda:3'), covar=tensor([0.3463, 0.4884, 0.4223, 0.3912, 0.5189, 0.6768, 0.4787, 0.4500], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0378, 0.0329, 0.0341, 0.0351, 0.0396, 0.0359, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:07:56,050 INFO [finetune.py:976] (3/7) Epoch 24, batch 1100, loss[loss=0.159, simple_loss=0.2455, pruned_loss=0.03623, over 4885.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2432, pruned_loss=0.04833, over 951463.43 frames. ], batch size: 43, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:08:14,644 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.617e+01 1.589e+02 1.846e+02 2.225e+02 4.028e+02, threshold=3.692e+02, percent-clipped=2.0 +2023-04-27 20:08:28,715 INFO [finetune.py:976] (3/7) Epoch 24, batch 1150, loss[loss=0.1727, simple_loss=0.2527, pruned_loss=0.04638, over 4857.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2437, pruned_loss=0.04797, over 953227.06 frames. ], batch size: 44, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:09:11,143 INFO [finetune.py:976] (3/7) Epoch 24, batch 1200, loss[loss=0.1412, simple_loss=0.202, pruned_loss=0.04026, over 4303.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2422, pruned_loss=0.04713, over 953380.86 frames. ], batch size: 19, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:09:48,366 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.568e+02 1.953e+02 2.291e+02 3.731e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-04-27 20:10:09,526 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.0957, 4.1350, 2.9051, 4.7589, 4.1984, 4.1493, 1.5922, 4.0114], + device='cuda:3'), covar=tensor([0.1748, 0.1166, 0.3020, 0.1332, 0.3949, 0.1598, 0.6266, 0.2423], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0219, 0.0252, 0.0304, 0.0293, 0.0246, 0.0273, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:10:17,856 INFO [finetune.py:976] (3/7) Epoch 24, batch 1250, loss[loss=0.1491, simple_loss=0.216, pruned_loss=0.0411, over 4819.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2397, pruned_loss=0.0467, over 952876.27 frames. ], batch size: 30, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:10:30,024 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.48 vs. limit=5.0 +2023-04-27 20:10:53,171 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133016.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:11:17,182 INFO [finetune.py:976] (3/7) Epoch 24, batch 1300, loss[loss=0.1437, simple_loss=0.2225, pruned_loss=0.0325, over 4774.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2373, pruned_loss=0.0461, over 951874.50 frames. ], batch size: 27, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:11:37,887 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.644e+02 1.911e+02 2.318e+02 6.826e+02, threshold=3.821e+02, percent-clipped=2.0 +2023-04-27 20:11:50,600 INFO [finetune.py:976] (3/7) Epoch 24, batch 1350, loss[loss=0.1734, simple_loss=0.2459, pruned_loss=0.05049, over 4814.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2384, pruned_loss=0.04717, over 953346.97 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:12:07,604 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133102.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:12:10,560 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9061, 2.1202, 2.0251, 2.2359, 2.0422, 2.1416, 2.1186, 2.0941], + device='cuda:3'), covar=tensor([0.3832, 0.6318, 0.5145, 0.4470, 0.5840, 0.7259, 0.6303, 0.5284], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0378, 0.0330, 0.0341, 0.0350, 0.0396, 0.0360, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:12:52,452 INFO [finetune.py:976] (3/7) Epoch 24, batch 1400, loss[loss=0.1689, simple_loss=0.2531, pruned_loss=0.04238, over 4828.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2396, pruned_loss=0.04711, over 951875.21 frames. ], batch size: 33, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:13:29,242 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.717e+02 2.018e+02 2.257e+02 3.714e+02, threshold=4.037e+02, percent-clipped=0.0 +2023-04-27 20:13:41,457 INFO [finetune.py:976] (3/7) Epoch 24, batch 1450, loss[loss=0.1801, simple_loss=0.2639, pruned_loss=0.04812, over 4855.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2417, pruned_loss=0.04793, over 952066.68 frames. ], batch size: 44, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:13:42,799 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5374, 1.7014, 1.6182, 2.0620, 1.9716, 1.8519, 1.5136, 4.4162], + device='cuda:3'), covar=tensor([0.0493, 0.0798, 0.0764, 0.1146, 0.0615, 0.0576, 0.0702, 0.0099], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 20:14:31,510 INFO [finetune.py:976] (3/7) Epoch 24, batch 1500, loss[loss=0.1927, simple_loss=0.2328, pruned_loss=0.07631, over 3978.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2431, pruned_loss=0.0487, over 951970.47 frames. ], batch size: 17, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:15:14,562 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.609e+02 1.860e+02 2.213e+02 4.460e+02, threshold=3.721e+02, percent-clipped=1.0 +2023-04-27 20:15:44,449 INFO [finetune.py:976] (3/7) Epoch 24, batch 1550, loss[loss=0.1762, simple_loss=0.2485, pruned_loss=0.05193, over 4781.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2438, pruned_loss=0.04874, over 953077.40 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:16:10,959 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133316.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:16:34,613 INFO [finetune.py:976] (3/7) Epoch 24, batch 1600, loss[loss=0.2008, simple_loss=0.2744, pruned_loss=0.06355, over 4825.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2432, pruned_loss=0.04901, over 952421.60 frames. ], batch size: 33, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:16:43,860 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9543, 1.6990, 1.8841, 2.2842, 2.3123, 1.8211, 1.5823, 2.0062], + device='cuda:3'), covar=tensor([0.0764, 0.1187, 0.0780, 0.0567, 0.0507, 0.0820, 0.0746, 0.0557], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0205, 0.0187, 0.0174, 0.0179, 0.0181, 0.0151, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 20:16:54,207 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3928, 3.3498, 2.4617, 3.8983, 3.3995, 3.4529, 1.3651, 3.3359], + device='cuda:3'), covar=tensor([0.1838, 0.1433, 0.3258, 0.2104, 0.2788, 0.1863, 0.5812, 0.2341], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0217, 0.0250, 0.0303, 0.0293, 0.0245, 0.0272, 0.0270], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:17:17,456 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133364.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:17:19,226 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.451e+02 1.794e+02 2.296e+02 6.131e+02, threshold=3.588e+02, percent-clipped=2.0 +2023-04-27 20:17:29,312 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133374.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 20:17:30,582 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133376.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:17:42,836 INFO [finetune.py:976] (3/7) Epoch 24, batch 1650, loss[loss=0.1348, simple_loss=0.2116, pruned_loss=0.02897, over 4865.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2402, pruned_loss=0.0477, over 952396.75 frames. ], batch size: 31, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:18:01,035 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8769, 2.1268, 2.1569, 2.3158, 2.1644, 2.1787, 2.1864, 2.1410], + device='cuda:3'), covar=tensor([0.3815, 0.6468, 0.5062, 0.4564, 0.5530, 0.7030, 0.6493, 0.5871], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0376, 0.0327, 0.0339, 0.0347, 0.0393, 0.0357, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:18:02,184 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133402.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:18:23,685 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2151, 2.8233, 2.3297, 2.7603, 1.9603, 2.5070, 2.7375, 2.0203], + device='cuda:3'), covar=tensor([0.2107, 0.1548, 0.1068, 0.1386, 0.3146, 0.1179, 0.1623, 0.2458], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0303, 0.0217, 0.0277, 0.0316, 0.0257, 0.0249, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1483e-04, 1.1978e-04, 8.5416e-05, 1.0917e-04, 1.2732e-04, 1.0132e-04, + 1.0069e-04, 1.0485e-04], device='cuda:3') +2023-04-27 20:18:48,098 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133435.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 20:18:49,169 INFO [finetune.py:976] (3/7) Epoch 24, batch 1700, loss[loss=0.2289, simple_loss=0.2882, pruned_loss=0.08477, over 4817.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2371, pruned_loss=0.04709, over 952508.67 frames. ], batch size: 41, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:18:49,289 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133437.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:19:07,623 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 20:19:08,014 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133450.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:19:19,710 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8633, 1.5216, 1.4655, 1.6722, 2.0590, 1.6425, 1.4288, 1.3881], + device='cuda:3'), covar=tensor([0.1593, 0.1198, 0.1711, 0.1155, 0.0780, 0.1538, 0.1752, 0.2021], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0307, 0.0350, 0.0284, 0.0327, 0.0306, 0.0299, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.3920e-05, 6.3234e-05, 7.3539e-05, 5.6960e-05, 6.7306e-05, 6.4024e-05, + 6.2226e-05, 7.8938e-05], device='cuda:3') +2023-04-27 20:19:31,793 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.276e+01 1.467e+02 1.697e+02 2.216e+02 4.193e+02, threshold=3.394e+02, percent-clipped=4.0 +2023-04-27 20:19:40,731 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-04-27 20:19:41,274 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133474.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:19:55,043 INFO [finetune.py:976] (3/7) Epoch 24, batch 1750, loss[loss=0.2047, simple_loss=0.2874, pruned_loss=0.06099, over 4734.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.241, pruned_loss=0.04884, over 952483.48 frames. ], batch size: 59, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:21:01,614 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133535.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:21:08,175 INFO [finetune.py:976] (3/7) Epoch 24, batch 1800, loss[loss=0.2338, simple_loss=0.2991, pruned_loss=0.08426, over 4757.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.244, pruned_loss=0.04961, over 953990.69 frames. ], batch size: 59, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:21:29,314 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1791, 2.0679, 1.7100, 1.7730, 2.1389, 1.8785, 2.5687, 1.4441], + device='cuda:3'), covar=tensor([0.3387, 0.1806, 0.4427, 0.2867, 0.1615, 0.2262, 0.1315, 0.4605], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0351, 0.0424, 0.0348, 0.0375, 0.0373, 0.0367, 0.0418], + device='cuda:3'), out_proj_covar=tensor([9.9669e-05, 1.0476e-04, 1.2838e-04, 1.0478e-04, 1.1151e-04, 1.1123e-04, + 1.0760e-04, 1.2582e-04], device='cuda:3') +2023-04-27 20:21:44,188 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.423e+01 1.601e+02 1.910e+02 2.304e+02 3.946e+02, threshold=3.820e+02, percent-clipped=1.0 +2023-04-27 20:22:08,895 INFO [finetune.py:976] (3/7) Epoch 24, batch 1850, loss[loss=0.1816, simple_loss=0.2541, pruned_loss=0.05455, over 4830.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2446, pruned_loss=0.04928, over 954000.69 frames. ], batch size: 47, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:22:42,123 INFO [finetune.py:976] (3/7) Epoch 24, batch 1900, loss[loss=0.1502, simple_loss=0.2303, pruned_loss=0.03508, over 4788.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2462, pruned_loss=0.04999, over 952125.80 frames. ], batch size: 29, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:22:51,310 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133652.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:23:12,128 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.575e+02 1.806e+02 2.357e+02 5.107e+02, threshold=3.612e+02, percent-clipped=3.0 +2023-04-27 20:23:18,606 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2329, 1.3887, 1.6744, 1.8022, 1.7291, 1.8207, 1.7338, 1.7621], + device='cuda:3'), covar=tensor([0.3837, 0.5059, 0.3884, 0.3893, 0.5083, 0.6416, 0.4741, 0.4593], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0326, 0.0337, 0.0346, 0.0392, 0.0356, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:23:25,441 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 20:23:26,868 INFO [finetune.py:976] (3/7) Epoch 24, batch 1950, loss[loss=0.1492, simple_loss=0.2205, pruned_loss=0.03893, over 4928.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2443, pruned_loss=0.04878, over 951141.04 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:23:31,882 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8446, 2.2275, 1.8131, 1.6520, 1.3212, 1.3658, 1.9343, 1.3120], + device='cuda:3'), covar=tensor([0.1591, 0.1314, 0.1366, 0.1648, 0.2228, 0.1873, 0.0878, 0.1946], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0212, 0.0170, 0.0206, 0.0201, 0.0186, 0.0157, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 20:23:43,298 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133713.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:23:51,047 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6637, 3.6614, 0.9564, 1.9334, 2.0787, 2.5464, 2.1246, 0.9858], + device='cuda:3'), covar=tensor([0.1235, 0.0764, 0.1913, 0.1160, 0.0962, 0.0958, 0.1354, 0.1883], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0239, 0.0137, 0.0121, 0.0133, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:23:56,174 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133730.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 20:23:57,386 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133732.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:24:00,818 INFO [finetune.py:976] (3/7) Epoch 24, batch 2000, loss[loss=0.1637, simple_loss=0.2316, pruned_loss=0.04786, over 4796.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.241, pruned_loss=0.04772, over 953929.98 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:24:25,086 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.642e+01 1.427e+02 1.749e+02 2.000e+02 3.515e+02, threshold=3.497e+02, percent-clipped=0.0 +2023-04-27 20:24:55,311 INFO [finetune.py:976] (3/7) Epoch 24, batch 2050, loss[loss=0.1719, simple_loss=0.2437, pruned_loss=0.05006, over 4817.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2379, pruned_loss=0.04665, over 954587.37 frames. ], batch size: 39, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:25:48,585 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133830.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:25:57,783 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5608, 3.6380, 0.8820, 1.8554, 2.0781, 2.5260, 2.0337, 0.9980], + device='cuda:3'), covar=tensor([0.1425, 0.1289, 0.2033, 0.1319, 0.1120, 0.1110, 0.1612, 0.1998], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0121, 0.0133, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:25:58,915 INFO [finetune.py:976] (3/7) Epoch 24, batch 2100, loss[loss=0.1704, simple_loss=0.2335, pruned_loss=0.05367, over 4719.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2376, pruned_loss=0.04717, over 955230.52 frames. ], batch size: 23, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:26:07,184 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6725, 3.6100, 2.6805, 4.2117, 3.6863, 3.6267, 1.5739, 3.5359], + device='cuda:3'), covar=tensor([0.1671, 0.1339, 0.3159, 0.1834, 0.3574, 0.1757, 0.5956, 0.2357], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0220, 0.0253, 0.0306, 0.0296, 0.0247, 0.0275, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:26:22,914 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.658e+02 2.018e+02 2.431e+02 5.100e+02, threshold=4.036e+02, percent-clipped=6.0 +2023-04-27 20:26:37,510 INFO [finetune.py:976] (3/7) Epoch 24, batch 2150, loss[loss=0.1838, simple_loss=0.2666, pruned_loss=0.05056, over 4814.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2408, pruned_loss=0.04826, over 954513.03 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:26:50,714 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7258, 3.7136, 2.7587, 4.3165, 3.8097, 3.6601, 1.6180, 3.6779], + device='cuda:3'), covar=tensor([0.1677, 0.1413, 0.3189, 0.1786, 0.3212, 0.1813, 0.5710, 0.2182], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0219, 0.0253, 0.0306, 0.0296, 0.0247, 0.0275, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:27:09,768 INFO [finetune.py:976] (3/7) Epoch 24, batch 2200, loss[loss=0.1668, simple_loss=0.2385, pruned_loss=0.04755, over 4814.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2439, pruned_loss=0.04921, over 955042.22 frames. ], batch size: 25, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:27:48,470 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.369e+01 1.555e+02 1.795e+02 2.184e+02 7.045e+02, threshold=3.590e+02, percent-clipped=2.0 +2023-04-27 20:28:06,399 INFO [finetune.py:976] (3/7) Epoch 24, batch 2250, loss[loss=0.1266, simple_loss=0.1933, pruned_loss=0.02995, over 3872.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2463, pruned_loss=0.05039, over 955742.66 frames. ], batch size: 16, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:28:06,474 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3080, 3.3382, 2.5183, 3.8370, 3.4092, 3.3182, 1.5146, 3.2541], + device='cuda:3'), covar=tensor([0.2054, 0.1409, 0.3441, 0.2266, 0.2813, 0.1962, 0.5849, 0.2895], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0218, 0.0252, 0.0306, 0.0295, 0.0247, 0.0274, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:28:06,518 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5524, 1.4840, 0.6622, 1.3103, 1.4293, 1.4358, 1.3425, 1.4409], + device='cuda:3'), covar=tensor([0.0496, 0.0374, 0.0362, 0.0550, 0.0287, 0.0503, 0.0474, 0.0551], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 20:28:14,964 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133997.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:28:23,002 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134008.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:28:36,972 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134030.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 20:28:38,195 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:28:41,585 INFO [finetune.py:976] (3/7) Epoch 24, batch 2300, loss[loss=0.1859, simple_loss=0.2477, pruned_loss=0.06208, over 4922.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2466, pruned_loss=0.05015, over 955568.80 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:28:55,877 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134058.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:29:01,261 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.437e+01 1.459e+02 1.887e+02 2.167e+02 3.394e+02, threshold=3.773e+02, percent-clipped=1.0 +2023-04-27 20:29:08,091 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134078.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 20:29:09,323 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:29:14,456 INFO [finetune.py:976] (3/7) Epoch 24, batch 2350, loss[loss=0.1916, simple_loss=0.2615, pruned_loss=0.06085, over 4818.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2435, pruned_loss=0.04891, over 956410.43 frames. ], batch size: 30, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:29:42,024 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134130.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:29:46,678 INFO [finetune.py:976] (3/7) Epoch 24, batch 2400, loss[loss=0.1584, simple_loss=0.2272, pruned_loss=0.04479, over 4836.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2408, pruned_loss=0.04849, over 954139.81 frames. ], batch size: 44, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:30:06,807 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.831e+01 1.496e+02 1.764e+02 2.072e+02 3.760e+02, threshold=3.528e+02, percent-clipped=0.0 +2023-04-27 20:30:13,577 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134178.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:30:19,543 INFO [finetune.py:976] (3/7) Epoch 24, batch 2450, loss[loss=0.1681, simple_loss=0.244, pruned_loss=0.04609, over 4818.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.239, pruned_loss=0.04809, over 955726.47 frames. ], batch size: 39, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:30:21,502 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-04-27 20:31:01,100 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9166, 1.5783, 4.0342, 3.7974, 3.5125, 3.7184, 3.6347, 3.5631], + device='cuda:3'), covar=tensor([0.5916, 0.4949, 0.0961, 0.1351, 0.1077, 0.1303, 0.2956, 0.1411], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0305, 0.0405, 0.0405, 0.0348, 0.0408, 0.0316, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 20:31:07,210 INFO [finetune.py:976] (3/7) Epoch 24, batch 2500, loss[loss=0.1828, simple_loss=0.2739, pruned_loss=0.04588, over 4777.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2423, pruned_loss=0.04932, over 954964.99 frames. ], batch size: 59, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:31:39,320 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.581e+02 1.827e+02 2.358e+02 4.176e+02, threshold=3.655e+02, percent-clipped=4.0 +2023-04-27 20:32:02,532 INFO [finetune.py:976] (3/7) Epoch 24, batch 2550, loss[loss=0.1779, simple_loss=0.2396, pruned_loss=0.05809, over 4049.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2462, pruned_loss=0.05041, over 955138.10 frames. ], batch size: 66, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:32:26,932 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134304.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:32:34,601 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134308.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:32:35,207 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134309.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:33:07,652 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8118, 4.1620, 0.9509, 2.3038, 2.5158, 3.1144, 2.4556, 0.9534], + device='cuda:3'), covar=tensor([0.1410, 0.1042, 0.1963, 0.1253, 0.1021, 0.0937, 0.1449, 0.2182], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0121, 0.0133, 0.0152, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:33:08,338 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2394, 1.4019, 1.7590, 1.8428, 1.7441, 1.7717, 1.7962, 1.7719], + device='cuda:3'), covar=tensor([0.3931, 0.5159, 0.4173, 0.3880, 0.5415, 0.7151, 0.4814, 0.4146], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0337, 0.0347, 0.0392, 0.0357, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:33:08,803 INFO [finetune.py:976] (3/7) Epoch 24, batch 2600, loss[loss=0.2048, simple_loss=0.2718, pruned_loss=0.06894, over 4927.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.246, pruned_loss=0.05018, over 955465.57 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:33:23,516 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7960, 1.9977, 2.0790, 2.1963, 1.9646, 2.0492, 2.1240, 2.0206], + device='cuda:3'), covar=tensor([0.3741, 0.6183, 0.4559, 0.4174, 0.5877, 0.6955, 0.5807, 0.5417], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0337, 0.0346, 0.0392, 0.0357, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:33:27,035 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134353.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:33:30,513 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:33:42,153 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134365.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:33:43,228 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.588e+02 1.844e+02 2.194e+02 4.309e+02, threshold=3.689e+02, percent-clipped=2.0 +2023-04-27 20:33:45,774 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134370.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:34:07,320 INFO [finetune.py:976] (3/7) Epoch 24, batch 2650, loss[loss=0.1751, simple_loss=0.2566, pruned_loss=0.0468, over 4847.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2467, pruned_loss=0.04998, over 954407.99 frames. ], batch size: 44, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:34:16,580 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-27 20:35:02,816 INFO [finetune.py:976] (3/7) Epoch 24, batch 2700, loss[loss=0.1795, simple_loss=0.2592, pruned_loss=0.0499, over 4901.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2452, pruned_loss=0.04901, over 951980.07 frames. ], batch size: 43, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:35:23,806 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.513e+01 1.459e+02 1.776e+02 2.011e+02 3.451e+02, threshold=3.553e+02, percent-clipped=0.0 +2023-04-27 20:35:36,506 INFO [finetune.py:976] (3/7) Epoch 24, batch 2750, loss[loss=0.1851, simple_loss=0.2491, pruned_loss=0.06055, over 4914.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2421, pruned_loss=0.04813, over 952457.92 frames. ], batch size: 37, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:35:54,524 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0812, 0.8258, 0.9283, 0.8629, 1.2571, 1.0120, 0.8697, 0.9833], + device='cuda:3'), covar=tensor([0.1597, 0.1393, 0.2240, 0.1607, 0.1077, 0.1482, 0.1655, 0.2541], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0304, 0.0347, 0.0281, 0.0325, 0.0303, 0.0296, 0.0369], + device='cuda:3'), out_proj_covar=tensor([6.3156e-05, 6.2699e-05, 7.2919e-05, 5.6413e-05, 6.6977e-05, 6.3568e-05, + 6.1595e-05, 7.8228e-05], device='cuda:3') +2023-04-27 20:36:16,398 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 20:36:43,196 INFO [finetune.py:976] (3/7) Epoch 24, batch 2800, loss[loss=0.1573, simple_loss=0.2254, pruned_loss=0.04459, over 4835.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2377, pruned_loss=0.04656, over 953414.52 frames. ], batch size: 30, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:37:21,563 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1769, 2.2488, 1.9210, 1.8625, 2.2884, 1.9367, 2.8436, 1.7080], + device='cuda:3'), covar=tensor([0.3461, 0.1833, 0.4498, 0.2762, 0.1611, 0.2323, 0.1292, 0.4059], + device='cuda:3'), in_proj_covar=tensor([0.0335, 0.0347, 0.0422, 0.0346, 0.0375, 0.0370, 0.0365, 0.0417], + device='cuda:3'), out_proj_covar=tensor([9.9059e-05, 1.0374e-04, 1.2773e-04, 1.0390e-04, 1.1138e-04, 1.1024e-04, + 1.0725e-04, 1.2544e-04], device='cuda:3') +2023-04-27 20:37:23,286 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.232e+01 1.527e+02 1.797e+02 2.170e+02 3.863e+02, threshold=3.594e+02, percent-clipped=3.0 +2023-04-27 20:37:48,299 INFO [finetune.py:976] (3/7) Epoch 24, batch 2850, loss[loss=0.1228, simple_loss=0.1997, pruned_loss=0.02298, over 4762.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2362, pruned_loss=0.04619, over 955445.20 frames. ], batch size: 28, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:37:55,986 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7781, 2.0178, 1.9508, 2.6609, 2.7955, 2.2769, 2.2909, 1.9032], + device='cuda:3'), covar=tensor([0.1619, 0.1565, 0.1931, 0.1365, 0.0991, 0.1565, 0.1537, 0.2162], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0305, 0.0348, 0.0282, 0.0326, 0.0304, 0.0297, 0.0370], + device='cuda:3'), out_proj_covar=tensor([6.3286e-05, 6.2977e-05, 7.3205e-05, 5.6567e-05, 6.7070e-05, 6.3695e-05, + 6.1770e-05, 7.8507e-05], device='cuda:3') +2023-04-27 20:38:59,963 INFO [finetune.py:976] (3/7) Epoch 24, batch 2900, loss[loss=0.1974, simple_loss=0.2737, pruned_loss=0.06059, over 4793.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2391, pruned_loss=0.04693, over 954146.21 frames. ], batch size: 51, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:39:10,407 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 20:39:15,071 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134653.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:39:24,563 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134660.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:39:33,042 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134665.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:39:34,176 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.576e+02 1.923e+02 2.277e+02 4.210e+02, threshold=3.845e+02, percent-clipped=2.0 +2023-04-27 20:40:04,275 INFO [finetune.py:976] (3/7) Epoch 24, batch 2950, loss[loss=0.1631, simple_loss=0.2432, pruned_loss=0.04153, over 4755.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2405, pruned_loss=0.0469, over 953300.71 frames. ], batch size: 54, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:40:18,489 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134701.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:41:09,548 INFO [finetune.py:976] (3/7) Epoch 24, batch 3000, loss[loss=0.1891, simple_loss=0.2647, pruned_loss=0.05673, over 4818.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2426, pruned_loss=0.04822, over 952765.61 frames. ], batch size: 40, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:41:09,548 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 20:41:19,116 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5176, 3.0284, 0.9515, 1.8448, 1.8993, 2.3795, 1.8749, 1.0335], + device='cuda:3'), covar=tensor([0.1165, 0.0906, 0.1767, 0.1087, 0.0944, 0.0791, 0.1406, 0.1613], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0238, 0.0135, 0.0120, 0.0132, 0.0151, 0.0116, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:41:25,508 INFO [finetune.py:1010] (3/7) Epoch 24, validation: loss=0.1526, simple_loss=0.2221, pruned_loss=0.04154, over 2265189.00 frames. +2023-04-27 20:41:25,509 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 20:41:44,219 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.565e+02 1.917e+02 2.257e+02 3.857e+02, threshold=3.833e+02, percent-clipped=1.0 +2023-04-27 20:41:57,438 INFO [finetune.py:976] (3/7) Epoch 24, batch 3050, loss[loss=0.1858, simple_loss=0.2594, pruned_loss=0.05615, over 4774.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2446, pruned_loss=0.04885, over 953372.76 frames. ], batch size: 26, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:42:07,841 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9269, 1.8314, 2.2200, 2.3446, 1.6885, 1.4953, 1.8850, 1.0757], + device='cuda:3'), covar=tensor([0.0565, 0.0615, 0.0443, 0.0730, 0.0685, 0.1077, 0.0660, 0.0678], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0074, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:42:15,784 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1918, 2.8669, 2.1826, 2.4020, 1.5615, 1.6238, 2.3375, 1.5277], + device='cuda:3'), covar=tensor([0.1779, 0.1497, 0.1470, 0.1645, 0.2443, 0.1972, 0.0964, 0.2107], + device='cuda:3'), in_proj_covar=tensor([0.0195, 0.0208, 0.0167, 0.0203, 0.0198, 0.0184, 0.0155, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 20:42:30,086 INFO [finetune.py:976] (3/7) Epoch 24, batch 3100, loss[loss=0.1889, simple_loss=0.2633, pruned_loss=0.05725, over 4900.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2431, pruned_loss=0.04862, over 954008.94 frames. ], batch size: 43, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:42:47,427 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134862.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:42:50,371 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.815e+01 1.536e+02 1.798e+02 2.119e+02 3.796e+02, threshold=3.595e+02, percent-clipped=0.0 +2023-04-27 20:43:02,616 INFO [finetune.py:976] (3/7) Epoch 24, batch 3150, loss[loss=0.1557, simple_loss=0.2192, pruned_loss=0.04606, over 4258.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2405, pruned_loss=0.04775, over 952187.82 frames. ], batch size: 18, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:43:28,107 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134923.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:43:36,615 INFO [finetune.py:976] (3/7) Epoch 24, batch 3200, loss[loss=0.1549, simple_loss=0.2227, pruned_loss=0.04354, over 4812.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2368, pruned_loss=0.04653, over 952645.37 frames. ], batch size: 25, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:43:53,560 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134960.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:43:56,582 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134965.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:43:57,657 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.515e+02 1.832e+02 2.291e+02 6.391e+02, threshold=3.663e+02, percent-clipped=4.0 +2023-04-27 20:44:10,015 INFO [finetune.py:976] (3/7) Epoch 24, batch 3250, loss[loss=0.1901, simple_loss=0.2612, pruned_loss=0.05949, over 4820.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2385, pruned_loss=0.04777, over 952623.90 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:44:11,972 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 20:44:25,317 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135008.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:44:28,802 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135013.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:44:43,393 INFO [finetune.py:976] (3/7) Epoch 24, batch 3300, loss[loss=0.1497, simple_loss=0.2334, pruned_loss=0.03298, over 4799.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2424, pruned_loss=0.04935, over 951596.27 frames. ], batch size: 51, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:45:15,631 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.494e+02 1.834e+02 2.222e+02 6.565e+02, threshold=3.668e+02, percent-clipped=2.0 +2023-04-27 20:45:44,564 INFO [finetune.py:976] (3/7) Epoch 24, batch 3350, loss[loss=0.1672, simple_loss=0.2489, pruned_loss=0.04273, over 4897.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2421, pruned_loss=0.04861, over 951991.07 frames. ], batch size: 35, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:46:15,702 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7367, 1.6766, 0.7285, 1.3857, 1.5151, 1.5405, 1.4256, 1.4933], + device='cuda:3'), covar=tensor([0.0490, 0.0368, 0.0374, 0.0560, 0.0294, 0.0491, 0.0486, 0.0571], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 20:46:40,852 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8167, 1.3245, 1.9244, 2.2401, 1.9028, 1.8000, 1.8691, 1.8410], + device='cuda:3'), covar=tensor([0.4860, 0.7406, 0.6687, 0.6260, 0.6362, 0.8775, 0.8729, 0.9076], + device='cuda:3'), in_proj_covar=tensor([0.0435, 0.0418, 0.0510, 0.0506, 0.0465, 0.0498, 0.0502, 0.0513], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 20:46:49,141 INFO [finetune.py:976] (3/7) Epoch 24, batch 3400, loss[loss=0.1642, simple_loss=0.2354, pruned_loss=0.04654, over 4830.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2437, pruned_loss=0.0489, over 952990.21 frames. ], batch size: 49, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:46:51,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0294, 1.2865, 1.1782, 1.5125, 1.4086, 1.4595, 1.2375, 2.3694], + device='cuda:3'), covar=tensor([0.0666, 0.0802, 0.0781, 0.1169, 0.0634, 0.0577, 0.0732, 0.0247], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 20:46:54,294 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 20:47:25,517 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.583e+02 1.888e+02 2.345e+02 4.447e+02, threshold=3.777e+02, percent-clipped=3.0 +2023-04-27 20:47:33,954 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:47:35,833 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 20:47:42,927 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 20:47:54,154 INFO [finetune.py:976] (3/7) Epoch 24, batch 3450, loss[loss=0.1575, simple_loss=0.2325, pruned_loss=0.04128, over 4909.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2422, pruned_loss=0.04797, over 954014.30 frames. ], batch size: 32, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:48:27,447 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135218.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:48:36,539 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:48:39,438 INFO [finetune.py:976] (3/7) Epoch 24, batch 3500, loss[loss=0.1643, simple_loss=0.2386, pruned_loss=0.04501, over 4815.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2409, pruned_loss=0.04753, over 953485.83 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:48:59,255 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.566e+02 1.818e+02 2.142e+02 4.699e+02, threshold=3.637e+02, percent-clipped=2.0 +2023-04-27 20:49:13,317 INFO [finetune.py:976] (3/7) Epoch 24, batch 3550, loss[loss=0.1341, simple_loss=0.2061, pruned_loss=0.03107, over 4826.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2387, pruned_loss=0.04709, over 953930.68 frames. ], batch size: 30, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:49:47,283 INFO [finetune.py:976] (3/7) Epoch 24, batch 3600, loss[loss=0.1912, simple_loss=0.2706, pruned_loss=0.05587, over 4865.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2366, pruned_loss=0.04636, over 953876.58 frames. ], batch size: 34, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:50:05,966 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.602e+02 1.884e+02 2.235e+02 3.277e+02, threshold=3.769e+02, percent-clipped=0.0 +2023-04-27 20:50:20,185 INFO [finetune.py:976] (3/7) Epoch 24, batch 3650, loss[loss=0.1544, simple_loss=0.2278, pruned_loss=0.04049, over 4819.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2375, pruned_loss=0.04684, over 951661.15 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:50:53,969 INFO [finetune.py:976] (3/7) Epoch 24, batch 3700, loss[loss=0.1786, simple_loss=0.2494, pruned_loss=0.05389, over 4827.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.241, pruned_loss=0.04789, over 952301.70 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:51:12,479 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.617e+02 1.866e+02 2.198e+02 4.179e+02, threshold=3.733e+02, percent-clipped=1.0 +2023-04-27 20:51:14,977 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6524, 1.9348, 1.9603, 2.0794, 1.9313, 2.0441, 2.0870, 2.0435], + device='cuda:3'), covar=tensor([0.3621, 0.5359, 0.4601, 0.4374, 0.5576, 0.7131, 0.4893, 0.4943], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0376, 0.0328, 0.0340, 0.0350, 0.0396, 0.0358, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 20:51:27,090 INFO [finetune.py:976] (3/7) Epoch 24, batch 3750, loss[loss=0.1732, simple_loss=0.2392, pruned_loss=0.05359, over 4861.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2433, pruned_loss=0.04854, over 954012.91 frames. ], batch size: 31, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:52:05,325 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135518.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:52:06,584 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0003, 2.4915, 2.0389, 2.4775, 1.6662, 2.1803, 2.1434, 1.6524], + device='cuda:3'), covar=tensor([0.1938, 0.1129, 0.0842, 0.1023, 0.3137, 0.1024, 0.1893, 0.2473], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0298, 0.0214, 0.0275, 0.0312, 0.0253, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1322e-04, 1.1777e-04, 8.4139e-05, 1.0876e-04, 1.2572e-04, 9.9630e-05, + 1.0031e-04, 1.0378e-04], device='cuda:3') +2023-04-27 20:52:16,497 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135527.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:52:29,079 INFO [finetune.py:976] (3/7) Epoch 24, batch 3800, loss[loss=0.1558, simple_loss=0.2333, pruned_loss=0.0392, over 4909.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2433, pruned_loss=0.04853, over 952184.74 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:53:08,363 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135566.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:53:08,917 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.795e+01 1.517e+02 1.799e+02 2.182e+02 4.922e+02, threshold=3.597e+02, percent-clipped=3.0 +2023-04-27 20:53:20,330 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 20:53:32,883 INFO [finetune.py:976] (3/7) Epoch 24, batch 3850, loss[loss=0.1914, simple_loss=0.2597, pruned_loss=0.06158, over 4903.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2423, pruned_loss=0.04811, over 951800.39 frames. ], batch size: 35, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:54:08,516 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0709, 1.9790, 1.8708, 1.6857, 2.1534, 1.7910, 2.7449, 1.6994], + device='cuda:3'), covar=tensor([0.3722, 0.2064, 0.4621, 0.3107, 0.1707, 0.2643, 0.1410, 0.4520], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0352, 0.0424, 0.0351, 0.0378, 0.0376, 0.0368, 0.0421], + device='cuda:3'), out_proj_covar=tensor([9.9941e-05, 1.0504e-04, 1.2868e-04, 1.0544e-04, 1.1230e-04, 1.1203e-04, + 1.0801e-04, 1.2675e-04], device='cuda:3') +2023-04-27 20:54:28,821 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135630.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:54:33,843 INFO [finetune.py:976] (3/7) Epoch 24, batch 3900, loss[loss=0.1608, simple_loss=0.236, pruned_loss=0.04284, over 4910.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2404, pruned_loss=0.04761, over 954106.77 frames. ], batch size: 35, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:55:15,856 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.536e+02 1.810e+02 2.224e+02 3.785e+02, threshold=3.619e+02, percent-clipped=2.0 +2023-04-27 20:55:44,884 INFO [finetune.py:976] (3/7) Epoch 24, batch 3950, loss[loss=0.1483, simple_loss=0.2154, pruned_loss=0.04056, over 4825.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2374, pruned_loss=0.0471, over 953348.82 frames. ], batch size: 30, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:55:47,435 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135691.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:56:19,343 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1958, 2.5592, 2.1916, 2.5656, 1.7563, 2.2214, 2.1996, 1.6904], + device='cuda:3'), covar=tensor([0.1791, 0.1026, 0.0741, 0.0971, 0.3474, 0.1072, 0.1917, 0.2695], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0299, 0.0214, 0.0276, 0.0312, 0.0254, 0.0249, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1337e-04, 1.1783e-04, 8.4362e-05, 1.0875e-04, 1.2594e-04, 9.9947e-05, + 1.0053e-04, 1.0400e-04], device='cuda:3') +2023-04-27 20:56:44,335 INFO [finetune.py:976] (3/7) Epoch 24, batch 4000, loss[loss=0.1509, simple_loss=0.23, pruned_loss=0.03588, over 4742.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.237, pruned_loss=0.04688, over 956017.91 frames. ], batch size: 27, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:57:22,843 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.579e+02 1.856e+02 2.321e+02 4.338e+02, threshold=3.712e+02, percent-clipped=1.0 +2023-04-27 20:57:52,037 INFO [finetune.py:976] (3/7) Epoch 24, batch 4050, loss[loss=0.1981, simple_loss=0.2637, pruned_loss=0.0662, over 4824.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2407, pruned_loss=0.04808, over 955993.29 frames. ], batch size: 39, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:57:56,894 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4490, 2.5680, 2.1555, 2.3131, 2.6135, 2.3285, 3.5147, 2.0221], + device='cuda:3'), covar=tensor([0.3879, 0.2245, 0.4603, 0.3251, 0.1842, 0.2288, 0.1262, 0.4124], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0354, 0.0428, 0.0352, 0.0379, 0.0378, 0.0369, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 20:58:05,489 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6577, 1.5394, 2.0269, 2.0544, 1.4923, 1.3375, 1.5670, 0.9790], + device='cuda:3'), covar=tensor([0.0471, 0.0632, 0.0302, 0.0541, 0.0658, 0.1022, 0.0600, 0.0576], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0074, 0.0095, 0.0073, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 20:58:26,672 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7821, 2.1548, 1.8840, 2.1686, 1.5126, 1.8950, 1.8448, 1.4778], + device='cuda:3'), covar=tensor([0.1691, 0.1121, 0.0818, 0.0976, 0.3323, 0.0935, 0.1718, 0.2435], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0299, 0.0215, 0.0276, 0.0312, 0.0254, 0.0249, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1317e-04, 1.1821e-04, 8.4523e-05, 1.0871e-04, 1.2603e-04, 9.9961e-05, + 1.0046e-04, 1.0397e-04], device='cuda:3') +2023-04-27 20:58:39,906 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135827.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:58:47,827 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4418, 1.1989, 1.2302, 1.1805, 1.6086, 1.3530, 1.1653, 1.1059], + device='cuda:3'), covar=tensor([0.1470, 0.1100, 0.1533, 0.1294, 0.0711, 0.1203, 0.1522, 0.2095], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0304, 0.0345, 0.0280, 0.0322, 0.0301, 0.0294, 0.0366], + device='cuda:3'), out_proj_covar=tensor([6.3035e-05, 6.2545e-05, 7.2528e-05, 5.6198e-05, 6.6187e-05, 6.2992e-05, + 6.1169e-05, 7.7659e-05], device='cuda:3') +2023-04-27 20:58:52,538 INFO [finetune.py:976] (3/7) Epoch 24, batch 4100, loss[loss=0.1921, simple_loss=0.2738, pruned_loss=0.05523, over 4804.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2427, pruned_loss=0.04814, over 957398.70 frames. ], batch size: 40, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:59:20,815 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.04 vs. limit=5.0 +2023-04-27 20:59:21,924 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 20:59:24,286 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7824, 1.2965, 1.8483, 2.2717, 1.8953, 1.7784, 1.8433, 1.7544], + device='cuda:3'), covar=tensor([0.4143, 0.6512, 0.6263, 0.5761, 0.5553, 0.6980, 0.7311, 0.8227], + device='cuda:3'), in_proj_covar=tensor([0.0434, 0.0417, 0.0509, 0.0506, 0.0465, 0.0497, 0.0501, 0.0513], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 20:59:28,241 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.687e+02 1.984e+02 2.360e+02 5.004e+02, threshold=3.968e+02, percent-clipped=3.0 +2023-04-27 20:59:33,115 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135875.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:59:40,397 INFO [finetune.py:976] (3/7) Epoch 24, batch 4150, loss[loss=0.2186, simple_loss=0.2846, pruned_loss=0.07631, over 4819.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2436, pruned_loss=0.04845, over 956536.51 frames. ], batch size: 39, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:00:14,127 INFO [finetune.py:976] (3/7) Epoch 24, batch 4200, loss[loss=0.1511, simple_loss=0.2216, pruned_loss=0.04031, over 4810.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2433, pruned_loss=0.04811, over 955186.19 frames. ], batch size: 39, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:00:35,177 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.525e+02 1.774e+02 2.091e+02 5.050e+02, threshold=3.548e+02, percent-clipped=3.0 +2023-04-27 21:00:44,761 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135982.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:00:47,169 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135986.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:00:47,703 INFO [finetune.py:976] (3/7) Epoch 24, batch 4250, loss[loss=0.164, simple_loss=0.2363, pruned_loss=0.04587, over 4823.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2426, pruned_loss=0.04833, over 954870.40 frames. ], batch size: 30, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:00:52,088 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135994.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:01:22,185 INFO [finetune.py:976] (3/7) Epoch 24, batch 4300, loss[loss=0.1629, simple_loss=0.2293, pruned_loss=0.0482, over 4799.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2401, pruned_loss=0.04784, over 955460.09 frames. ], batch size: 25, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:01:26,013 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136043.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:01:34,849 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:01:42,938 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.901e+01 1.479e+02 1.753e+02 2.087e+02 3.589e+02, threshold=3.506e+02, percent-clipped=1.0 +2023-04-27 21:01:55,530 INFO [finetune.py:976] (3/7) Epoch 24, batch 4350, loss[loss=0.1454, simple_loss=0.2189, pruned_loss=0.03592, over 4725.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2375, pruned_loss=0.04723, over 956416.06 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:01:55,856 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 21:02:33,985 INFO [finetune.py:976] (3/7) Epoch 24, batch 4400, loss[loss=0.161, simple_loss=0.2383, pruned_loss=0.04188, over 4830.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2384, pruned_loss=0.04789, over 953434.32 frames. ], batch size: 30, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:02:41,372 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6050, 2.3355, 2.6674, 3.1063, 2.9540, 2.5305, 2.3592, 2.8129], + device='cuda:3'), covar=tensor([0.0835, 0.1003, 0.0562, 0.0485, 0.0558, 0.0795, 0.0654, 0.0434], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0205, 0.0187, 0.0176, 0.0180, 0.0181, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:02:42,594 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136142.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:02:52,792 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 21:03:16,195 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.622e+02 1.920e+02 2.374e+02 4.375e+02, threshold=3.840e+02, percent-clipped=3.0 +2023-04-27 21:03:40,889 INFO [finetune.py:976] (3/7) Epoch 24, batch 4450, loss[loss=0.1506, simple_loss=0.2355, pruned_loss=0.03288, over 4881.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2406, pruned_loss=0.048, over 953077.82 frames. ], batch size: 32, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:04:02,272 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136203.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:04:13,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2597, 2.9098, 0.9789, 1.7452, 2.2246, 1.3248, 4.0052, 2.0151], + device='cuda:3'), covar=tensor([0.0677, 0.0819, 0.0837, 0.1184, 0.0511, 0.1052, 0.0246, 0.0608], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0048, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 21:04:53,811 INFO [finetune.py:976] (3/7) Epoch 24, batch 4500, loss[loss=0.1966, simple_loss=0.2706, pruned_loss=0.0613, over 4831.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2427, pruned_loss=0.04828, over 953146.00 frames. ], batch size: 47, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:05:04,671 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 21:05:05,181 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4093, 2.2057, 2.5324, 2.8670, 2.7439, 2.2428, 1.9870, 2.5492], + device='cuda:3'), covar=tensor([0.0847, 0.0985, 0.0626, 0.0565, 0.0621, 0.0862, 0.0756, 0.0543], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0204, 0.0187, 0.0175, 0.0179, 0.0180, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:05:07,025 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7011, 2.0507, 1.8829, 2.0157, 1.4865, 1.7262, 1.6416, 1.3086], + device='cuda:3'), covar=tensor([0.1718, 0.1034, 0.0734, 0.1071, 0.3427, 0.1010, 0.1984, 0.2317], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0298, 0.0215, 0.0275, 0.0312, 0.0255, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1323e-04, 1.1757e-04, 8.4674e-05, 1.0853e-04, 1.2575e-04, 1.0035e-04, + 1.0039e-04, 1.0372e-04], device='cuda:3') +2023-04-27 21:05:15,798 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.36 vs. limit=5.0 +2023-04-27 21:05:29,628 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.760e+01 1.501e+02 1.833e+02 2.147e+02 4.216e+02, threshold=3.666e+02, percent-clipped=1.0 +2023-04-27 21:05:42,160 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6602, 1.3952, 1.2932, 1.4332, 1.8781, 1.4876, 1.3212, 1.2208], + device='cuda:3'), covar=tensor([0.1789, 0.1415, 0.1891, 0.1516, 0.0825, 0.1880, 0.2113, 0.2571], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0306, 0.0349, 0.0283, 0.0324, 0.0303, 0.0295, 0.0370], + device='cuda:3'), out_proj_covar=tensor([6.3581e-05, 6.2945e-05, 7.3388e-05, 5.6767e-05, 6.6518e-05, 6.3447e-05, + 6.1409e-05, 7.8486e-05], device='cuda:3') +2023-04-27 21:05:58,029 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136286.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:05:58,526 INFO [finetune.py:976] (3/7) Epoch 24, batch 4550, loss[loss=0.161, simple_loss=0.2327, pruned_loss=0.04462, over 4827.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2435, pruned_loss=0.04814, over 952027.34 frames. ], batch size: 49, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:06:56,483 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 21:06:56,737 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136334.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:06:57,403 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136335.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:07:04,272 INFO [finetune.py:976] (3/7) Epoch 24, batch 4600, loss[loss=0.1557, simple_loss=0.2209, pruned_loss=0.04528, over 4216.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2431, pruned_loss=0.0483, over 953005.53 frames. ], batch size: 18, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:07:04,993 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136338.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:07:18,412 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136350.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:07:40,420 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.579e+02 1.858e+02 2.127e+02 3.942e+02, threshold=3.716e+02, percent-clipped=1.0 +2023-04-27 21:08:03,894 INFO [finetune.py:976] (3/7) Epoch 24, batch 4650, loss[loss=0.1432, simple_loss=0.2221, pruned_loss=0.03214, over 4890.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2399, pruned_loss=0.04746, over 953675.49 frames. ], batch size: 35, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:08:20,793 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:08:26,277 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1389, 1.8827, 2.1382, 2.4812, 2.4336, 2.0595, 1.7791, 2.2476], + device='cuda:3'), covar=tensor([0.0807, 0.1028, 0.0672, 0.0527, 0.0558, 0.0784, 0.0717, 0.0532], + device='cuda:3'), in_proj_covar=tensor([0.0189, 0.0204, 0.0187, 0.0175, 0.0179, 0.0180, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:08:56,538 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.50 vs. limit=5.0 +2023-04-27 21:09:13,810 INFO [finetune.py:976] (3/7) Epoch 24, batch 4700, loss[loss=0.1569, simple_loss=0.2352, pruned_loss=0.03926, over 4894.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2375, pruned_loss=0.04683, over 954336.37 frames. ], batch size: 35, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:09:16,373 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2604, 1.3566, 3.8331, 3.6078, 3.3882, 3.6497, 3.6410, 3.3959], + device='cuda:3'), covar=tensor([0.6825, 0.5035, 0.1075, 0.1492, 0.1112, 0.1844, 0.1908, 0.1500], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0306, 0.0406, 0.0408, 0.0351, 0.0410, 0.0319, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:09:45,097 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.039e+02 1.527e+02 1.865e+02 2.204e+02 5.324e+02, threshold=3.729e+02, percent-clipped=2.0 +2023-04-27 21:09:58,843 INFO [finetune.py:976] (3/7) Epoch 24, batch 4750, loss[loss=0.1802, simple_loss=0.2609, pruned_loss=0.04977, over 4843.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2365, pruned_loss=0.04663, over 954819.93 frames. ], batch size: 49, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:10:07,139 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136498.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:10:24,988 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1798, 1.6296, 1.4231, 2.0309, 2.2470, 1.8044, 1.8251, 1.5247], + device='cuda:3'), covar=tensor([0.1885, 0.1738, 0.2039, 0.1520, 0.1102, 0.1964, 0.1975, 0.2494], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0306, 0.0350, 0.0284, 0.0325, 0.0304, 0.0297, 0.0371], + device='cuda:3'), out_proj_covar=tensor([6.3759e-05, 6.3065e-05, 7.3638e-05, 5.6960e-05, 6.6889e-05, 6.3649e-05, + 6.1736e-05, 7.8716e-05], device='cuda:3') +2023-04-27 21:10:47,888 INFO [finetune.py:976] (3/7) Epoch 24, batch 4800, loss[loss=0.1634, simple_loss=0.2391, pruned_loss=0.04383, over 4831.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2398, pruned_loss=0.04771, over 957159.69 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:11:29,607 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.561e+02 1.788e+02 2.083e+02 3.546e+02, threshold=3.576e+02, percent-clipped=0.0 +2023-04-27 21:11:53,502 INFO [finetune.py:976] (3/7) Epoch 24, batch 4850, loss[loss=0.2093, simple_loss=0.2786, pruned_loss=0.07, over 4884.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2434, pruned_loss=0.04829, over 957297.57 frames. ], batch size: 35, lr: 3.04e-03, grad_scale: 64.0 +2023-04-27 21:12:00,728 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2632, 1.7364, 2.1485, 2.5214, 2.1584, 1.6637, 1.5154, 1.9674], + device='cuda:3'), covar=tensor([0.3030, 0.3163, 0.1456, 0.2087, 0.2441, 0.2523, 0.3759, 0.1831], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0243, 0.0227, 0.0312, 0.0220, 0.0233, 0.0226, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 21:12:10,501 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5769, 1.8013, 1.4707, 1.1750, 1.2270, 1.2120, 1.4885, 1.1746], + device='cuda:3'), covar=tensor([0.1687, 0.1353, 0.1456, 0.1705, 0.2344, 0.1845, 0.1016, 0.2017], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0211, 0.0169, 0.0206, 0.0201, 0.0186, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 21:12:15,282 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1594, 1.5210, 1.4535, 1.6083, 1.5658, 1.8483, 1.3664, 3.3850], + device='cuda:3'), covar=tensor([0.0602, 0.0833, 0.0753, 0.1236, 0.0635, 0.0541, 0.0733, 0.0150], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 21:12:26,524 INFO [finetune.py:976] (3/7) Epoch 24, batch 4900, loss[loss=0.1671, simple_loss=0.2426, pruned_loss=0.04585, over 4811.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2443, pruned_loss=0.04869, over 955902.82 frames. ], batch size: 39, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:12:27,228 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136638.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:12:30,258 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:12:35,595 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:12:46,886 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.604e+02 1.891e+02 2.229e+02 4.573e+02, threshold=3.781e+02, percent-clipped=1.0 +2023-04-27 21:12:58,274 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136686.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:12:58,835 INFO [finetune.py:976] (3/7) Epoch 24, batch 4950, loss[loss=0.2225, simple_loss=0.2887, pruned_loss=0.07813, over 4095.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2463, pruned_loss=0.0491, over 956730.62 frames. ], batch size: 65, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:13:02,153 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136691.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:13:07,456 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136698.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:13:11,651 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:13:18,245 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.5729, 1.3792, 1.3865, 1.0061, 1.4005, 1.2023, 1.6949, 1.3271], + device='cuda:3'), covar=tensor([0.3131, 0.1731, 0.4820, 0.2466, 0.1416, 0.1998, 0.1618, 0.4117], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0349, 0.0421, 0.0348, 0.0375, 0.0373, 0.0365, 0.0419], + device='cuda:3'), out_proj_covar=tensor([9.9383e-05, 1.0434e-04, 1.2766e-04, 1.0445e-04, 1.1130e-04, 1.1100e-04, + 1.0711e-04, 1.2597e-04], device='cuda:3') +2023-04-27 21:13:20,540 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7510, 2.0224, 1.9039, 2.6570, 2.7603, 2.2909, 2.3251, 1.9683], + device='cuda:3'), covar=tensor([0.1821, 0.1748, 0.1843, 0.1738, 0.1161, 0.1634, 0.1892, 0.2363], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0308, 0.0351, 0.0284, 0.0326, 0.0304, 0.0298, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.4191e-05, 6.3369e-05, 7.3853e-05, 5.7042e-05, 6.6980e-05, 6.3756e-05, + 6.1909e-05, 7.8847e-05], device='cuda:3') +2023-04-27 21:13:23,888 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 21:13:32,460 INFO [finetune.py:976] (3/7) Epoch 24, batch 5000, loss[loss=0.1696, simple_loss=0.2466, pruned_loss=0.04626, over 4926.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2455, pruned_loss=0.04911, over 956426.57 frames. ], batch size: 33, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:13:53,643 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.571e+02 1.815e+02 2.179e+02 3.304e+02, threshold=3.630e+02, percent-clipped=0.0 +2023-04-27 21:14:05,880 INFO [finetune.py:976] (3/7) Epoch 24, batch 5050, loss[loss=0.1587, simple_loss=0.2284, pruned_loss=0.04446, over 4922.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2415, pruned_loss=0.04787, over 956333.93 frames. ], batch size: 36, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:14:18,799 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136798.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:14:29,862 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136806.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:15:01,715 INFO [finetune.py:976] (3/7) Epoch 24, batch 5100, loss[loss=0.1522, simple_loss=0.2224, pruned_loss=0.04099, over 4823.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2381, pruned_loss=0.04722, over 956375.02 frames. ], batch size: 39, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:15:07,735 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136846.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:15:10,085 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0692, 4.5041, 1.6253, 2.3197, 2.5855, 3.2554, 2.7266, 1.0245], + device='cuda:3'), covar=tensor([0.1322, 0.0928, 0.1680, 0.1260, 0.1009, 0.0959, 0.1333, 0.2088], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0236, 0.0135, 0.0120, 0.0131, 0.0151, 0.0116, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:15:22,626 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136867.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:15:23,110 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.537e+02 1.873e+02 2.196e+02 3.846e+02, threshold=3.746e+02, percent-clipped=1.0 +2023-04-27 21:15:35,186 INFO [finetune.py:976] (3/7) Epoch 24, batch 5150, loss[loss=0.1781, simple_loss=0.2675, pruned_loss=0.04431, over 4809.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2386, pruned_loss=0.04758, over 957475.23 frames. ], batch size: 40, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:16:16,659 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 21:16:29,341 INFO [finetune.py:976] (3/7) Epoch 24, batch 5200, loss[loss=0.1835, simple_loss=0.2644, pruned_loss=0.05136, over 4896.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.241, pruned_loss=0.04835, over 956010.92 frames. ], batch size: 35, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:16:40,009 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.6875, 4.5441, 2.9789, 5.3735, 4.7575, 4.6632, 2.0299, 4.7158], + device='cuda:3'), covar=tensor([0.1560, 0.1020, 0.3347, 0.0945, 0.3152, 0.1794, 0.5811, 0.1813], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0223, 0.0255, 0.0311, 0.0301, 0.0251, 0.0279, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 21:16:51,332 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 21:16:59,706 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.230e+02 1.748e+02 2.026e+02 2.510e+02 4.483e+02, threshold=4.051e+02, percent-clipped=3.0 +2023-04-27 21:17:22,717 INFO [finetune.py:976] (3/7) Epoch 24, batch 5250, loss[loss=0.1778, simple_loss=0.2479, pruned_loss=0.05381, over 4861.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2435, pruned_loss=0.04906, over 956625.32 frames. ], batch size: 31, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:17:24,147 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-27 21:17:25,238 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136991.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:17:36,066 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136999.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:18:01,204 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 21:18:10,124 INFO [finetune.py:976] (3/7) Epoch 24, batch 5300, loss[loss=0.1788, simple_loss=0.2519, pruned_loss=0.05281, over 4917.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2452, pruned_loss=0.04979, over 957732.42 frames. ], batch size: 38, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:18:11,866 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137039.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:18:19,095 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-27 21:18:30,928 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.584e+02 1.901e+02 2.207e+02 4.060e+02, threshold=3.801e+02, percent-clipped=1.0 +2023-04-27 21:18:44,060 INFO [finetune.py:976] (3/7) Epoch 24, batch 5350, loss[loss=0.1709, simple_loss=0.2442, pruned_loss=0.04886, over 4886.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2442, pruned_loss=0.04892, over 956363.48 frames. ], batch size: 32, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:19:16,917 INFO [finetune.py:976] (3/7) Epoch 24, batch 5400, loss[loss=0.2077, simple_loss=0.275, pruned_loss=0.07021, over 4820.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2416, pruned_loss=0.04819, over 953629.75 frames. ], batch size: 39, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:19:33,099 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=137162.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:19:37,582 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.430e+02 1.764e+02 2.146e+02 4.768e+02, threshold=3.527e+02, percent-clipped=1.0 +2023-04-27 21:19:50,714 INFO [finetune.py:976] (3/7) Epoch 24, batch 5450, loss[loss=0.1842, simple_loss=0.2485, pruned_loss=0.05992, over 4921.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.238, pruned_loss=0.04704, over 954991.55 frames. ], batch size: 37, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:20:03,061 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5835, 1.4213, 1.9192, 1.9246, 1.4571, 1.3541, 1.5215, 0.9331], + device='cuda:3'), covar=tensor([0.0563, 0.0721, 0.0375, 0.0555, 0.0763, 0.1122, 0.0646, 0.0640], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0075, 0.0095, 0.0073, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:20:13,009 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9902, 2.3551, 1.0618, 1.3841, 1.8213, 1.1878, 2.9262, 1.4709], + device='cuda:3'), covar=tensor([0.0668, 0.0589, 0.0768, 0.1170, 0.0443, 0.0947, 0.0224, 0.0631], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0048, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0007, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 21:20:23,702 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.75 vs. limit=5.0 +2023-04-27 21:20:24,732 INFO [finetune.py:976] (3/7) Epoch 24, batch 5500, loss[loss=0.1469, simple_loss=0.2055, pruned_loss=0.04416, over 4766.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2357, pruned_loss=0.04634, over 954536.05 frames. ], batch size: 27, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:20:29,781 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9514, 2.6097, 1.9393, 1.8786, 1.3966, 1.4519, 2.0136, 1.3634], + device='cuda:3'), covar=tensor([0.1606, 0.1281, 0.1258, 0.1594, 0.2208, 0.1844, 0.0945, 0.1934], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0210, 0.0168, 0.0203, 0.0198, 0.0184, 0.0156, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 21:20:31,563 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0788, 1.7443, 1.8917, 2.3673, 2.3911, 1.9602, 1.6861, 2.1443], + device='cuda:3'), covar=tensor([0.0761, 0.1078, 0.0824, 0.0537, 0.0554, 0.0784, 0.0730, 0.0521], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0203, 0.0187, 0.0174, 0.0180, 0.0179, 0.0153, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:20:44,123 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.686e+01 1.504e+02 1.776e+02 2.176e+02 3.969e+02, threshold=3.551e+02, percent-clipped=4.0 +2023-04-27 21:20:57,635 INFO [finetune.py:976] (3/7) Epoch 24, batch 5550, loss[loss=0.1712, simple_loss=0.2413, pruned_loss=0.05054, over 4759.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2375, pruned_loss=0.047, over 952869.99 frames. ], batch size: 27, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:21:02,597 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0492, 1.4184, 1.2675, 1.8771, 1.5378, 1.8999, 1.3985, 4.1704], + device='cuda:3'), covar=tensor([0.0766, 0.1115, 0.1128, 0.1351, 0.0880, 0.0725, 0.0994, 0.0177], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 21:21:05,071 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=137299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:21:34,762 INFO [finetune.py:976] (3/7) Epoch 24, batch 5600, loss[loss=0.197, simple_loss=0.2684, pruned_loss=0.06279, over 4818.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2415, pruned_loss=0.04814, over 952668.44 frames. ], batch size: 51, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:21:46,035 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137347.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:22:14,764 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.541e+02 1.789e+02 2.150e+02 5.679e+02, threshold=3.578e+02, percent-clipped=2.0 +2023-04-27 21:22:37,498 INFO [finetune.py:976] (3/7) Epoch 24, batch 5650, loss[loss=0.1189, simple_loss=0.1864, pruned_loss=0.02571, over 4198.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2439, pruned_loss=0.04861, over 949760.21 frames. ], batch size: 18, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:23:26,255 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-04-27 21:23:28,789 INFO [finetune.py:976] (3/7) Epoch 24, batch 5700, loss[loss=0.1408, simple_loss=0.2085, pruned_loss=0.03653, over 4256.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2411, pruned_loss=0.04829, over 934749.55 frames. ], batch size: 18, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:23:43,835 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=137462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:23:57,748 INFO [finetune.py:976] (3/7) Epoch 25, batch 0, loss[loss=0.2091, simple_loss=0.2708, pruned_loss=0.07371, over 4814.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2708, pruned_loss=0.07371, over 4814.00 frames. ], batch size: 33, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:23:57,748 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 21:24:02,355 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7502, 2.1264, 1.8319, 2.0853, 1.6346, 1.7028, 1.7850, 1.3240], + device='cuda:3'), covar=tensor([0.1871, 0.1265, 0.0887, 0.1035, 0.4059, 0.1162, 0.1914, 0.2767], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0299, 0.0215, 0.0276, 0.0313, 0.0255, 0.0250, 0.0264], + device='cuda:3'), out_proj_covar=tensor([1.1367e-04, 1.1800e-04, 8.4676e-05, 1.0887e-04, 1.2620e-04, 1.0035e-04, + 1.0069e-04, 1.0427e-04], device='cuda:3') +2023-04-27 21:24:08,111 INFO [finetune.py:1010] (3/7) Epoch 25, validation: loss=0.155, simple_loss=0.224, pruned_loss=0.04295, over 2265189.00 frames. +2023-04-27 21:24:08,112 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 21:24:09,917 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.060e+01 1.467e+02 1.816e+02 2.337e+02 4.208e+02, threshold=3.632e+02, percent-clipped=2.0 +2023-04-27 21:24:37,393 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137510.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:24:40,446 INFO [finetune.py:976] (3/7) Epoch 25, batch 50, loss[loss=0.13, simple_loss=0.2026, pruned_loss=0.02867, over 4752.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2446, pruned_loss=0.04825, over 217001.74 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:25:13,441 INFO [finetune.py:976] (3/7) Epoch 25, batch 100, loss[loss=0.1678, simple_loss=0.2357, pruned_loss=0.04998, over 4827.00 frames. ], tot_loss[loss=0.168, simple_loss=0.24, pruned_loss=0.04795, over 381254.55 frames. ], batch size: 30, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:25:15,238 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.734e+01 1.567e+02 1.856e+02 2.214e+02 3.687e+02, threshold=3.711e+02, percent-clipped=3.0 +2023-04-27 21:25:46,404 INFO [finetune.py:976] (3/7) Epoch 25, batch 150, loss[loss=0.2039, simple_loss=0.2654, pruned_loss=0.07118, over 4863.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.236, pruned_loss=0.04731, over 508605.09 frames. ], batch size: 31, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:25:48,352 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5364, 3.5044, 1.0939, 1.9420, 1.9225, 2.4702, 1.9968, 0.9985], + device='cuda:3'), covar=tensor([0.1285, 0.0823, 0.1768, 0.1197, 0.1008, 0.0998, 0.1445, 0.2042], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0237, 0.0136, 0.0120, 0.0132, 0.0152, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:25:51,115 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4573, 1.7686, 1.8898, 2.0011, 1.8269, 1.8610, 1.9628, 1.8873], + device='cuda:3'), covar=tensor([0.4017, 0.5018, 0.3925, 0.3807, 0.5258, 0.6508, 0.4686, 0.4198], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0373, 0.0325, 0.0339, 0.0348, 0.0393, 0.0357, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 21:25:53,375 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3128, 2.9515, 0.9686, 1.6412, 2.0133, 1.2686, 3.8156, 1.7607], + device='cuda:3'), covar=tensor([0.0646, 0.0772, 0.0937, 0.1195, 0.0519, 0.0945, 0.0186, 0.0600], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0048, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0007, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 21:26:20,277 INFO [finetune.py:976] (3/7) Epoch 25, batch 200, loss[loss=0.1777, simple_loss=0.2482, pruned_loss=0.05367, over 4750.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2358, pruned_loss=0.04716, over 607657.01 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:26:22,071 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.527e+02 1.776e+02 2.195e+02 3.612e+02, threshold=3.551e+02, percent-clipped=0.0 +2023-04-27 21:26:33,144 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 21:27:09,383 INFO [finetune.py:976] (3/7) Epoch 25, batch 250, loss[loss=0.1879, simple_loss=0.2676, pruned_loss=0.05411, over 4738.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2359, pruned_loss=0.04665, over 685806.30 frames. ], batch size: 59, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:27:19,172 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5237, 1.8210, 1.9602, 2.0463, 1.8799, 1.9510, 2.0476, 1.9897], + device='cuda:3'), covar=tensor([0.3759, 0.4793, 0.3805, 0.3945, 0.5121, 0.6400, 0.4565, 0.4054], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0372, 0.0324, 0.0338, 0.0348, 0.0392, 0.0356, 0.0329], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 21:28:04,799 INFO [finetune.py:976] (3/7) Epoch 25, batch 300, loss[loss=0.152, simple_loss=0.2249, pruned_loss=0.0395, over 4774.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2398, pruned_loss=0.04736, over 746701.20 frames. ], batch size: 26, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:28:05,562 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6741, 1.0991, 1.7258, 2.1598, 1.7416, 1.6035, 1.6963, 1.6220], + device='cuda:3'), covar=tensor([0.4309, 0.6810, 0.6135, 0.5420, 0.5570, 0.7567, 0.7618, 0.8639], + device='cuda:3'), in_proj_covar=tensor([0.0436, 0.0420, 0.0512, 0.0508, 0.0466, 0.0500, 0.0503, 0.0516], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:28:06,622 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.558e+02 1.872e+02 2.368e+02 4.247e+02, threshold=3.743e+02, percent-clipped=2.0 +2023-04-27 21:28:43,570 INFO [finetune.py:976] (3/7) Epoch 25, batch 350, loss[loss=0.1754, simple_loss=0.2486, pruned_loss=0.05106, over 4732.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2423, pruned_loss=0.04842, over 793053.67 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:29:09,500 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-04-27 21:29:21,925 INFO [finetune.py:976] (3/7) Epoch 25, batch 400, loss[loss=0.1355, simple_loss=0.2089, pruned_loss=0.0311, over 4761.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2423, pruned_loss=0.04848, over 829376.82 frames. ], batch size: 28, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:29:29,043 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.617e+02 1.875e+02 2.197e+02 4.707e+02, threshold=3.750e+02, percent-clipped=1.0 +2023-04-27 21:30:01,129 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137892.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:30:02,937 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137895.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:30:27,172 INFO [finetune.py:976] (3/7) Epoch 25, batch 450, loss[loss=0.1867, simple_loss=0.2543, pruned_loss=0.05953, over 4910.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2418, pruned_loss=0.04869, over 857305.67 frames. ], batch size: 35, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:30:29,000 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6258, 2.9853, 1.0925, 1.7929, 2.4620, 1.6297, 4.3752, 1.9353], + device='cuda:3'), covar=tensor([0.0607, 0.0667, 0.0791, 0.1261, 0.0460, 0.0982, 0.0211, 0.0622], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0048, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0007, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 21:31:27,311 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137953.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:31:33,976 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137956.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:31:36,571 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-27 21:31:44,823 INFO [finetune.py:976] (3/7) Epoch 25, batch 500, loss[loss=0.1673, simple_loss=0.2395, pruned_loss=0.04759, over 4921.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2393, pruned_loss=0.04743, over 879003.67 frames. ], batch size: 37, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:31:46,604 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.991e+01 1.587e+02 1.886e+02 2.282e+02 3.867e+02, threshold=3.771e+02, percent-clipped=1.0 +2023-04-27 21:32:09,718 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2259, 1.5967, 1.5355, 1.7985, 1.8102, 1.9103, 1.4792, 3.5767], + device='cuda:3'), covar=tensor([0.0641, 0.0826, 0.0829, 0.1222, 0.0633, 0.0517, 0.0724, 0.0132], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 21:32:30,090 INFO [finetune.py:976] (3/7) Epoch 25, batch 550, loss[loss=0.1985, simple_loss=0.2686, pruned_loss=0.06418, over 4915.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2358, pruned_loss=0.04605, over 896033.10 frames. ], batch size: 37, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:32:32,682 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9963, 1.5293, 1.8199, 1.7149, 1.8532, 1.4787, 0.7846, 1.4804], + device='cuda:3'), covar=tensor([0.3170, 0.3036, 0.1679, 0.2104, 0.2464, 0.2621, 0.4129, 0.1994], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0246, 0.0228, 0.0314, 0.0222, 0.0235, 0.0228, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 21:32:50,135 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-04-27 21:33:20,386 INFO [finetune.py:976] (3/7) Epoch 25, batch 600, loss[loss=0.1394, simple_loss=0.2031, pruned_loss=0.03786, over 3502.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2362, pruned_loss=0.04643, over 907968.08 frames. ], batch size: 15, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:33:22,209 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.469e+02 1.786e+02 2.001e+02 2.950e+02, threshold=3.571e+02, percent-clipped=0.0 +2023-04-27 21:33:31,199 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7667, 1.3522, 1.8806, 2.2851, 1.8626, 1.7471, 1.8547, 1.7518], + device='cuda:3'), covar=tensor([0.4270, 0.6671, 0.5826, 0.5427, 0.5779, 0.7570, 0.7508, 0.8893], + device='cuda:3'), in_proj_covar=tensor([0.0436, 0.0420, 0.0511, 0.0507, 0.0466, 0.0500, 0.0502, 0.0516], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:33:52,672 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6947, 2.0115, 1.7564, 1.9944, 1.4661, 1.7381, 1.6592, 1.3439], + device='cuda:3'), covar=tensor([0.1554, 0.1172, 0.0808, 0.1046, 0.3304, 0.1102, 0.1774, 0.2261], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0299, 0.0213, 0.0276, 0.0311, 0.0254, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1304e-04, 1.1822e-04, 8.3951e-05, 1.0875e-04, 1.2531e-04, 1.0016e-04, + 1.0014e-04, 1.0399e-04], device='cuda:3') +2023-04-27 21:33:53,155 INFO [finetune.py:976] (3/7) Epoch 25, batch 650, loss[loss=0.2419, simple_loss=0.3064, pruned_loss=0.08868, over 4826.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2427, pruned_loss=0.04922, over 918156.29 frames. ], batch size: 40, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:34:26,503 INFO [finetune.py:976] (3/7) Epoch 25, batch 700, loss[loss=0.2264, simple_loss=0.289, pruned_loss=0.08191, over 4132.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2444, pruned_loss=0.04919, over 924468.19 frames. ], batch size: 65, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:34:28,315 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.605e+02 1.838e+02 2.214e+02 4.494e+02, threshold=3.677e+02, percent-clipped=3.0 +2023-04-27 21:35:26,623 INFO [finetune.py:976] (3/7) Epoch 25, batch 750, loss[loss=0.1396, simple_loss=0.2276, pruned_loss=0.02584, over 4814.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2447, pruned_loss=0.04887, over 931903.14 frames. ], batch size: 39, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:36:09,234 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138248.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:36:16,832 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138251.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:36:25,762 INFO [finetune.py:976] (3/7) Epoch 25, batch 800, loss[loss=0.1668, simple_loss=0.2484, pruned_loss=0.04259, over 4848.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2431, pruned_loss=0.04776, over 938194.33 frames. ], batch size: 44, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:36:27,568 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.037e+01 1.577e+02 1.890e+02 2.276e+02 6.092e+02, threshold=3.780e+02, percent-clipped=1.0 +2023-04-27 21:36:28,918 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138270.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:37:09,964 INFO [finetune.py:976] (3/7) Epoch 25, batch 850, loss[loss=0.1656, simple_loss=0.2399, pruned_loss=0.04566, over 4731.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2414, pruned_loss=0.04737, over 940929.27 frames. ], batch size: 59, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:37:14,321 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138322.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:37:24,955 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138331.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:38:00,617 INFO [finetune.py:976] (3/7) Epoch 25, batch 900, loss[loss=0.1788, simple_loss=0.2504, pruned_loss=0.05357, over 4766.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2385, pruned_loss=0.04613, over 945238.12 frames. ], batch size: 26, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:38:02,478 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.023e+01 1.492e+02 1.759e+02 2.095e+02 3.711e+02, threshold=3.518e+02, percent-clipped=0.0 +2023-04-27 21:38:22,005 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9868, 1.8428, 2.3485, 2.4448, 1.8209, 1.4997, 1.9731, 1.2011], + device='cuda:3'), covar=tensor([0.0649, 0.0646, 0.0416, 0.0582, 0.0749, 0.1177, 0.0659, 0.0686], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0068, 0.0075, 0.0096, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:38:22,008 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138383.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:38:23,203 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1525, 2.5814, 1.1433, 1.5894, 2.1938, 1.1936, 3.5577, 1.7300], + device='cuda:3'), covar=tensor([0.0649, 0.0665, 0.0756, 0.1151, 0.0440, 0.1009, 0.0240, 0.0626], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0046, 0.0045, 0.0048, 0.0050, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0007, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 21:39:01,961 INFO [finetune.py:976] (3/7) Epoch 25, batch 950, loss[loss=0.1406, simple_loss=0.2134, pruned_loss=0.03383, over 4823.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2372, pruned_loss=0.04645, over 948924.68 frames. ], batch size: 41, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:39:03,322 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9861, 1.8318, 1.7213, 1.5576, 2.0524, 1.6072, 2.4821, 1.5787], + device='cuda:3'), covar=tensor([0.3390, 0.1951, 0.4773, 0.2992, 0.1670, 0.2478, 0.1451, 0.4513], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0352, 0.0428, 0.0349, 0.0378, 0.0376, 0.0369, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:39:16,789 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9082, 3.1545, 1.2035, 1.9730, 2.6989, 1.8329, 4.3873, 2.3810], + device='cuda:3'), covar=tensor([0.0512, 0.0647, 0.0747, 0.1118, 0.0422, 0.0887, 0.0227, 0.0538], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0045, 0.0048, 0.0050, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0007, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 21:39:25,805 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6676, 1.5609, 2.0070, 2.0619, 1.4731, 1.3494, 1.5783, 1.0323], + device='cuda:3'), covar=tensor([0.0558, 0.0759, 0.0395, 0.0653, 0.0745, 0.1136, 0.0679, 0.0703], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:39:46,002 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7468, 2.4034, 1.2081, 1.4306, 2.4850, 1.5518, 1.5194, 1.6603], + device='cuda:3'), covar=tensor([0.0632, 0.0304, 0.0292, 0.0625, 0.0224, 0.0653, 0.0628, 0.0631], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 21:40:06,146 INFO [finetune.py:976] (3/7) Epoch 25, batch 1000, loss[loss=0.2121, simple_loss=0.2744, pruned_loss=0.0749, over 4759.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2383, pruned_loss=0.04722, over 950730.53 frames. ], batch size: 59, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:40:07,974 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.516e+02 1.776e+02 2.068e+02 3.891e+02, threshold=3.551e+02, percent-clipped=2.0 +2023-04-27 21:41:00,724 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0065, 1.6082, 2.0812, 2.4834, 2.0392, 1.9734, 2.0172, 1.9481], + device='cuda:3'), covar=tensor([0.4684, 0.7182, 0.6688, 0.5376, 0.6058, 0.8185, 0.8193, 0.8708], + device='cuda:3'), in_proj_covar=tensor([0.0437, 0.0421, 0.0513, 0.0509, 0.0467, 0.0502, 0.0504, 0.0518], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:41:09,819 INFO [finetune.py:976] (3/7) Epoch 25, batch 1050, loss[loss=0.1626, simple_loss=0.2485, pruned_loss=0.03832, over 4808.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2399, pruned_loss=0.04683, over 951262.36 frames. ], batch size: 41, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:41:10,632 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 21:41:52,745 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138548.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:41:53,381 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5670, 1.9390, 0.8778, 1.3470, 1.8321, 1.4220, 1.3961, 1.4205], + device='cuda:3'), covar=tensor([0.0564, 0.0274, 0.0338, 0.0557, 0.0269, 0.0601, 0.0553, 0.0586], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 21:41:55,018 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138551.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:42:14,665 INFO [finetune.py:976] (3/7) Epoch 25, batch 1100, loss[loss=0.1891, simple_loss=0.2538, pruned_loss=0.06216, over 4881.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2425, pruned_loss=0.04788, over 951317.41 frames. ], batch size: 32, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:42:16,466 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.782e+01 1.591e+02 1.867e+02 2.331e+02 5.511e+02, threshold=3.734e+02, percent-clipped=3.0 +2023-04-27 21:42:55,898 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138596.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:42:57,699 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138599.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:43:16,327 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9766, 1.4528, 1.7861, 1.7172, 1.8052, 1.4581, 0.8489, 1.4544], + device='cuda:3'), covar=tensor([0.3169, 0.2939, 0.1604, 0.1978, 0.2279, 0.2406, 0.3825, 0.1834], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0245, 0.0227, 0.0312, 0.0221, 0.0233, 0.0227, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 21:43:18,594 INFO [finetune.py:976] (3/7) Epoch 25, batch 1150, loss[loss=0.1453, simple_loss=0.2148, pruned_loss=0.03795, over 4816.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2442, pruned_loss=0.04853, over 954494.46 frames. ], batch size: 25, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:43:38,293 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138626.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:44:12,094 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8413, 2.2701, 1.9010, 1.7137, 1.3831, 1.4102, 1.8760, 1.3226], + device='cuda:3'), covar=tensor([0.1682, 0.1402, 0.1370, 0.1685, 0.2374, 0.1933, 0.1037, 0.2103], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0209, 0.0168, 0.0204, 0.0199, 0.0185, 0.0156, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 21:44:23,367 INFO [finetune.py:976] (3/7) Epoch 25, batch 1200, loss[loss=0.1977, simple_loss=0.2592, pruned_loss=0.06809, over 4873.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2426, pruned_loss=0.04797, over 955562.34 frames. ], batch size: 34, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:44:26,074 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.537e+02 1.836e+02 2.355e+02 3.778e+02, threshold=3.672e+02, percent-clipped=1.0 +2023-04-27 21:44:43,178 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138678.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:44:44,975 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138681.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:44:46,889 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1871, 1.7972, 2.2456, 2.6006, 2.2105, 2.0715, 2.1979, 2.0743], + device='cuda:3'), covar=tensor([0.4025, 0.5857, 0.5296, 0.5054, 0.5410, 0.7487, 0.7152, 0.6833], + device='cuda:3'), in_proj_covar=tensor([0.0436, 0.0421, 0.0511, 0.0507, 0.0466, 0.0500, 0.0502, 0.0515], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:44:53,114 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138686.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:45:05,140 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9499, 2.1276, 1.2777, 1.6592, 2.2957, 1.7551, 1.7315, 1.8903], + device='cuda:3'), covar=tensor([0.0448, 0.0343, 0.0270, 0.0492, 0.0215, 0.0472, 0.0467, 0.0510], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 21:45:05,151 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138697.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:45:14,114 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3349, 1.2917, 0.4538, 1.1625, 1.3502, 1.2125, 1.1975, 1.2373], + device='cuda:3'), covar=tensor([0.0550, 0.0301, 0.0411, 0.0506, 0.0313, 0.0511, 0.0510, 0.0571], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 21:45:29,241 INFO [finetune.py:976] (3/7) Epoch 25, batch 1250, loss[loss=0.1766, simple_loss=0.2445, pruned_loss=0.05438, over 4897.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2414, pruned_loss=0.0486, over 952740.68 frames. ], batch size: 46, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:46:07,984 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138742.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:46:11,062 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138747.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:46:23,973 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138758.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:46:32,247 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7010, 1.6822, 0.8061, 1.3872, 1.8320, 1.5487, 1.4273, 1.5299], + device='cuda:3'), covar=tensor([0.0468, 0.0352, 0.0347, 0.0528, 0.0268, 0.0496, 0.0488, 0.0516], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 21:46:33,341 INFO [finetune.py:976] (3/7) Epoch 25, batch 1300, loss[loss=0.184, simple_loss=0.2564, pruned_loss=0.05582, over 4825.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2382, pruned_loss=0.04749, over 951526.85 frames. ], batch size: 39, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:46:40,294 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.540e+02 1.806e+02 2.216e+02 3.604e+02, threshold=3.612e+02, percent-clipped=0.0 +2023-04-27 21:47:44,176 INFO [finetune.py:976] (3/7) Epoch 25, batch 1350, loss[loss=0.1513, simple_loss=0.2442, pruned_loss=0.02915, over 4795.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2394, pruned_loss=0.04803, over 952843.24 frames. ], batch size: 29, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:47:57,448 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5624, 3.0605, 2.7116, 2.8777, 2.1590, 2.5837, 2.6401, 2.0824], + device='cuda:3'), covar=tensor([0.1998, 0.1166, 0.0688, 0.1287, 0.3141, 0.1189, 0.2099, 0.2692], + device='cuda:3'), in_proj_covar=tensor([0.0287, 0.0301, 0.0214, 0.0278, 0.0313, 0.0257, 0.0250, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1460e-04, 1.1869e-04, 8.4422e-05, 1.0972e-04, 1.2631e-04, 1.0128e-04, + 1.0095e-04, 1.0470e-04], device='cuda:3') +2023-04-27 21:48:39,839 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9009, 1.6564, 2.0906, 2.3546, 1.9237, 1.8234, 1.9704, 1.9114], + device='cuda:3'), covar=tensor([0.4368, 0.7490, 0.6998, 0.5575, 0.5958, 0.8515, 0.8476, 0.9692], + device='cuda:3'), in_proj_covar=tensor([0.0435, 0.0420, 0.0510, 0.0506, 0.0464, 0.0499, 0.0500, 0.0514], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:48:48,891 INFO [finetune.py:976] (3/7) Epoch 25, batch 1400, loss[loss=0.1718, simple_loss=0.2561, pruned_loss=0.04369, over 4800.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2428, pruned_loss=0.04896, over 953484.39 frames. ], batch size: 51, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:48:50,719 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.518e+02 1.853e+02 2.271e+02 4.217e+02, threshold=3.705e+02, percent-clipped=3.0 +2023-04-27 21:49:44,990 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:49:54,328 INFO [finetune.py:976] (3/7) Epoch 25, batch 1450, loss[loss=0.1622, simple_loss=0.2473, pruned_loss=0.03854, over 4818.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.244, pruned_loss=0.04906, over 951362.52 frames. ], batch size: 39, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:49:55,584 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7315, 2.4074, 1.8883, 1.9004, 1.2287, 1.2971, 1.9094, 1.2201], + device='cuda:3'), covar=tensor([0.1736, 0.1417, 0.1522, 0.1678, 0.2465, 0.2162, 0.1014, 0.2227], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0209, 0.0168, 0.0204, 0.0199, 0.0186, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 21:50:06,697 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2795, 2.8271, 2.3075, 2.6775, 1.7440, 2.2841, 2.5610, 1.8402], + device='cuda:3'), covar=tensor([0.2034, 0.1103, 0.0885, 0.1196, 0.3882, 0.1191, 0.1949, 0.2543], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0299, 0.0213, 0.0276, 0.0311, 0.0255, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1353e-04, 1.1776e-04, 8.3891e-05, 1.0876e-04, 1.2554e-04, 1.0052e-04, + 1.0010e-04, 1.0378e-04], device='cuda:3') +2023-04-27 21:50:07,877 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138926.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:51:01,096 INFO [finetune.py:976] (3/7) Epoch 25, batch 1500, loss[loss=0.184, simple_loss=0.2579, pruned_loss=0.05506, over 4832.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2449, pruned_loss=0.0491, over 951813.58 frames. ], batch size: 30, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:51:03,883 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.504e+02 1.740e+02 2.143e+02 4.195e+02, threshold=3.481e+02, percent-clipped=2.0 +2023-04-27 21:51:10,097 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138970.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 21:51:12,535 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:51:19,581 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138978.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:02,847 INFO [finetune.py:976] (3/7) Epoch 25, batch 1550, loss[loss=0.2005, simple_loss=0.2565, pruned_loss=0.07227, over 4891.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2456, pruned_loss=0.04958, over 953012.66 frames. ], batch size: 35, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:52:11,152 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139026.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:11,844 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5555, 1.0955, 0.3939, 1.2308, 1.1885, 1.3963, 1.3109, 1.3395], + device='cuda:3'), covar=tensor([0.0493, 0.0405, 0.0411, 0.0549, 0.0306, 0.0491, 0.0514, 0.0551], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 21:52:17,669 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3584, 3.0553, 0.8593, 1.5811, 1.7370, 2.1599, 1.7687, 0.9572], + device='cuda:3'), covar=tensor([0.1345, 0.1041, 0.1915, 0.1344, 0.1075, 0.1017, 0.1567, 0.1903], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0239, 0.0137, 0.0121, 0.0132, 0.0153, 0.0117, 0.0120], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:52:18,851 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139037.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:22,337 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139042.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:23,941 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 21:52:29,078 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139053.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:41,925 INFO [finetune.py:976] (3/7) Epoch 25, batch 1600, loss[loss=0.1565, simple_loss=0.219, pruned_loss=0.04705, over 4889.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2427, pruned_loss=0.04855, over 954312.59 frames. ], batch size: 43, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:52:44,232 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.576e+02 1.832e+02 2.139e+02 4.092e+02, threshold=3.665e+02, percent-clipped=3.0 +2023-04-27 21:53:24,793 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5264, 2.1181, 2.4678, 3.1237, 2.5452, 2.0030, 2.0732, 2.3870], + device='cuda:3'), covar=tensor([0.2854, 0.2679, 0.1474, 0.1958, 0.2330, 0.2371, 0.3280, 0.1823], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0244, 0.0227, 0.0311, 0.0221, 0.0234, 0.0226, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 21:53:47,319 INFO [finetune.py:976] (3/7) Epoch 25, batch 1650, loss[loss=0.1804, simple_loss=0.2369, pruned_loss=0.06197, over 4867.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2401, pruned_loss=0.0479, over 956150.70 frames. ], batch size: 31, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:54:33,489 INFO [finetune.py:976] (3/7) Epoch 25, batch 1700, loss[loss=0.1915, simple_loss=0.2537, pruned_loss=0.06463, over 4901.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2378, pruned_loss=0.04702, over 957199.88 frames. ], batch size: 35, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:54:35,335 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.477e+02 1.739e+02 2.189e+02 4.097e+02, threshold=3.477e+02, percent-clipped=1.0 +2023-04-27 21:54:40,426 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 21:55:07,048 INFO [finetune.py:976] (3/7) Epoch 25, batch 1750, loss[loss=0.2087, simple_loss=0.2862, pruned_loss=0.06561, over 4808.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2385, pruned_loss=0.04704, over 957678.73 frames. ], batch size: 51, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:55:23,348 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139238.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:55:29,611 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 21:55:37,047 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6996, 1.2603, 1.8200, 2.0168, 1.7235, 1.6217, 1.6595, 1.6844], + device='cuda:3'), covar=tensor([0.6041, 0.8721, 0.7890, 0.8748, 0.7219, 1.0452, 0.9770, 1.1604], + device='cuda:3'), in_proj_covar=tensor([0.0437, 0.0420, 0.0510, 0.0508, 0.0466, 0.0500, 0.0502, 0.0516], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:55:40,699 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9292, 1.9351, 2.3025, 2.4632, 1.8211, 1.5533, 1.9478, 1.0470], + device='cuda:3'), covar=tensor([0.0607, 0.0573, 0.0436, 0.0680, 0.0708, 0.1093, 0.0656, 0.0722], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0067, 0.0068, 0.0074, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:55:41,183 INFO [finetune.py:976] (3/7) Epoch 25, batch 1800, loss[loss=0.1743, simple_loss=0.2432, pruned_loss=0.05267, over 4911.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.241, pruned_loss=0.04737, over 957976.79 frames. ], batch size: 37, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:55:41,248 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139265.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 21:55:42,983 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.657e+01 1.551e+02 1.902e+02 2.363e+02 3.513e+02, threshold=3.803e+02, percent-clipped=1.0 +2023-04-27 21:55:44,923 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6737, 3.7567, 1.1156, 1.8106, 2.0858, 2.6047, 1.9625, 1.0145], + device='cuda:3'), covar=tensor([0.1378, 0.1020, 0.2023, 0.1342, 0.1050, 0.1040, 0.1613, 0.1922], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0237, 0.0136, 0.0120, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 21:55:48,006 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139276.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:55:57,491 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:04,612 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:07,558 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3853, 2.6892, 1.0318, 1.6043, 2.1123, 1.5640, 3.5585, 2.1005], + device='cuda:3'), covar=tensor([0.0584, 0.0565, 0.0777, 0.1266, 0.0491, 0.0899, 0.0305, 0.0517], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:3') +2023-04-27 21:56:14,571 INFO [finetune.py:976] (3/7) Epoch 25, batch 1850, loss[loss=0.1505, simple_loss=0.2392, pruned_loss=0.03084, over 4889.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2427, pruned_loss=0.04848, over 956132.40 frames. ], batch size: 36, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:56:25,770 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8645, 1.1549, 4.6320, 4.3385, 4.1004, 4.3216, 4.2206, 4.0461], + device='cuda:3'), covar=tensor([0.7052, 0.6137, 0.1050, 0.1855, 0.1074, 0.1271, 0.2157, 0.1592], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0309, 0.0408, 0.0412, 0.0351, 0.0414, 0.0321, 0.0370], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 21:56:29,234 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:29,283 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139337.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 21:56:33,191 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139342.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:43,863 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139350.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:51,967 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139353.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:57:05,212 INFO [finetune.py:976] (3/7) Epoch 25, batch 1900, loss[loss=0.1959, simple_loss=0.2767, pruned_loss=0.05759, over 4814.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2432, pruned_loss=0.0478, over 955415.42 frames. ], batch size: 40, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 21:57:07,684 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.239e+02 1.510e+02 1.790e+02 2.163e+02 4.429e+02, threshold=3.581e+02, percent-clipped=1.0 +2023-04-27 21:57:13,418 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7432, 1.9641, 1.8523, 2.5394, 2.6635, 2.2007, 2.2003, 1.8015], + device='cuda:3'), covar=tensor([0.1929, 0.1978, 0.2077, 0.2028, 0.1471, 0.2233, 0.2470, 0.2652], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0308, 0.0350, 0.0286, 0.0327, 0.0306, 0.0299, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.3913e-05, 6.3401e-05, 7.3633e-05, 5.7281e-05, 6.7239e-05, 6.4070e-05, + 6.2088e-05, 7.8941e-05], device='cuda:3') +2023-04-27 21:57:28,904 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:57:36,783 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:57:49,497 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139401.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:57:56,056 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6230, 3.6490, 2.7119, 4.2643, 3.6752, 3.6124, 1.5130, 3.6427], + device='cuda:3'), covar=tensor([0.1986, 0.1421, 0.3455, 0.1687, 0.3953, 0.2063, 0.6680, 0.2549], + device='cuda:3'), in_proj_covar=tensor([0.0242, 0.0217, 0.0249, 0.0303, 0.0295, 0.0246, 0.0272, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 21:58:09,750 INFO [finetune.py:976] (3/7) Epoch 25, batch 1950, loss[loss=0.1746, simple_loss=0.2425, pruned_loss=0.05337, over 4795.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2425, pruned_loss=0.04783, over 956401.37 frames. ], batch size: 45, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 21:59:13,359 INFO [finetune.py:976] (3/7) Epoch 25, batch 2000, loss[loss=0.169, simple_loss=0.2349, pruned_loss=0.05152, over 4907.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2415, pruned_loss=0.04811, over 956410.31 frames. ], batch size: 36, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 21:59:13,475 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4555, 1.7429, 1.6057, 1.9752, 2.0343, 2.1328, 1.5526, 3.9358], + device='cuda:3'), covar=tensor([0.0574, 0.0789, 0.0789, 0.1154, 0.0579, 0.0454, 0.0738, 0.0134], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 21:59:15,797 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.572e+02 1.799e+02 2.251e+02 3.942e+02, threshold=3.599e+02, percent-clipped=2.0 +2023-04-27 21:59:25,649 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-27 21:59:56,676 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-27 22:00:17,502 INFO [finetune.py:976] (3/7) Epoch 25, batch 2050, loss[loss=0.1657, simple_loss=0.2409, pruned_loss=0.04527, over 4820.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2376, pruned_loss=0.04652, over 957965.36 frames. ], batch size: 38, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:00:50,413 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 22:01:10,392 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 22:01:21,674 INFO [finetune.py:976] (3/7) Epoch 25, batch 2100, loss[loss=0.1292, simple_loss=0.2008, pruned_loss=0.02878, over 4786.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2362, pruned_loss=0.04586, over 958528.88 frames. ], batch size: 26, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:01:21,784 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139565.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 22:01:24,121 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.088e+01 1.541e+02 1.795e+02 2.157e+02 3.840e+02, threshold=3.589e+02, percent-clipped=1.0 +2023-04-27 22:01:30,513 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139571.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:01:37,251 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139582.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:01:44,578 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139594.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:01:57,941 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:01:59,115 INFO [finetune.py:976] (3/7) Epoch 25, batch 2150, loss[loss=0.2223, simple_loss=0.2886, pruned_loss=0.07804, over 4916.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2397, pruned_loss=0.04737, over 955655.39 frames. ], batch size: 36, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:02:10,545 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139632.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 22:02:10,577 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139632.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:02:17,307 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:02:18,459 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139645.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:02:31,934 INFO [finetune.py:976] (3/7) Epoch 25, batch 2200, loss[loss=0.1619, simple_loss=0.231, pruned_loss=0.04645, over 4761.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2431, pruned_loss=0.04823, over 957144.66 frames. ], batch size: 26, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:02:34,822 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.529e+02 1.754e+02 2.258e+02 5.468e+02, threshold=3.509e+02, percent-clipped=4.0 +2023-04-27 22:03:04,782 INFO [finetune.py:976] (3/7) Epoch 25, batch 2250, loss[loss=0.149, simple_loss=0.2316, pruned_loss=0.03318, over 4818.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2448, pruned_loss=0.04898, over 957124.51 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:03:07,802 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1781, 1.3668, 1.2446, 1.6453, 1.4619, 1.5588, 1.3055, 2.9791], + device='cuda:3'), covar=tensor([0.0753, 0.1155, 0.1092, 0.1452, 0.0857, 0.0659, 0.1018, 0.0261], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 22:03:08,446 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7394, 2.1046, 1.6578, 1.4344, 1.2758, 1.3011, 1.6442, 1.2152], + device='cuda:3'), covar=tensor([0.1698, 0.1254, 0.1447, 0.1698, 0.2284, 0.1972, 0.1074, 0.2015], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0210, 0.0168, 0.0204, 0.0199, 0.0186, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 22:03:38,678 INFO [finetune.py:976] (3/7) Epoch 25, batch 2300, loss[loss=0.1371, simple_loss=0.2111, pruned_loss=0.03156, over 4842.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2458, pruned_loss=0.04924, over 955611.99 frames. ], batch size: 49, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:03:41,540 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.526e+02 1.763e+02 2.328e+02 5.368e+02, threshold=3.526e+02, percent-clipped=6.0 +2023-04-27 22:03:41,659 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139769.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:03:54,742 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139789.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:04:03,987 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139804.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:04:05,247 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139806.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:04:23,019 INFO [finetune.py:976] (3/7) Epoch 25, batch 2350, loss[loss=0.2193, simple_loss=0.2745, pruned_loss=0.08208, over 4891.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.243, pruned_loss=0.04809, over 956947.85 frames. ], batch size: 35, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:04:44,834 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139830.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:06,894 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:06,903 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:09,120 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 22:05:17,201 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1898, 2.5636, 2.1205, 2.4681, 1.8312, 2.1150, 2.2560, 1.6275], + device='cuda:3'), covar=tensor([0.1826, 0.1236, 0.0715, 0.1042, 0.3184, 0.1165, 0.1888, 0.2731], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0300, 0.0213, 0.0277, 0.0311, 0.0255, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1344e-04, 1.1850e-04, 8.4008e-05, 1.0923e-04, 1.2556e-04, 1.0058e-04, + 1.0005e-04, 1.0379e-04], device='cuda:3') +2023-04-27 22:05:17,688 INFO [finetune.py:976] (3/7) Epoch 25, batch 2400, loss[loss=0.1539, simple_loss=0.2167, pruned_loss=0.04554, over 4929.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2402, pruned_loss=0.04757, over 958643.42 frames. ], batch size: 38, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:05:17,797 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:19,539 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139867.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:20,604 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.574e+01 1.558e+02 1.905e+02 2.229e+02 5.557e+02, threshold=3.809e+02, percent-clipped=2.0 +2023-04-27 22:05:23,147 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5939, 1.5570, 1.7450, 2.0316, 2.0693, 1.5344, 1.2872, 1.8583], + device='cuda:3'), covar=tensor([0.0923, 0.1195, 0.0809, 0.0630, 0.0623, 0.0922, 0.0809, 0.0585], + device='cuda:3'), in_proj_covar=tensor([0.0188, 0.0206, 0.0189, 0.0175, 0.0181, 0.0180, 0.0153, 0.0181], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:05:37,261 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139894.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:53,351 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139911.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:55,629 INFO [finetune.py:976] (3/7) Epoch 25, batch 2450, loss[loss=0.1985, simple_loss=0.2672, pruned_loss=0.06494, over 4809.00 frames. ], tot_loss[loss=0.165, simple_loss=0.237, pruned_loss=0.04651, over 955143.03 frames. ], batch size: 39, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:06:05,158 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 22:06:15,191 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139927.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:18,757 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139932.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 22:06:27,880 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139938.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:35,934 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139942.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:37,755 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:59,672 INFO [finetune.py:976] (3/7) Epoch 25, batch 2500, loss[loss=0.1449, simple_loss=0.2195, pruned_loss=0.03514, over 4783.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2376, pruned_loss=0.04689, over 953782.15 frames. ], batch size: 29, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:07:01,352 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 22:07:04,982 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.582e+02 1.792e+02 2.204e+02 3.647e+02, threshold=3.585e+02, percent-clipped=0.0 +2023-04-27 22:07:11,615 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:07:21,152 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139993.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:07:35,748 INFO [finetune.py:976] (3/7) Epoch 25, batch 2550, loss[loss=0.183, simple_loss=0.2461, pruned_loss=0.05992, over 4909.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2423, pruned_loss=0.04842, over 954078.39 frames. ], batch size: 28, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:07:49,704 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8814, 2.2265, 1.1031, 1.5605, 2.2374, 1.6367, 1.5702, 1.7197], + device='cuda:3'), covar=tensor([0.0446, 0.0306, 0.0276, 0.0507, 0.0235, 0.0481, 0.0476, 0.0506], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 22:08:09,506 INFO [finetune.py:976] (3/7) Epoch 25, batch 2600, loss[loss=0.1721, simple_loss=0.2474, pruned_loss=0.04842, over 4827.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2453, pruned_loss=0.04987, over 953572.20 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:08:09,632 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7727, 1.3817, 1.4147, 1.6141, 2.0134, 1.6170, 1.4045, 1.3684], + device='cuda:3'), covar=tensor([0.1780, 0.1592, 0.1864, 0.1199, 0.0899, 0.1594, 0.2039, 0.2430], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0309, 0.0351, 0.0286, 0.0329, 0.0306, 0.0299, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.3815e-05, 6.3671e-05, 7.3925e-05, 5.7355e-05, 6.7640e-05, 6.4118e-05, + 6.1933e-05, 7.9229e-05], device='cuda:3') +2023-04-27 22:08:12,526 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.514e+02 1.836e+02 2.215e+02 4.899e+02, threshold=3.672e+02, percent-clipped=3.0 +2023-04-27 22:08:13,447 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 22:08:33,240 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9881, 1.7840, 2.2201, 2.3217, 2.0011, 1.9066, 2.0528, 1.9899], + device='cuda:3'), covar=tensor([0.4964, 0.7498, 0.7128, 0.5948, 0.6152, 0.8844, 0.9220, 1.0833], + device='cuda:3'), in_proj_covar=tensor([0.0438, 0.0421, 0.0512, 0.0509, 0.0466, 0.0499, 0.0504, 0.0518], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:08:35,645 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140103.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:08:39,910 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:08:42,875 INFO [finetune.py:976] (3/7) Epoch 25, batch 2650, loss[loss=0.1494, simple_loss=0.2373, pruned_loss=0.03071, over 4807.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2462, pruned_loss=0.05005, over 953498.16 frames. ], batch size: 40, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:08:49,897 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140125.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:04,026 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140145.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:13,157 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140160.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:14,366 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140162.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:15,620 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140164.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:16,092 INFO [finetune.py:976] (3/7) Epoch 25, batch 2700, loss[loss=0.1764, simple_loss=0.2369, pruned_loss=0.05801, over 4799.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2447, pruned_loss=0.04906, over 954038.59 frames. ], batch size: 51, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:09:19,142 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.968e+01 1.589e+02 1.838e+02 2.247e+02 3.608e+02, threshold=3.675e+02, percent-clipped=0.0 +2023-04-27 22:09:19,892 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:55,086 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:10:00,542 INFO [finetune.py:976] (3/7) Epoch 25, batch 2750, loss[loss=0.1539, simple_loss=0.2262, pruned_loss=0.04079, over 4890.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2422, pruned_loss=0.04849, over 954207.19 frames. ], batch size: 35, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:10:18,554 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140227.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:10:32,156 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140238.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:11:02,663 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5267, 3.0662, 2.6116, 2.9026, 2.2180, 2.7063, 2.7675, 2.1608], + device='cuda:3'), covar=tensor([0.1971, 0.1078, 0.0663, 0.1207, 0.2962, 0.0914, 0.1887, 0.2351], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0299, 0.0212, 0.0275, 0.0310, 0.0253, 0.0247, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1324e-04, 1.1813e-04, 8.3532e-05, 1.0851e-04, 1.2495e-04, 9.9874e-05, + 9.9577e-05, 1.0353e-04], device='cuda:3') +2023-04-27 22:11:03,264 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140259.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:11:12,261 INFO [finetune.py:976] (3/7) Epoch 25, batch 2800, loss[loss=0.1582, simple_loss=0.2396, pruned_loss=0.0384, over 4826.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2375, pruned_loss=0.04673, over 952604.97 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:11:15,316 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.465e+02 1.747e+02 2.235e+02 4.062e+02, threshold=3.495e+02, percent-clipped=1.0 +2023-04-27 22:11:20,366 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140275.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:11:21,687 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4461, 1.3640, 1.6998, 1.7226, 1.3563, 1.2507, 1.3454, 0.8389], + device='cuda:3'), covar=tensor([0.0602, 0.0701, 0.0363, 0.0556, 0.0834, 0.1233, 0.0662, 0.0624], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0074, 0.0095, 0.0073, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 22:11:32,990 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140286.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:12:14,089 INFO [finetune.py:976] (3/7) Epoch 25, batch 2850, loss[loss=0.1834, simple_loss=0.2635, pruned_loss=0.05163, over 4809.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2356, pruned_loss=0.04667, over 950172.92 frames. ], batch size: 41, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:12:17,263 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:12:48,659 INFO [finetune.py:976] (3/7) Epoch 25, batch 2900, loss[loss=0.1654, simple_loss=0.2312, pruned_loss=0.04978, over 4831.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2384, pruned_loss=0.047, over 950731.84 frames. ], batch size: 30, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:12:51,723 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.618e+02 1.938e+02 2.286e+02 3.478e+02, threshold=3.877e+02, percent-clipped=0.0 +2023-04-27 22:13:00,442 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5273, 1.8514, 1.9283, 2.0296, 1.8561, 1.9006, 1.9790, 1.8940], + device='cuda:3'), covar=tensor([0.3594, 0.5038, 0.4132, 0.4158, 0.5439, 0.6993, 0.4984, 0.4996], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0339, 0.0349, 0.0393, 0.0358, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:13:02,720 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140387.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:20,921 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6438, 2.0816, 1.6986, 1.4990, 1.2629, 1.2585, 1.6931, 1.2000], + device='cuda:3'), covar=tensor([0.1797, 0.1358, 0.1412, 0.1714, 0.2372, 0.1996, 0.1031, 0.2086], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0211, 0.0169, 0.0204, 0.0200, 0.0187, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 22:13:22,623 INFO [finetune.py:976] (3/7) Epoch 25, batch 2950, loss[loss=0.2077, simple_loss=0.2717, pruned_loss=0.07186, over 4906.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.242, pruned_loss=0.04806, over 949751.31 frames. ], batch size: 43, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:13:28,760 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:42,311 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140445.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:44,097 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140448.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:52,224 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:52,862 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140460.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:54,550 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:56,276 INFO [finetune.py:976] (3/7) Epoch 25, batch 3000, loss[loss=0.1334, simple_loss=0.2052, pruned_loss=0.03077, over 4765.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2432, pruned_loss=0.04846, over 950399.87 frames. ], batch size: 26, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:13:56,276 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 22:13:59,151 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4407, 1.3416, 3.8930, 3.6252, 3.5239, 3.7662, 3.8498, 3.4952], + device='cuda:3'), covar=tensor([0.6863, 0.5112, 0.1221, 0.2049, 0.1332, 0.1347, 0.0729, 0.1765], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0308, 0.0409, 0.0412, 0.0349, 0.0414, 0.0319, 0.0370], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:14:00,980 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2762, 1.6958, 1.4669, 1.9309, 1.7927, 1.7741, 1.5450, 3.1563], + device='cuda:3'), covar=tensor([0.0602, 0.0774, 0.0777, 0.1137, 0.0592, 0.0431, 0.0707, 0.0187], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 22:14:07,209 INFO [finetune.py:1010] (3/7) Epoch 25, validation: loss=0.1531, simple_loss=0.2225, pruned_loss=0.04184, over 2265189.00 frames. +2023-04-27 22:14:07,209 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 22:14:07,885 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:10,185 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.175e+02 1.715e+02 2.080e+02 2.596e+02 3.715e+02, threshold=4.161e+02, percent-clipped=0.0 +2023-04-27 22:14:12,097 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140473.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:21,880 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140489.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:24,277 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140493.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:27,587 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 22:14:33,155 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140506.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:34,804 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140508.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:35,486 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140509.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:36,042 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140510.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:39,579 INFO [finetune.py:976] (3/7) Epoch 25, batch 3050, loss[loss=0.138, simple_loss=0.2124, pruned_loss=0.03183, over 4732.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2447, pruned_loss=0.04889, over 951286.97 frames. ], batch size: 27, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:15:02,555 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140550.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:15:05,405 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140554.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:15:12,554 INFO [finetune.py:976] (3/7) Epoch 25, batch 3100, loss[loss=0.1895, simple_loss=0.2494, pruned_loss=0.06483, over 4923.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2439, pruned_loss=0.04887, over 954640.76 frames. ], batch size: 33, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:15:16,048 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.508e+02 1.726e+02 2.242e+02 3.598e+02, threshold=3.452e+02, percent-clipped=0.0 +2023-04-27 22:15:16,194 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140570.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:15:21,480 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6874, 1.3388, 4.1101, 3.8371, 3.5736, 3.7949, 3.7424, 3.6187], + device='cuda:3'), covar=tensor([0.6702, 0.5640, 0.1027, 0.1698, 0.1071, 0.1753, 0.2517, 0.1450], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0308, 0.0408, 0.0411, 0.0348, 0.0413, 0.0319, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:16:07,368 INFO [finetune.py:976] (3/7) Epoch 25, batch 3150, loss[loss=0.1403, simple_loss=0.209, pruned_loss=0.03575, over 4825.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.241, pruned_loss=0.04803, over 954068.23 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:16:07,435 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140615.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:16:19,069 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0827, 2.4003, 2.1402, 2.3573, 1.7184, 2.0778, 2.0908, 1.7138], + device='cuda:3'), covar=tensor([0.1849, 0.1174, 0.0730, 0.1123, 0.3258, 0.1067, 0.2060, 0.2365], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0301, 0.0213, 0.0277, 0.0311, 0.0255, 0.0249, 0.0264], + device='cuda:3'), out_proj_covar=tensor([1.1390e-04, 1.1894e-04, 8.3926e-05, 1.0907e-04, 1.2545e-04, 1.0054e-04, + 1.0041e-04, 1.0440e-04], device='cuda:3') +2023-04-27 22:17:13,348 INFO [finetune.py:976] (3/7) Epoch 25, batch 3200, loss[loss=0.1553, simple_loss=0.2244, pruned_loss=0.04305, over 4771.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2384, pruned_loss=0.04744, over 955990.87 frames. ], batch size: 26, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:17:21,871 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.528e+02 1.799e+02 2.210e+02 5.000e+02, threshold=3.599e+02, percent-clipped=4.0 +2023-04-27 22:17:34,515 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7642, 1.2342, 1.8184, 2.1969, 1.8363, 1.7007, 1.7833, 1.6894], + device='cuda:3'), covar=tensor([0.4464, 0.6875, 0.6201, 0.5806, 0.5833, 0.8268, 0.7972, 0.9241], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0424, 0.0517, 0.0513, 0.0470, 0.0504, 0.0508, 0.0521], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:18:19,768 INFO [finetune.py:976] (3/7) Epoch 25, batch 3250, loss[loss=0.1639, simple_loss=0.2393, pruned_loss=0.04429, over 4833.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2383, pruned_loss=0.04748, over 951420.12 frames. ], batch size: 30, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:18:26,958 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140723.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:18:55,177 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-27 22:18:55,996 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140743.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:19:16,152 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140759.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:19:20,223 INFO [finetune.py:976] (3/7) Epoch 25, batch 3300, loss[loss=0.1878, simple_loss=0.2631, pruned_loss=0.05627, over 4871.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2432, pruned_loss=0.04909, over 953544.28 frames. ], batch size: 34, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:19:21,453 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:19:27,805 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.701e+02 1.963e+02 2.275e+02 6.006e+02, threshold=3.926e+02, percent-clipped=4.0 +2023-04-27 22:19:29,804 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1389, 1.5803, 2.0692, 2.5835, 2.0682, 1.6098, 1.4243, 1.9116], + device='cuda:3'), covar=tensor([0.3683, 0.3412, 0.1762, 0.2350, 0.2690, 0.2735, 0.4077, 0.1955], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0244, 0.0225, 0.0311, 0.0220, 0.0232, 0.0225, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 22:19:48,001 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:19,458 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140807.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:20,735 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140809.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:20,763 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8863, 1.9150, 2.3525, 2.5278, 1.7669, 1.5328, 1.9291, 1.2829], + device='cuda:3'), covar=tensor([0.0681, 0.0593, 0.0396, 0.0595, 0.0759, 0.1180, 0.0677, 0.0634], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0067, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 22:20:23,777 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140814.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:24,323 INFO [finetune.py:976] (3/7) Epoch 25, batch 3350, loss[loss=0.218, simple_loss=0.2887, pruned_loss=0.07366, over 4821.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2455, pruned_loss=0.05003, over 953861.28 frames. ], batch size: 49, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:20:46,744 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140845.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:56,104 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4046, 1.6425, 1.8255, 1.9281, 1.8094, 1.8482, 1.8317, 1.8430], + device='cuda:3'), covar=tensor([0.3882, 0.5799, 0.4542, 0.4219, 0.5557, 0.7157, 0.5138, 0.5178], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0376, 0.0328, 0.0340, 0.0350, 0.0392, 0.0359, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:20:59,011 INFO [finetune.py:976] (3/7) Epoch 25, batch 3400, loss[loss=0.2162, simple_loss=0.2829, pruned_loss=0.07476, over 4796.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2466, pruned_loss=0.05012, over 954546.80 frames. ], batch size: 45, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:20:59,081 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:02,010 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.526e+02 1.810e+02 2.121e+02 3.949e+02, threshold=3.620e+02, percent-clipped=1.0 +2023-04-27 22:21:02,139 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140870.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:10,342 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6139, 1.9160, 1.8557, 2.1813, 2.1440, 2.3446, 1.8553, 4.7143], + device='cuda:3'), covar=tensor([0.0523, 0.0777, 0.0722, 0.1128, 0.0603, 0.0461, 0.0671, 0.0104], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 22:21:32,399 INFO [finetune.py:976] (3/7) Epoch 25, batch 3450, loss[loss=0.1551, simple_loss=0.2364, pruned_loss=0.03687, over 4883.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2454, pruned_loss=0.04952, over 955388.30 frames. ], batch size: 43, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:21:32,501 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140915.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:33,722 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140917.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:55,372 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140947.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:59,616 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1251, 1.4312, 1.3305, 1.7899, 1.5591, 1.7189, 1.4184, 3.1430], + device='cuda:3'), covar=tensor([0.0689, 0.0967, 0.0932, 0.1266, 0.0766, 0.0523, 0.0845, 0.0231], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 22:22:03,299 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0189, 1.5158, 4.4049, 4.1402, 3.8087, 4.1165, 3.8785, 3.8928], + device='cuda:3'), covar=tensor([0.6553, 0.5521, 0.1063, 0.1544, 0.0991, 0.1586, 0.2946, 0.1476], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0309, 0.0409, 0.0411, 0.0349, 0.0414, 0.0320, 0.0370], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:22:05,109 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140963.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:22:05,753 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0361, 1.3790, 1.6597, 2.4281, 2.4522, 1.8387, 1.5375, 2.0833], + device='cuda:3'), covar=tensor([0.0845, 0.1690, 0.1089, 0.0566, 0.0637, 0.0908, 0.0858, 0.0638], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0186, 0.0171, 0.0179, 0.0177, 0.0150, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:22:06,247 INFO [finetune.py:976] (3/7) Epoch 25, batch 3500, loss[loss=0.1662, simple_loss=0.2347, pruned_loss=0.04884, over 4822.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2418, pruned_loss=0.04796, over 956218.03 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:22:09,318 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.472e+02 1.838e+02 2.111e+02 1.137e+03, threshold=3.676e+02, percent-clipped=2.0 +2023-04-27 22:22:14,764 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140978.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:22:17,655 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6379, 2.6703, 2.2204, 2.4230, 2.7173, 2.3903, 3.5682, 2.1044], + device='cuda:3'), covar=tensor([0.4019, 0.2580, 0.4466, 0.3845, 0.1826, 0.2709, 0.1619, 0.4247], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0353, 0.0428, 0.0351, 0.0381, 0.0376, 0.0370, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:22:35,397 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141008.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:22:39,554 INFO [finetune.py:976] (3/7) Epoch 25, batch 3550, loss[loss=0.1477, simple_loss=0.2313, pruned_loss=0.03202, over 4912.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.239, pruned_loss=0.04688, over 956731.57 frames. ], batch size: 32, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:22:58,188 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141043.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:23:13,445 INFO [finetune.py:976] (3/7) Epoch 25, batch 3600, loss[loss=0.1371, simple_loss=0.1994, pruned_loss=0.03742, over 3983.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2374, pruned_loss=0.04667, over 956364.58 frames. ], batch size: 17, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:23:16,474 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.519e+02 1.759e+02 2.110e+02 6.340e+02, threshold=3.519e+02, percent-clipped=2.0 +2023-04-27 22:23:32,934 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141079.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:23:46,971 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141091.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:24:08,384 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8069, 1.1535, 3.2433, 2.9974, 2.9023, 3.1540, 3.1029, 2.8423], + device='cuda:3'), covar=tensor([0.7732, 0.5769, 0.1652, 0.2434, 0.1477, 0.1861, 0.2152, 0.1836], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0310, 0.0410, 0.0412, 0.0351, 0.0415, 0.0320, 0.0371], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:24:19,423 INFO [finetune.py:976] (3/7) Epoch 25, batch 3650, loss[loss=0.2032, simple_loss=0.2978, pruned_loss=0.05426, over 4819.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2398, pruned_loss=0.04778, over 953623.00 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:24:21,410 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2838, 1.5684, 1.7675, 1.8541, 1.7152, 1.7869, 1.7975, 1.7614], + device='cuda:3'), covar=tensor([0.3672, 0.4923, 0.3974, 0.3961, 0.5030, 0.6574, 0.4749, 0.4397], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0375, 0.0328, 0.0341, 0.0349, 0.0393, 0.0359, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:24:42,735 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141143.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:24:43,923 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141145.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:24:54,460 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6304, 1.5099, 1.7072, 2.0354, 2.0394, 1.6702, 1.2813, 1.8111], + device='cuda:3'), covar=tensor([0.0883, 0.1291, 0.0849, 0.0556, 0.0695, 0.0861, 0.0810, 0.0606], + device='cuda:3'), in_proj_covar=tensor([0.0182, 0.0200, 0.0185, 0.0170, 0.0177, 0.0176, 0.0149, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:24:57,943 INFO [finetune.py:976] (3/7) Epoch 25, batch 3700, loss[loss=0.1786, simple_loss=0.2448, pruned_loss=0.05622, over 4818.00 frames. ], tot_loss[loss=0.17, simple_loss=0.243, pruned_loss=0.04852, over 953150.09 frames. ], batch size: 25, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:24:58,015 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:24:58,034 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:25:00,977 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.179e+02 1.634e+02 1.921e+02 2.262e+02 4.366e+02, threshold=3.843e+02, percent-clipped=2.0 +2023-04-27 22:25:15,813 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141193.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:25:22,645 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141204.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:25:30,032 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141213.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:25:31,152 INFO [finetune.py:976] (3/7) Epoch 25, batch 3750, loss[loss=0.1782, simple_loss=0.2541, pruned_loss=0.0512, over 4775.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2443, pruned_loss=0.04885, over 953422.23 frames. ], batch size: 28, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:26:00,408 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 22:26:23,951 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0942, 1.4296, 1.9294, 2.3543, 2.0835, 1.5862, 1.4292, 1.8242], + device='cuda:3'), covar=tensor([0.3531, 0.4029, 0.2041, 0.2830, 0.2896, 0.2872, 0.4674, 0.2270], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0246, 0.0228, 0.0312, 0.0220, 0.0234, 0.0227, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 22:26:35,278 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0939, 1.7469, 1.9090, 2.4067, 2.4200, 2.0314, 1.4752, 2.1871], + device='cuda:3'), covar=tensor([0.0713, 0.1192, 0.0865, 0.0497, 0.0544, 0.0728, 0.0786, 0.0530], + device='cuda:3'), in_proj_covar=tensor([0.0182, 0.0200, 0.0184, 0.0170, 0.0176, 0.0175, 0.0148, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:26:37,437 INFO [finetune.py:976] (3/7) Epoch 25, batch 3800, loss[loss=0.1867, simple_loss=0.2615, pruned_loss=0.05596, over 4842.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2439, pruned_loss=0.04855, over 952703.44 frames. ], batch size: 49, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:26:45,874 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.561e+02 1.876e+02 2.402e+02 4.277e+02, threshold=3.753e+02, percent-clipped=1.0 +2023-04-27 22:26:47,827 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:27:28,694 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141303.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:27:38,904 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 22:27:42,126 INFO [finetune.py:976] (3/7) Epoch 25, batch 3850, loss[loss=0.2032, simple_loss=0.2666, pruned_loss=0.06987, over 4261.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2426, pruned_loss=0.04824, over 952335.39 frames. ], batch size: 65, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:28:00,745 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 22:28:02,218 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:28:10,370 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3157, 1.8547, 2.2066, 2.6039, 2.2678, 1.7698, 1.4397, 2.0389], + device='cuda:3'), covar=tensor([0.3340, 0.3259, 0.1848, 0.1934, 0.2502, 0.2744, 0.4034, 0.1861], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0245, 0.0227, 0.0312, 0.0220, 0.0233, 0.0226, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 22:28:11,264 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-04-27 22:28:27,561 INFO [finetune.py:976] (3/7) Epoch 25, batch 3900, loss[loss=0.1267, simple_loss=0.2087, pruned_loss=0.0224, over 4835.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2396, pruned_loss=0.04689, over 952398.91 frames. ], batch size: 47, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:28:27,638 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4525, 3.3682, 2.5077, 3.9606, 3.4404, 3.4674, 1.4541, 3.3843], + device='cuda:3'), covar=tensor([0.1802, 0.1449, 0.2838, 0.2075, 0.3894, 0.1849, 0.5632, 0.2596], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0220, 0.0254, 0.0307, 0.0300, 0.0251, 0.0275, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:28:31,461 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.108e+01 1.538e+02 1.758e+02 2.132e+02 4.697e+02, threshold=3.516e+02, percent-clipped=1.0 +2023-04-27 22:28:37,517 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141379.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:28:44,169 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:28:47,196 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7412, 1.6315, 0.7964, 1.4198, 1.7191, 1.5687, 1.5052, 1.5615], + device='cuda:3'), covar=tensor([0.0446, 0.0352, 0.0335, 0.0513, 0.0282, 0.0481, 0.0465, 0.0519], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 22:29:00,436 INFO [finetune.py:976] (3/7) Epoch 25, batch 3950, loss[loss=0.1545, simple_loss=0.2111, pruned_loss=0.04897, over 4135.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.237, pruned_loss=0.04622, over 953395.29 frames. ], batch size: 17, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:29:01,812 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141417.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 22:29:09,298 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141427.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:17,826 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141441.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:33,822 INFO [finetune.py:976] (3/7) Epoch 25, batch 4000, loss[loss=0.1641, simple_loss=0.232, pruned_loss=0.04817, over 4738.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2368, pruned_loss=0.0463, over 953145.91 frames. ], batch size: 27, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:29:33,904 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141465.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:36,894 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.301e+01 1.493e+02 1.709e+02 2.038e+02 4.561e+02, threshold=3.417e+02, percent-clipped=1.0 +2023-04-27 22:29:43,322 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141478.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 22:29:55,961 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141499.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:57,868 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141502.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:30:05,459 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141513.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:30:06,624 INFO [finetune.py:976] (3/7) Epoch 25, batch 4050, loss[loss=0.1715, simple_loss=0.2471, pruned_loss=0.04796, over 4776.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2398, pruned_loss=0.04721, over 952119.73 frames. ], batch size: 28, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:30:25,371 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2154, 1.7081, 2.1353, 2.4890, 2.1158, 1.6573, 1.3738, 1.8584], + device='cuda:3'), covar=tensor([0.3489, 0.3124, 0.1658, 0.2093, 0.2516, 0.2736, 0.3969, 0.1956], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0243, 0.0226, 0.0311, 0.0219, 0.0232, 0.0225, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 22:30:39,499 INFO [finetune.py:976] (3/7) Epoch 25, batch 4100, loss[loss=0.1501, simple_loss=0.2283, pruned_loss=0.03594, over 4755.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2417, pruned_loss=0.04736, over 954303.18 frames. ], batch size: 54, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:30:40,811 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5288, 1.8323, 1.4198, 1.2137, 1.1627, 1.1744, 1.4598, 1.1354], + device='cuda:3'), covar=tensor([0.1616, 0.1162, 0.1364, 0.1587, 0.2235, 0.1877, 0.0946, 0.2015], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0209, 0.0167, 0.0203, 0.0198, 0.0185, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 22:30:42,488 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.442e+01 1.586e+02 1.949e+02 2.336e+02 4.544e+02, threshold=3.898e+02, percent-clipped=7.0 +2023-04-27 22:30:44,326 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141573.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:30:46,253 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8065, 1.3914, 1.9265, 2.3000, 1.9648, 1.8172, 1.9310, 1.8084], + device='cuda:3'), covar=tensor([0.4566, 0.7219, 0.6148, 0.5190, 0.5671, 0.7746, 0.7750, 0.9525], + device='cuda:3'), in_proj_covar=tensor([0.0438, 0.0421, 0.0512, 0.0507, 0.0466, 0.0500, 0.0503, 0.0516], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:31:04,700 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141603.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:31:17,880 INFO [finetune.py:976] (3/7) Epoch 25, batch 4150, loss[loss=0.1435, simple_loss=0.224, pruned_loss=0.0315, over 4920.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2433, pruned_loss=0.0478, over 953517.77 frames. ], batch size: 28, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:31:21,627 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141621.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:31:53,143 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 22:32:02,964 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141651.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:32:13,575 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 22:32:23,210 INFO [finetune.py:976] (3/7) Epoch 25, batch 4200, loss[loss=0.1842, simple_loss=0.2565, pruned_loss=0.05591, over 4787.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2444, pruned_loss=0.04842, over 954510.42 frames. ], batch size: 51, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:32:26,264 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.638e+02 1.894e+02 2.201e+02 5.051e+02, threshold=3.789e+02, percent-clipped=1.0 +2023-04-27 22:32:47,766 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:32:56,484 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9829, 1.9072, 1.8096, 1.5534, 2.0393, 1.7504, 2.5638, 1.7074], + device='cuda:3'), covar=tensor([0.3779, 0.2285, 0.5002, 0.3237, 0.1930, 0.2496, 0.1447, 0.4469], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0353, 0.0428, 0.0353, 0.0383, 0.0376, 0.0371, 0.0425], + device='cuda:3'), out_proj_covar=tensor([9.9912e-05, 1.0531e-04, 1.2976e-04, 1.0598e-04, 1.1378e-04, 1.1198e-04, + 1.0878e-04, 1.2790e-04], device='cuda:3') +2023-04-27 22:33:28,597 INFO [finetune.py:976] (3/7) Epoch 25, batch 4250, loss[loss=0.1428, simple_loss=0.2107, pruned_loss=0.03745, over 4768.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2431, pruned_loss=0.04842, over 955417.28 frames. ], batch size: 26, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:34:30,331 INFO [finetune.py:976] (3/7) Epoch 25, batch 4300, loss[loss=0.1602, simple_loss=0.227, pruned_loss=0.04665, over 4919.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.24, pruned_loss=0.04756, over 953722.28 frames. ], batch size: 37, lr: 2.99e-03, grad_scale: 16.0 +2023-04-27 22:34:39,430 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.343e+01 1.507e+02 1.747e+02 2.193e+02 4.425e+02, threshold=3.494e+02, percent-clipped=2.0 +2023-04-27 22:34:41,376 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141773.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 22:34:55,279 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-27 22:35:15,326 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:35:16,548 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141799.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:35:36,800 INFO [finetune.py:976] (3/7) Epoch 25, batch 4350, loss[loss=0.1915, simple_loss=0.2639, pruned_loss=0.05955, over 4861.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2379, pruned_loss=0.04725, over 954947.19 frames. ], batch size: 44, lr: 2.99e-03, grad_scale: 16.0 +2023-04-27 22:36:20,236 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141847.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:36:42,216 INFO [finetune.py:976] (3/7) Epoch 25, batch 4400, loss[loss=0.1357, simple_loss=0.217, pruned_loss=0.02724, over 4904.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2388, pruned_loss=0.04796, over 954324.53 frames. ], batch size: 35, lr: 2.99e-03, grad_scale: 16.0 +2023-04-27 22:36:50,566 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.541e+02 1.847e+02 2.232e+02 3.824e+02, threshold=3.693e+02, percent-clipped=5.0 +2023-04-27 22:36:54,752 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2495, 1.6982, 2.1496, 2.7383, 2.2202, 1.6701, 1.5844, 2.0210], + device='cuda:3'), covar=tensor([0.3052, 0.3201, 0.1615, 0.1890, 0.2403, 0.2616, 0.3943, 0.1793], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0246, 0.0229, 0.0315, 0.0222, 0.0236, 0.0228, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 22:37:21,472 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0665, 1.8087, 2.0889, 2.5064, 2.4731, 2.0439, 1.7424, 2.2409], + device='cuda:3'), covar=tensor([0.0866, 0.1216, 0.0749, 0.0565, 0.0624, 0.0837, 0.0737, 0.0547], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0185, 0.0172, 0.0178, 0.0177, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:37:46,056 INFO [finetune.py:976] (3/7) Epoch 25, batch 4450, loss[loss=0.1929, simple_loss=0.2633, pruned_loss=0.06129, over 4872.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2426, pruned_loss=0.04882, over 956971.17 frames. ], batch size: 34, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:37:46,786 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:38:04,877 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141928.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:38:51,125 INFO [finetune.py:976] (3/7) Epoch 25, batch 4500, loss[loss=0.1543, simple_loss=0.2309, pruned_loss=0.03885, over 4885.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2429, pruned_loss=0.0489, over 953820.82 frames. ], batch size: 35, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:38:59,405 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.537e+02 1.854e+02 2.228e+02 3.851e+02, threshold=3.709e+02, percent-clipped=1.0 +2023-04-27 22:39:09,402 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:39:20,817 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8640, 2.8036, 2.1344, 3.2650, 2.8615, 2.8215, 1.0903, 2.8381], + device='cuda:3'), covar=tensor([0.2145, 0.1887, 0.3880, 0.3296, 0.2861, 0.2572, 0.6059, 0.2783], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0220, 0.0253, 0.0307, 0.0301, 0.0251, 0.0276, 0.0275], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:39:21,486 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141985.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:39:23,963 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:39:57,191 INFO [finetune.py:976] (3/7) Epoch 25, batch 4550, loss[loss=0.1829, simple_loss=0.2487, pruned_loss=0.05858, over 4924.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2442, pruned_loss=0.04916, over 953259.07 frames. ], batch size: 42, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:40:19,390 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142033.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:41:01,793 INFO [finetune.py:976] (3/7) Epoch 25, batch 4600, loss[loss=0.1418, simple_loss=0.218, pruned_loss=0.03278, over 4780.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2431, pruned_loss=0.04819, over 954868.80 frames. ], batch size: 26, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:41:10,130 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.365e+01 1.636e+02 1.871e+02 2.331e+02 4.472e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-04-27 22:41:11,068 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-27 22:41:12,042 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142073.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 22:41:43,194 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:42:06,624 INFO [finetune.py:976] (3/7) Epoch 25, batch 4650, loss[loss=0.2108, simple_loss=0.2743, pruned_loss=0.07363, over 4824.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2409, pruned_loss=0.0482, over 952869.87 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:42:15,790 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142121.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 22:42:28,649 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6051, 1.4736, 4.4033, 4.1094, 3.7972, 4.1937, 4.0738, 3.8618], + device='cuda:3'), covar=tensor([0.7593, 0.5922, 0.1144, 0.1871, 0.1237, 0.1336, 0.1205, 0.1891], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0308, 0.0408, 0.0411, 0.0350, 0.0414, 0.0319, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:42:47,434 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142145.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:43:11,598 INFO [finetune.py:976] (3/7) Epoch 25, batch 4700, loss[loss=0.1581, simple_loss=0.2306, pruned_loss=0.04281, over 4872.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2389, pruned_loss=0.04785, over 955203.29 frames. ], batch size: 34, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:43:19,815 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.574e+02 1.870e+02 2.251e+02 4.397e+02, threshold=3.741e+02, percent-clipped=1.0 +2023-04-27 22:44:17,216 INFO [finetune.py:976] (3/7) Epoch 25, batch 4750, loss[loss=0.2123, simple_loss=0.2658, pruned_loss=0.07946, over 4925.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2375, pruned_loss=0.04758, over 956075.40 frames. ], batch size: 38, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:44:39,400 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6314, 1.3973, 4.5120, 4.1926, 3.9044, 4.2712, 4.1203, 3.9661], + device='cuda:3'), covar=tensor([0.7292, 0.6036, 0.1033, 0.1957, 0.1212, 0.1691, 0.1400, 0.1691], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0307, 0.0405, 0.0409, 0.0348, 0.0413, 0.0318, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:44:57,373 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5631, 2.4559, 2.8048, 3.1759, 2.9431, 2.4019, 2.1758, 2.8123], + device='cuda:3'), covar=tensor([0.0941, 0.1007, 0.0641, 0.0607, 0.0704, 0.0959, 0.0755, 0.0573], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0201, 0.0185, 0.0171, 0.0177, 0.0177, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:45:20,314 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0522, 1.1060, 1.7487, 1.8530, 1.7524, 1.8696, 1.7567, 1.7799], + device='cuda:3'), covar=tensor([0.3729, 0.5006, 0.4512, 0.4775, 0.5670, 0.6987, 0.4397, 0.4396], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0374, 0.0327, 0.0340, 0.0348, 0.0392, 0.0360, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:45:29,863 INFO [finetune.py:976] (3/7) Epoch 25, batch 4800, loss[loss=0.2023, simple_loss=0.2689, pruned_loss=0.06786, over 4743.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.239, pruned_loss=0.0478, over 956191.50 frames. ], batch size: 54, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:45:34,011 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.626e+02 1.868e+02 2.200e+02 5.189e+02, threshold=3.736e+02, percent-clipped=2.0 +2023-04-27 22:45:40,861 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:45:54,146 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142284.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:46:03,984 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.87 vs. limit=5.0 +2023-04-27 22:46:36,614 INFO [finetune.py:976] (3/7) Epoch 25, batch 4850, loss[loss=0.1652, simple_loss=0.2269, pruned_loss=0.05173, over 4247.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2425, pruned_loss=0.0482, over 954996.59 frames. ], batch size: 18, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:47:42,702 INFO [finetune.py:976] (3/7) Epoch 25, batch 4900, loss[loss=0.1839, simple_loss=0.2657, pruned_loss=0.051, over 4718.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2439, pruned_loss=0.0486, over 954857.12 frames. ], batch size: 59, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:47:51,886 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.944e+01 1.647e+02 1.933e+02 2.315e+02 7.407e+02, threshold=3.866e+02, percent-clipped=3.0 +2023-04-27 22:48:49,020 INFO [finetune.py:976] (3/7) Epoch 25, batch 4950, loss[loss=0.1788, simple_loss=0.2445, pruned_loss=0.05657, over 4281.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2448, pruned_loss=0.04909, over 954967.18 frames. ], batch size: 18, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:49:57,972 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1951, 1.6649, 2.0614, 2.3960, 2.1391, 1.6233, 1.3175, 1.8742], + device='cuda:3'), covar=tensor([0.2871, 0.2948, 0.1566, 0.2080, 0.2284, 0.2529, 0.4043, 0.1814], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0245, 0.0227, 0.0313, 0.0221, 0.0234, 0.0226, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 22:49:58,431 INFO [finetune.py:976] (3/7) Epoch 25, batch 5000, loss[loss=0.189, simple_loss=0.2658, pruned_loss=0.05609, over 4897.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2429, pruned_loss=0.04823, over 953666.54 frames. ], batch size: 43, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:50:01,481 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.520e+02 1.784e+02 2.172e+02 4.625e+02, threshold=3.567e+02, percent-clipped=1.0 +2023-04-27 22:50:13,339 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9560, 1.8639, 2.1454, 2.3977, 2.4445, 1.8568, 1.6436, 2.1480], + device='cuda:3'), covar=tensor([0.0894, 0.1166, 0.0725, 0.0610, 0.0616, 0.0943, 0.0844, 0.0612], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0203, 0.0186, 0.0172, 0.0178, 0.0179, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:50:45,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5547, 1.8196, 1.9433, 1.9956, 1.8688, 1.9734, 2.0011, 1.9773], + device='cuda:3'), covar=tensor([0.3561, 0.4835, 0.4142, 0.4090, 0.5085, 0.6051, 0.4730, 0.4509], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0375, 0.0327, 0.0340, 0.0348, 0.0392, 0.0360, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:51:03,145 INFO [finetune.py:976] (3/7) Epoch 25, batch 5050, loss[loss=0.1352, simple_loss=0.212, pruned_loss=0.02918, over 4825.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2394, pruned_loss=0.04709, over 955082.74 frames. ], batch size: 41, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:51:04,468 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4752, 1.4760, 3.6317, 3.2131, 3.3086, 3.2823, 3.4090, 3.0159], + device='cuda:3'), covar=tensor([0.9279, 0.7809, 0.1981, 0.3777, 0.2111, 0.3908, 0.2691, 0.3504], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0307, 0.0407, 0.0409, 0.0350, 0.0414, 0.0318, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:51:28,032 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.11 vs. limit=5.0 +2023-04-27 22:51:48,460 INFO [finetune.py:976] (3/7) Epoch 25, batch 5100, loss[loss=0.1779, simple_loss=0.2353, pruned_loss=0.06024, over 4833.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2371, pruned_loss=0.04666, over 954446.81 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:51:51,979 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.504e+02 1.862e+02 2.404e+02 6.312e+02, threshold=3.723e+02, percent-clipped=2.0 +2023-04-27 22:51:53,295 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142572.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:51:53,936 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=142573.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:01,655 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142584.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:21,630 INFO [finetune.py:976] (3/7) Epoch 25, batch 5150, loss[loss=0.2499, simple_loss=0.2979, pruned_loss=0.101, over 4822.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2361, pruned_loss=0.04681, over 952980.15 frames. ], batch size: 38, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:52:24,148 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-27 22:52:26,239 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142620.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:33,587 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142632.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:35,334 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142634.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:43,072 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3940, 1.3832, 1.7760, 1.7099, 1.2569, 1.1606, 1.4440, 0.8837], + device='cuda:3'), covar=tensor([0.0498, 0.0584, 0.0341, 0.0563, 0.0738, 0.1070, 0.0577, 0.0585], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0068, 0.0067, 0.0069, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 22:52:56,102 INFO [finetune.py:976] (3/7) Epoch 25, batch 5200, loss[loss=0.1895, simple_loss=0.2523, pruned_loss=0.06329, over 4830.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2396, pruned_loss=0.04734, over 951958.96 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:53:00,173 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.581e+02 1.903e+02 2.348e+02 3.515e+02, threshold=3.805e+02, percent-clipped=0.0 +2023-04-27 22:53:29,661 INFO [finetune.py:976] (3/7) Epoch 25, batch 5250, loss[loss=0.1739, simple_loss=0.2557, pruned_loss=0.04603, over 4813.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2412, pruned_loss=0.04724, over 953602.61 frames. ], batch size: 45, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:53:36,256 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=142724.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:53:42,361 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6069, 1.2678, 0.5677, 1.2982, 1.4650, 1.4621, 1.3709, 1.3397], + device='cuda:3'), covar=tensor([0.0462, 0.0374, 0.0367, 0.0538, 0.0276, 0.0463, 0.0452, 0.0551], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 22:54:19,562 INFO [finetune.py:976] (3/7) Epoch 25, batch 5300, loss[loss=0.1739, simple_loss=0.2492, pruned_loss=0.04935, over 4837.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2434, pruned_loss=0.04832, over 952623.62 frames. ], batch size: 49, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:54:22,608 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.572e+02 1.803e+02 2.243e+02 5.104e+02, threshold=3.606e+02, percent-clipped=2.0 +2023-04-27 22:54:44,815 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142785.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:55:27,672 INFO [finetune.py:976] (3/7) Epoch 25, batch 5350, loss[loss=0.1501, simple_loss=0.2166, pruned_loss=0.04184, over 4887.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2431, pruned_loss=0.04786, over 952174.59 frames. ], batch size: 32, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:55:29,649 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7400, 1.3537, 1.4333, 1.4995, 1.8428, 1.5676, 1.3294, 1.3583], + device='cuda:3'), covar=tensor([0.1396, 0.1390, 0.1802, 0.1304, 0.0933, 0.1370, 0.1749, 0.2263], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0308, 0.0352, 0.0286, 0.0328, 0.0305, 0.0298, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.4081e-05, 6.3245e-05, 7.3943e-05, 5.7294e-05, 6.7268e-05, 6.3837e-05, + 6.1752e-05, 7.9203e-05], device='cuda:3') +2023-04-27 22:56:34,016 INFO [finetune.py:976] (3/7) Epoch 25, batch 5400, loss[loss=0.1694, simple_loss=0.2357, pruned_loss=0.05152, over 4872.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2411, pruned_loss=0.04737, over 954254.95 frames. ], batch size: 34, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:56:42,409 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.482e+02 1.784e+02 2.095e+02 4.679e+02, threshold=3.568e+02, percent-clipped=3.0 +2023-04-27 22:57:18,642 INFO [finetune.py:976] (3/7) Epoch 25, batch 5450, loss[loss=0.1436, simple_loss=0.221, pruned_loss=0.03311, over 4744.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2389, pruned_loss=0.04687, over 955846.28 frames. ], batch size: 26, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:57:27,229 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:57:33,077 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0696, 1.4379, 1.5190, 1.7229, 2.0022, 1.6290, 1.5277, 1.3864], + device='cuda:3'), covar=tensor([0.1376, 0.1714, 0.1906, 0.1484, 0.1037, 0.1885, 0.1866, 0.2489], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0306, 0.0350, 0.0284, 0.0326, 0.0304, 0.0297, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.3780e-05, 6.2885e-05, 7.3465e-05, 5.6896e-05, 6.6921e-05, 6.3552e-05, + 6.1459e-05, 7.8755e-05], device='cuda:3') +2023-04-27 22:57:36,999 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 22:57:48,116 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 22:57:52,210 INFO [finetune.py:976] (3/7) Epoch 25, batch 5500, loss[loss=0.1751, simple_loss=0.2422, pruned_loss=0.05404, over 4901.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2364, pruned_loss=0.04603, over 954370.81 frames. ], batch size: 43, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:57:55,643 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.918e+01 1.380e+02 1.724e+02 2.155e+02 4.005e+02, threshold=3.448e+02, percent-clipped=1.0 +2023-04-27 22:58:03,967 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.05 vs. limit=2.0 +2023-04-27 22:58:26,144 INFO [finetune.py:976] (3/7) Epoch 25, batch 5550, loss[loss=0.1864, simple_loss=0.248, pruned_loss=0.06237, over 4020.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2385, pruned_loss=0.04718, over 953432.49 frames. ], batch size: 17, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:58:37,219 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8514, 1.1854, 3.2403, 2.9876, 2.9185, 3.1942, 3.1648, 2.8754], + device='cuda:3'), covar=tensor([0.7720, 0.5752, 0.1641, 0.2340, 0.1594, 0.2082, 0.1849, 0.1847], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0308, 0.0408, 0.0409, 0.0348, 0.0414, 0.0318, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 22:58:37,522 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.09 vs. limit=2.0 +2023-04-27 22:58:57,671 INFO [finetune.py:976] (3/7) Epoch 25, batch 5600, loss[loss=0.1797, simple_loss=0.2488, pruned_loss=0.05533, over 4819.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.243, pruned_loss=0.04845, over 953263.51 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:58:57,788 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9849, 2.4363, 1.9316, 1.7017, 1.4826, 1.4835, 1.9838, 1.4217], + device='cuda:3'), covar=tensor([0.1699, 0.1186, 0.1390, 0.1723, 0.2204, 0.1909, 0.1007, 0.2044], + device='cuda:3'), in_proj_covar=tensor([0.0196, 0.0208, 0.0167, 0.0202, 0.0198, 0.0184, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 22:58:58,317 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8926, 2.4161, 1.0704, 1.4317, 1.8778, 1.2123, 3.1776, 1.6236], + device='cuda:3'), covar=tensor([0.0744, 0.0618, 0.0741, 0.1236, 0.0530, 0.1013, 0.0208, 0.0616], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 22:59:00,543 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.590e+02 1.837e+02 2.169e+02 3.781e+02, threshold=3.675e+02, percent-clipped=1.0 +2023-04-27 22:59:06,456 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:59:19,189 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2747, 1.4859, 1.3565, 1.7311, 1.6047, 1.7646, 1.3761, 3.4389], + device='cuda:3'), covar=tensor([0.0620, 0.0826, 0.0840, 0.1254, 0.0658, 0.0513, 0.0761, 0.0140], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 22:59:26,510 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7673, 3.7494, 2.7520, 4.3362, 3.7419, 3.7129, 1.4859, 3.6729], + device='cuda:3'), covar=tensor([0.1649, 0.1195, 0.3277, 0.1607, 0.3120, 0.1901, 0.5797, 0.2539], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0218, 0.0250, 0.0304, 0.0297, 0.0247, 0.0272, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 22:59:27,655 INFO [finetune.py:976] (3/7) Epoch 25, batch 5650, loss[loss=0.1667, simple_loss=0.2526, pruned_loss=0.04046, over 4862.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2458, pruned_loss=0.04902, over 953247.15 frames. ], batch size: 44, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 23:00:12,852 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5177, 2.4630, 2.2962, 2.2107, 2.6302, 2.2411, 3.0899, 1.9660], + device='cuda:3'), covar=tensor([0.2889, 0.1664, 0.3114, 0.2148, 0.1274, 0.1926, 0.1461, 0.3283], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0351, 0.0424, 0.0349, 0.0381, 0.0374, 0.0368, 0.0424], + device='cuda:3'), out_proj_covar=tensor([9.9702e-05, 1.0469e-04, 1.2831e-04, 1.0486e-04, 1.1302e-04, 1.1112e-04, + 1.0790e-04, 1.2741e-04], device='cuda:3') +2023-04-27 23:00:24,803 INFO [finetune.py:976] (3/7) Epoch 25, batch 5700, loss[loss=0.1447, simple_loss=0.2039, pruned_loss=0.04278, over 4273.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.241, pruned_loss=0.04826, over 934855.97 frames. ], batch size: 18, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:00:27,756 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.938e+01 1.487e+02 1.759e+02 2.216e+02 4.830e+02, threshold=3.518e+02, percent-clipped=2.0 +2023-04-27 23:01:04,603 INFO [finetune.py:976] (3/7) Epoch 26, batch 0, loss[loss=0.1646, simple_loss=0.2372, pruned_loss=0.04603, over 4758.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2372, pruned_loss=0.04603, over 4758.00 frames. ], batch size: 26, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:01:04,604 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 23:01:10,385 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3038, 1.5691, 1.4584, 1.7522, 1.6754, 1.7399, 1.4442, 3.0445], + device='cuda:3'), covar=tensor([0.0552, 0.0787, 0.0726, 0.1185, 0.0575, 0.0472, 0.0682, 0.0176], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 23:01:11,499 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8802, 1.8032, 1.9817, 2.2806, 2.2645, 1.8226, 1.4290, 2.0624], + device='cuda:3'), covar=tensor([0.0849, 0.1131, 0.0739, 0.0526, 0.0590, 0.0814, 0.0703, 0.0541], + device='cuda:3'), in_proj_covar=tensor([0.0183, 0.0201, 0.0184, 0.0170, 0.0175, 0.0176, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:01:12,609 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7812, 2.0156, 1.9520, 1.6926, 2.0803, 1.6911, 2.4467, 1.7234], + device='cuda:3'), covar=tensor([0.3159, 0.1459, 0.4278, 0.2169, 0.1207, 0.1997, 0.1452, 0.4056], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0351, 0.0423, 0.0349, 0.0380, 0.0373, 0.0368, 0.0423], + device='cuda:3'), out_proj_covar=tensor([9.9582e-05, 1.0454e-04, 1.2819e-04, 1.0477e-04, 1.1284e-04, 1.1102e-04, + 1.0775e-04, 1.2729e-04], device='cuda:3') +2023-04-27 23:01:26,499 INFO [finetune.py:1010] (3/7) Epoch 26, validation: loss=0.1543, simple_loss=0.2237, pruned_loss=0.04251, over 2265189.00 frames. +2023-04-27 23:01:26,499 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 23:01:52,144 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7447, 2.0517, 2.0012, 2.3626, 2.2961, 2.3555, 1.8712, 4.7127], + device='cuda:3'), covar=tensor([0.0469, 0.0747, 0.0711, 0.1020, 0.0527, 0.0461, 0.0648, 0.0116], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 23:02:16,890 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:02:34,931 INFO [finetune.py:976] (3/7) Epoch 26, batch 50, loss[loss=0.1464, simple_loss=0.2221, pruned_loss=0.03534, over 4823.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2436, pruned_loss=0.04748, over 217183.99 frames. ], batch size: 33, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:02:57,495 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5092, 1.1038, 1.3019, 1.2597, 1.5806, 1.2870, 1.0852, 1.2534], + device='cuda:3'), covar=tensor([0.1296, 0.1246, 0.1477, 0.1225, 0.0924, 0.1294, 0.1518, 0.1801], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0305, 0.0348, 0.0284, 0.0325, 0.0302, 0.0295, 0.0371], + device='cuda:3'), out_proj_covar=tensor([6.3631e-05, 6.2701e-05, 7.3136e-05, 5.6802e-05, 6.6669e-05, 6.3230e-05, + 6.1165e-05, 7.8611e-05], device='cuda:3') +2023-04-27 23:03:09,361 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5938, 2.0791, 2.5394, 3.0656, 2.4274, 1.9966, 1.8745, 2.3282], + device='cuda:3'), covar=tensor([0.3018, 0.3002, 0.1471, 0.2137, 0.2486, 0.2567, 0.3582, 0.1916], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0230, 0.0315, 0.0221, 0.0235, 0.0228, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 23:03:09,818 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.418e+01 1.474e+02 1.785e+02 2.280e+02 3.483e+02, threshold=3.571e+02, percent-clipped=0.0 +2023-04-27 23:03:19,711 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143277.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:03:31,447 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5009, 2.1285, 2.4772, 2.8784, 2.4513, 1.9735, 1.7412, 2.3019], + device='cuda:3'), covar=tensor([0.3218, 0.2698, 0.1424, 0.1986, 0.2586, 0.2518, 0.3591, 0.1850], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0315, 0.0221, 0.0235, 0.0228, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 23:03:40,347 INFO [finetune.py:976] (3/7) Epoch 26, batch 100, loss[loss=0.1799, simple_loss=0.246, pruned_loss=0.05686, over 4938.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2371, pruned_loss=0.04567, over 382310.56 frames. ], batch size: 33, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:03:40,425 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6900, 3.5438, 2.7758, 4.2258, 3.6082, 3.7182, 1.5412, 3.6025], + device='cuda:3'), covar=tensor([0.1841, 0.1477, 0.3230, 0.1940, 0.2975, 0.1728, 0.5806, 0.2581], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0217, 0.0251, 0.0304, 0.0297, 0.0247, 0.0272, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:04:19,085 INFO [finetune.py:976] (3/7) Epoch 26, batch 150, loss[loss=0.1518, simple_loss=0.2232, pruned_loss=0.04015, over 4878.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2326, pruned_loss=0.04456, over 510983.00 frames. ], batch size: 34, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:04:25,880 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.65 vs. limit=5.0 +2023-04-27 23:04:37,630 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.476e+02 1.692e+02 2.051e+02 5.029e+02, threshold=3.384e+02, percent-clipped=1.0 +2023-04-27 23:04:43,891 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:04:52,921 INFO [finetune.py:976] (3/7) Epoch 26, batch 200, loss[loss=0.1696, simple_loss=0.2393, pruned_loss=0.04995, over 4776.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2332, pruned_loss=0.04609, over 610183.54 frames. ], batch size: 28, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:04:59,039 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=143400.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:05:16,580 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143428.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:05:16,661 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0668, 1.5300, 1.9531, 2.1972, 1.9054, 1.5285, 1.0462, 1.6542], + device='cuda:3'), covar=tensor([0.3591, 0.3456, 0.2001, 0.2196, 0.2714, 0.3046, 0.4280, 0.2131], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0315, 0.0221, 0.0235, 0.0227, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 23:05:31,680 INFO [finetune.py:976] (3/7) Epoch 26, batch 250, loss[loss=0.1759, simple_loss=0.2531, pruned_loss=0.04938, over 4859.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2365, pruned_loss=0.04651, over 687497.41 frames. ], batch size: 44, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:05:44,908 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143461.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:05:47,384 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1125, 2.0605, 1.6888, 1.8560, 2.2905, 1.8872, 2.7376, 1.6069], + device='cuda:3'), covar=tensor([0.3754, 0.2162, 0.4666, 0.3007, 0.1790, 0.2458, 0.1445, 0.4255], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0351, 0.0423, 0.0348, 0.0380, 0.0374, 0.0367, 0.0422], + device='cuda:3'), out_proj_covar=tensor([9.9711e-05, 1.0461e-04, 1.2805e-04, 1.0466e-04, 1.1269e-04, 1.1139e-04, + 1.0755e-04, 1.2697e-04], device='cuda:3') +2023-04-27 23:05:50,287 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.556e+02 1.824e+02 2.331e+02 6.380e+02, threshold=3.648e+02, percent-clipped=4.0 +2023-04-27 23:05:58,097 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=143473.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:06:15,190 INFO [finetune.py:976] (3/7) Epoch 26, batch 300, loss[loss=0.1826, simple_loss=0.2716, pruned_loss=0.04678, over 4909.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2407, pruned_loss=0.04788, over 746750.91 frames. ], batch size: 42, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:06:30,773 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3574, 3.0444, 2.6022, 2.8724, 2.0773, 2.7279, 2.6868, 2.0728], + device='cuda:3'), covar=tensor([0.2037, 0.1245, 0.0691, 0.1168, 0.3343, 0.0923, 0.1899, 0.2701], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0302, 0.0214, 0.0277, 0.0315, 0.0256, 0.0250, 0.0267], + device='cuda:3'), out_proj_covar=tensor([1.1415e-04, 1.1927e-04, 8.4152e-05, 1.0908e-04, 1.2682e-04, 1.0058e-04, + 1.0063e-04, 1.0539e-04], device='cuda:3') +2023-04-27 23:06:43,067 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143534.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:06:48,239 INFO [finetune.py:976] (3/7) Epoch 26, batch 350, loss[loss=0.155, simple_loss=0.2361, pruned_loss=0.03691, over 4896.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2428, pruned_loss=0.04815, over 793765.97 frames. ], batch size: 36, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:07:08,173 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.499e+02 1.727e+02 2.023e+02 3.986e+02, threshold=3.454e+02, percent-clipped=1.0 +2023-04-27 23:07:15,562 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3630, 3.3477, 2.4722, 3.8472, 3.3865, 3.3071, 1.4542, 3.3258], + device='cuda:3'), covar=tensor([0.1835, 0.1212, 0.3338, 0.2323, 0.2693, 0.1957, 0.5453, 0.2595], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0218, 0.0252, 0.0305, 0.0298, 0.0247, 0.0273, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:07:22,106 INFO [finetune.py:976] (3/7) Epoch 26, batch 400, loss[loss=0.1585, simple_loss=0.2317, pruned_loss=0.04263, over 4863.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2438, pruned_loss=0.04823, over 829226.24 frames. ], batch size: 31, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:07:55,518 INFO [finetune.py:976] (3/7) Epoch 26, batch 450, loss[loss=0.1624, simple_loss=0.2384, pruned_loss=0.0432, over 4892.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2421, pruned_loss=0.04741, over 858315.97 frames. ], batch size: 35, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:08:20,551 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.962e+01 1.532e+02 1.796e+02 2.161e+02 5.029e+02, threshold=3.592e+02, percent-clipped=5.0 +2023-04-27 23:08:43,963 INFO [finetune.py:976] (3/7) Epoch 26, batch 500, loss[loss=0.1665, simple_loss=0.2351, pruned_loss=0.04897, over 4818.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2396, pruned_loss=0.0469, over 879163.46 frames. ], batch size: 25, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:09:17,409 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2479, 2.7353, 2.1102, 2.1635, 1.6581, 1.5924, 2.2842, 1.5080], + device='cuda:3'), covar=tensor([0.1582, 0.1396, 0.1346, 0.1605, 0.2176, 0.1849, 0.0948, 0.1978], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0209, 0.0168, 0.0203, 0.0200, 0.0184, 0.0156, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 23:09:27,563 INFO [finetune.py:976] (3/7) Epoch 26, batch 550, loss[loss=0.2246, simple_loss=0.2799, pruned_loss=0.08465, over 4872.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2358, pruned_loss=0.04589, over 895266.56 frames. ], batch size: 34, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:09:37,712 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143756.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:09:47,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.995e+01 1.539e+02 1.829e+02 2.212e+02 3.034e+02, threshold=3.659e+02, percent-clipped=1.0 +2023-04-27 23:10:00,518 INFO [finetune.py:976] (3/7) Epoch 26, batch 600, loss[loss=0.146, simple_loss=0.2253, pruned_loss=0.03333, over 4915.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2378, pruned_loss=0.04697, over 908361.46 frames. ], batch size: 37, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:10:03,633 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3593, 3.2964, 2.5864, 3.8916, 3.4168, 3.3894, 1.4438, 3.2961], + device='cuda:3'), covar=tensor([0.2077, 0.1718, 0.3192, 0.2454, 0.4515, 0.1918, 0.6334, 0.2833], + device='cuda:3'), in_proj_covar=tensor([0.0245, 0.0218, 0.0251, 0.0305, 0.0297, 0.0246, 0.0273, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:10:25,722 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143829.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:10:33,615 INFO [finetune.py:976] (3/7) Epoch 26, batch 650, loss[loss=0.1377, simple_loss=0.2213, pruned_loss=0.02706, over 4779.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2397, pruned_loss=0.04673, over 916430.05 frames. ], batch size: 29, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:11:14,501 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.206e+02 1.587e+02 1.916e+02 2.318e+02 7.608e+02, threshold=3.833e+02, percent-clipped=3.0 +2023-04-27 23:11:39,516 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3961, 1.1357, 0.3835, 1.0937, 1.0230, 1.2756, 1.1638, 1.1727], + device='cuda:3'), covar=tensor([0.0521, 0.0399, 0.0427, 0.0567, 0.0314, 0.0521, 0.0502, 0.0582], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 23:11:40,022 INFO [finetune.py:976] (3/7) Epoch 26, batch 700, loss[loss=0.2092, simple_loss=0.2933, pruned_loss=0.06249, over 4800.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2416, pruned_loss=0.04726, over 923522.17 frames. ], batch size: 45, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:12:45,799 INFO [finetune.py:976] (3/7) Epoch 26, batch 750, loss[loss=0.1457, simple_loss=0.2303, pruned_loss=0.03057, over 4865.00 frames. ], tot_loss[loss=0.169, simple_loss=0.243, pruned_loss=0.04746, over 931451.69 frames. ], batch size: 34, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:13:02,610 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9908, 1.6700, 2.1483, 2.3987, 2.0534, 1.8928, 2.0054, 2.0074], + device='cuda:3'), covar=tensor([0.4838, 0.7495, 0.7213, 0.5581, 0.5832, 0.9173, 0.8778, 0.9923], + device='cuda:3'), in_proj_covar=tensor([0.0443, 0.0424, 0.0518, 0.0511, 0.0471, 0.0507, 0.0509, 0.0521], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:13:26,583 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.623e+02 1.843e+02 2.196e+02 3.710e+02, threshold=3.686e+02, percent-clipped=0.0 +2023-04-27 23:13:56,910 INFO [finetune.py:976] (3/7) Epoch 26, batch 800, loss[loss=0.1693, simple_loss=0.2335, pruned_loss=0.05255, over 4746.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2421, pruned_loss=0.04689, over 937875.34 frames. ], batch size: 23, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:13:58,266 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1803, 2.7004, 2.0747, 2.2018, 1.6232, 1.5924, 2.3037, 1.4905], + device='cuda:3'), covar=tensor([0.1494, 0.1362, 0.1320, 0.1592, 0.2100, 0.1749, 0.0882, 0.1931], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0209, 0.0168, 0.0204, 0.0201, 0.0185, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 23:14:51,264 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7228, 1.7087, 0.8204, 1.3896, 1.8107, 1.5487, 1.4250, 1.5584], + device='cuda:3'), covar=tensor([0.0476, 0.0362, 0.0307, 0.0523, 0.0257, 0.0494, 0.0483, 0.0563], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0038, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 23:15:01,726 INFO [finetune.py:976] (3/7) Epoch 26, batch 850, loss[loss=0.1409, simple_loss=0.2093, pruned_loss=0.0363, over 4775.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2391, pruned_loss=0.04639, over 940707.28 frames. ], batch size: 29, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:15:15,466 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144056.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:15:35,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.501e+02 1.661e+02 2.180e+02 3.367e+02, threshold=3.322e+02, percent-clipped=0.0 +2023-04-27 23:16:05,584 INFO [finetune.py:976] (3/7) Epoch 26, batch 900, loss[loss=0.1488, simple_loss=0.2223, pruned_loss=0.03761, over 4817.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2352, pruned_loss=0.04492, over 943432.77 frames. ], batch size: 38, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:16:06,935 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0184, 2.4332, 1.9824, 1.9400, 1.5048, 1.5236, 2.1054, 1.4208], + device='cuda:3'), covar=tensor([0.1331, 0.1287, 0.1304, 0.1485, 0.2050, 0.1635, 0.0862, 0.1854], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0209, 0.0168, 0.0204, 0.0200, 0.0184, 0.0157, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-27 23:16:12,934 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144104.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:16:16,905 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 23:16:17,157 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8831, 1.9151, 1.0802, 1.4962, 2.1424, 1.6802, 1.6150, 1.6976], + device='cuda:3'), covar=tensor([0.0477, 0.0346, 0.0274, 0.0525, 0.0221, 0.0474, 0.0462, 0.0524], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 23:16:34,941 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144129.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:16:53,903 INFO [finetune.py:976] (3/7) Epoch 26, batch 950, loss[loss=0.2219, simple_loss=0.29, pruned_loss=0.07692, over 4902.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2334, pruned_loss=0.04447, over 944891.73 frames. ], batch size: 36, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:17:28,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.443e+02 1.800e+02 2.192e+02 5.909e+02, threshold=3.600e+02, percent-clipped=3.0 +2023-04-27 23:17:37,856 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144177.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:17:59,696 INFO [finetune.py:976] (3/7) Epoch 26, batch 1000, loss[loss=0.1576, simple_loss=0.2353, pruned_loss=0.03999, over 4788.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2363, pruned_loss=0.04584, over 947435.38 frames. ], batch size: 59, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:18:29,346 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 23:18:32,531 INFO [finetune.py:976] (3/7) Epoch 26, batch 1050, loss[loss=0.1893, simple_loss=0.2772, pruned_loss=0.0507, over 4852.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2391, pruned_loss=0.04627, over 949819.04 frames. ], batch size: 49, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:18:51,248 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.491e+02 1.893e+02 2.144e+02 7.486e+02, threshold=3.787e+02, percent-clipped=1.0 +2023-04-27 23:19:06,553 INFO [finetune.py:976] (3/7) Epoch 26, batch 1100, loss[loss=0.1991, simple_loss=0.2631, pruned_loss=0.06755, over 4793.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2401, pruned_loss=0.04651, over 951132.09 frames. ], batch size: 26, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:19:31,365 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 23:19:32,260 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3621, 2.9582, 2.5434, 2.8454, 2.1137, 2.6821, 2.7032, 2.1141], + device='cuda:3'), covar=tensor([0.2083, 0.1339, 0.0754, 0.1185, 0.2946, 0.0970, 0.2002, 0.2212], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0301, 0.0214, 0.0276, 0.0314, 0.0255, 0.0248, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1428e-04, 1.1875e-04, 8.4096e-05, 1.0887e-04, 1.2673e-04, 1.0029e-04, + 1.0011e-04, 1.0475e-04], device='cuda:3') +2023-04-27 23:19:39,801 INFO [finetune.py:976] (3/7) Epoch 26, batch 1150, loss[loss=0.1504, simple_loss=0.2185, pruned_loss=0.04117, over 4774.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2404, pruned_loss=0.04605, over 952057.56 frames. ], batch size: 26, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:19:59,724 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.546e+02 1.832e+02 2.212e+02 3.348e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-04-27 23:20:13,153 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2823, 1.8775, 2.1663, 2.5845, 2.5483, 2.1195, 1.7831, 2.4341], + device='cuda:3'), covar=tensor([0.0859, 0.1219, 0.0748, 0.0619, 0.0598, 0.0893, 0.0809, 0.0553], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0203, 0.0185, 0.0171, 0.0176, 0.0177, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:20:14,235 INFO [finetune.py:976] (3/7) Epoch 26, batch 1200, loss[loss=0.1679, simple_loss=0.2353, pruned_loss=0.05031, over 4744.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2389, pruned_loss=0.04537, over 952647.91 frames. ], batch size: 59, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:20:22,677 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 23:20:32,357 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144411.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:21:13,637 INFO [finetune.py:976] (3/7) Epoch 26, batch 1250, loss[loss=0.1359, simple_loss=0.1935, pruned_loss=0.0391, over 4286.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2373, pruned_loss=0.0459, over 950918.19 frames. ], batch size: 18, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:21:26,407 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.54 vs. limit=2.0 +2023-04-27 23:21:45,825 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.274e+01 1.443e+02 1.675e+02 2.129e+02 3.494e+02, threshold=3.349e+02, percent-clipped=0.0 +2023-04-27 23:21:51,778 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144472.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:22:16,353 INFO [finetune.py:976] (3/7) Epoch 26, batch 1300, loss[loss=0.1702, simple_loss=0.2406, pruned_loss=0.04993, over 4815.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2349, pruned_loss=0.04523, over 951893.88 frames. ], batch size: 51, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:22:48,068 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7127, 1.7414, 1.7712, 1.3749, 1.8947, 1.5204, 2.3512, 1.6365], + device='cuda:3'), covar=tensor([0.3527, 0.1629, 0.4661, 0.2589, 0.1547, 0.2351, 0.1464, 0.4320], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0351, 0.0425, 0.0349, 0.0381, 0.0374, 0.0366, 0.0423], + device='cuda:3'), out_proj_covar=tensor([9.9781e-05, 1.0456e-04, 1.2856e-04, 1.0471e-04, 1.1307e-04, 1.1114e-04, + 1.0704e-04, 1.2740e-04], device='cuda:3') +2023-04-27 23:23:06,076 INFO [finetune.py:976] (3/7) Epoch 26, batch 1350, loss[loss=0.1844, simple_loss=0.2584, pruned_loss=0.05525, over 4823.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2355, pruned_loss=0.04561, over 951608.43 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:23:09,590 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4548, 3.3318, 2.4687, 3.9055, 3.4943, 3.3845, 1.6670, 3.3319], + device='cuda:3'), covar=tensor([0.1966, 0.1532, 0.3343, 0.2368, 0.2922, 0.2070, 0.5873, 0.2782], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0219, 0.0252, 0.0306, 0.0299, 0.0247, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:23:26,217 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.435e+01 1.590e+02 1.903e+02 2.333e+02 4.606e+02, threshold=3.805e+02, percent-clipped=5.0 +2023-04-27 23:23:39,995 INFO [finetune.py:976] (3/7) Epoch 26, batch 1400, loss[loss=0.176, simple_loss=0.2695, pruned_loss=0.04123, over 4793.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2389, pruned_loss=0.04686, over 951913.95 frames. ], batch size: 29, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:23:41,220 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7946, 2.0444, 2.0180, 2.2044, 1.9145, 2.0457, 2.0688, 2.0394], + device='cuda:3'), covar=tensor([0.3808, 0.6192, 0.4823, 0.4257, 0.6031, 0.7053, 0.6108, 0.5462], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0374, 0.0328, 0.0339, 0.0349, 0.0393, 0.0359, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:23:51,208 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 23:24:01,452 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.8575, 4.7430, 3.2259, 5.5019, 4.8838, 4.7143, 2.3334, 4.7562], + device='cuda:3'), covar=tensor([0.1491, 0.0853, 0.3000, 0.0951, 0.4053, 0.1780, 0.5425, 0.2019], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0219, 0.0253, 0.0307, 0.0300, 0.0248, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:24:12,868 INFO [finetune.py:976] (3/7) Epoch 26, batch 1450, loss[loss=0.1724, simple_loss=0.2472, pruned_loss=0.0488, over 4874.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2412, pruned_loss=0.04689, over 953689.03 frames. ], batch size: 32, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:24:33,310 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.523e+02 1.914e+02 2.224e+02 4.652e+02, threshold=3.827e+02, percent-clipped=2.0 +2023-04-27 23:24:37,061 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6209, 1.7153, 0.7417, 1.3434, 1.7610, 1.4610, 1.3873, 1.4834], + device='cuda:3'), covar=tensor([0.0503, 0.0381, 0.0344, 0.0569, 0.0266, 0.0525, 0.0510, 0.0575], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 23:24:46,127 INFO [finetune.py:976] (3/7) Epoch 26, batch 1500, loss[loss=0.198, simple_loss=0.2677, pruned_loss=0.06415, over 4902.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2435, pruned_loss=0.04811, over 951735.06 frames. ], batch size: 36, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:25:06,382 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 23:25:20,122 INFO [finetune.py:976] (3/7) Epoch 26, batch 1550, loss[loss=0.1766, simple_loss=0.2429, pruned_loss=0.0552, over 4901.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2433, pruned_loss=0.04791, over 951945.47 frames. ], batch size: 37, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:25:42,637 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=144767.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:25:45,421 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.907e+01 1.462e+02 1.740e+02 2.100e+02 5.227e+02, threshold=3.480e+02, percent-clipped=2.0 +2023-04-27 23:26:14,647 INFO [finetune.py:976] (3/7) Epoch 26, batch 1600, loss[loss=0.1504, simple_loss=0.2209, pruned_loss=0.03994, over 4700.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2412, pruned_loss=0.04716, over 953808.97 frames. ], batch size: 23, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:26:32,722 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.71 vs. limit=5.0 +2023-04-27 23:26:36,105 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2681, 1.7417, 1.6119, 2.1368, 2.3409, 2.0208, 1.9712, 1.6795], + device='cuda:3'), covar=tensor([0.2097, 0.2032, 0.1925, 0.1516, 0.1453, 0.2135, 0.2208, 0.2554], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0309, 0.0352, 0.0288, 0.0330, 0.0306, 0.0299, 0.0376], + device='cuda:3'), out_proj_covar=tensor([6.4685e-05, 6.3499e-05, 7.3903e-05, 5.7690e-05, 6.7657e-05, 6.3933e-05, + 6.2031e-05, 7.9819e-05], device='cuda:3') +2023-04-27 23:26:58,881 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 23:27:00,875 INFO [finetune.py:976] (3/7) Epoch 26, batch 1650, loss[loss=0.1757, simple_loss=0.2452, pruned_loss=0.05307, over 4830.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2397, pruned_loss=0.04748, over 953688.01 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:27:20,928 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.459e+02 1.734e+02 2.287e+02 5.147e+02, threshold=3.469e+02, percent-clipped=2.0 +2023-04-27 23:27:39,743 INFO [finetune.py:976] (3/7) Epoch 26, batch 1700, loss[loss=0.153, simple_loss=0.224, pruned_loss=0.04105, over 4820.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2377, pruned_loss=0.04661, over 953463.24 frames. ], batch size: 41, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:28:12,694 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7144, 1.4150, 1.3124, 1.5170, 1.8433, 1.4963, 1.3176, 1.2878], + device='cuda:3'), covar=tensor([0.1698, 0.1383, 0.1907, 0.1157, 0.0944, 0.1702, 0.1832, 0.2247], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0306, 0.0349, 0.0285, 0.0328, 0.0304, 0.0297, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.4168e-05, 6.2951e-05, 7.3361e-05, 5.7192e-05, 6.7230e-05, 6.3438e-05, + 6.1573e-05, 7.9340e-05], device='cuda:3') +2023-04-27 23:28:44,464 INFO [finetune.py:976] (3/7) Epoch 26, batch 1750, loss[loss=0.1645, simple_loss=0.2456, pruned_loss=0.04174, over 4904.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2393, pruned_loss=0.04739, over 953612.69 frames. ], batch size: 36, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:29:25,331 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.536e+02 1.830e+02 2.200e+02 7.306e+02, threshold=3.661e+02, percent-clipped=1.0 +2023-04-27 23:29:50,042 INFO [finetune.py:976] (3/7) Epoch 26, batch 1800, loss[loss=0.2303, simple_loss=0.3166, pruned_loss=0.07197, over 4904.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.241, pruned_loss=0.04739, over 953740.39 frames. ], batch size: 43, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:29:53,361 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144996.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:10,216 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9693, 2.4100, 1.1536, 1.4247, 1.9263, 1.1837, 3.1121, 1.6265], + device='cuda:3'), covar=tensor([0.0681, 0.0636, 0.0767, 0.1148, 0.0480, 0.0980, 0.0235, 0.0603], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-27 23:30:14,868 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-27 23:30:15,072 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:22,865 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.53 vs. limit=5.0 +2023-04-27 23:30:24,557 INFO [finetune.py:976] (3/7) Epoch 26, batch 1850, loss[loss=0.1552, simple_loss=0.2306, pruned_loss=0.0399, over 4895.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2426, pruned_loss=0.04829, over 954569.50 frames. ], batch size: 37, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:30:25,221 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 23:30:34,404 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145057.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 23:30:41,230 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145066.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:41,820 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145067.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:44,629 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.513e+02 1.816e+02 2.184e+02 4.128e+02, threshold=3.632e+02, percent-clipped=2.0 +2023-04-27 23:30:55,919 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:58,305 INFO [finetune.py:976] (3/7) Epoch 26, batch 1900, loss[loss=0.1186, simple_loss=0.201, pruned_loss=0.01814, over 4791.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2442, pruned_loss=0.04897, over 954540.78 frames. ], batch size: 25, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:31:02,550 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6627, 3.3930, 0.8480, 1.8643, 1.9076, 2.3640, 1.9792, 1.0137], + device='cuda:3'), covar=tensor([0.1393, 0.1106, 0.2316, 0.1313, 0.1126, 0.1211, 0.1662, 0.2065], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0237, 0.0135, 0.0120, 0.0131, 0.0151, 0.0116, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 23:31:11,149 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 23:31:13,799 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:31:22,155 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145127.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:31:32,057 INFO [finetune.py:976] (3/7) Epoch 26, batch 1950, loss[loss=0.1706, simple_loss=0.2332, pruned_loss=0.05398, over 4720.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2435, pruned_loss=0.04876, over 955057.61 frames. ], batch size: 59, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:32:03,277 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.496e+01 1.516e+02 2.001e+02 2.404e+02 4.376e+02, threshold=4.002e+02, percent-clipped=4.0 +2023-04-27 23:32:11,439 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7138, 1.7388, 1.7889, 1.3429, 1.9082, 1.5323, 2.3541, 1.6275], + device='cuda:3'), covar=tensor([0.3404, 0.1737, 0.4129, 0.2746, 0.1397, 0.2107, 0.1392, 0.3990], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0350, 0.0424, 0.0350, 0.0382, 0.0374, 0.0366, 0.0421], + device='cuda:3'), out_proj_covar=tensor([9.9631e-05, 1.0427e-04, 1.2843e-04, 1.0507e-04, 1.1353e-04, 1.1112e-04, + 1.0714e-04, 1.2646e-04], device='cuda:3') +2023-04-27 23:32:16,019 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 23:32:23,670 INFO [finetune.py:976] (3/7) Epoch 26, batch 2000, loss[loss=0.1751, simple_loss=0.2488, pruned_loss=0.05063, over 4799.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2403, pruned_loss=0.04778, over 952636.81 frames. ], batch size: 51, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:32:51,439 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-27 23:33:00,190 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 23:33:03,264 INFO [finetune.py:976] (3/7) Epoch 26, batch 2050, loss[loss=0.1717, simple_loss=0.2417, pruned_loss=0.05088, over 4800.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2378, pruned_loss=0.04719, over 952154.65 frames. ], batch size: 51, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:33:32,589 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 23:33:43,226 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.513e+02 1.827e+02 2.273e+02 3.950e+02, threshold=3.653e+02, percent-clipped=0.0 +2023-04-27 23:33:54,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3372, 3.3125, 2.4916, 3.8449, 3.3314, 3.3179, 1.4920, 3.2317], + device='cuda:3'), covar=tensor([0.1963, 0.1456, 0.3374, 0.2199, 0.3818, 0.1974, 0.5798, 0.2867], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0217, 0.0251, 0.0302, 0.0298, 0.0245, 0.0272, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:33:58,466 INFO [finetune.py:976] (3/7) Epoch 26, batch 2100, loss[loss=0.1339, simple_loss=0.2026, pruned_loss=0.03265, over 4802.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2377, pruned_loss=0.04741, over 952861.37 frames. ], batch size: 25, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:34:15,437 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6023, 1.9364, 1.7594, 2.4640, 2.5282, 2.1172, 2.1121, 1.8236], + device='cuda:3'), covar=tensor([0.1716, 0.1827, 0.2140, 0.1572, 0.1668, 0.2273, 0.1976, 0.2717], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0308, 0.0350, 0.0286, 0.0329, 0.0305, 0.0297, 0.0375], + device='cuda:3'), out_proj_covar=tensor([6.4134e-05, 6.3341e-05, 7.3609e-05, 5.7224e-05, 6.7501e-05, 6.3704e-05, + 6.1623e-05, 7.9581e-05], device='cuda:3') +2023-04-27 23:34:32,515 INFO [finetune.py:976] (3/7) Epoch 26, batch 2150, loss[loss=0.1548, simple_loss=0.2364, pruned_loss=0.03661, over 4757.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2404, pruned_loss=0.04784, over 954638.14 frames. ], batch size: 27, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:34:39,203 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145352.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 23:34:44,132 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5395, 1.8034, 1.9357, 2.0105, 1.8083, 1.8941, 2.0139, 1.9433], + device='cuda:3'), covar=tensor([0.3627, 0.5255, 0.4231, 0.4237, 0.5390, 0.6853, 0.4653, 0.4676], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0375, 0.0329, 0.0340, 0.0350, 0.0394, 0.0361, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:34:45,719 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7832, 1.5933, 4.6734, 4.3978, 4.0702, 4.4307, 4.3689, 4.0719], + device='cuda:3'), covar=tensor([0.6885, 0.5627, 0.0977, 0.1731, 0.1245, 0.1597, 0.1065, 0.1616], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0309, 0.0410, 0.0411, 0.0349, 0.0417, 0.0321, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:34:51,070 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.597e+02 2.010e+02 2.319e+02 4.178e+02, threshold=4.020e+02, percent-clipped=2.0 +2023-04-27 23:34:51,185 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145371.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:34:59,921 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145383.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:35:10,687 INFO [finetune.py:976] (3/7) Epoch 26, batch 2200, loss[loss=0.1555, simple_loss=0.2302, pruned_loss=0.04042, over 4814.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2415, pruned_loss=0.04741, over 954463.66 frames. ], batch size: 25, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:35:11,379 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6655, 3.7013, 0.9302, 2.0249, 1.9658, 2.6680, 2.2480, 1.1721], + device='cuda:3'), covar=tensor([0.1269, 0.0838, 0.1964, 0.1178, 0.1017, 0.0951, 0.1364, 0.2190], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0135, 0.0120, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 23:35:22,851 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3936, 1.4382, 1.8611, 1.8029, 1.3276, 1.2293, 1.5520, 0.9452], + device='cuda:3'), covar=tensor([0.0607, 0.0572, 0.0346, 0.0663, 0.0749, 0.1068, 0.0582, 0.0602], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0069, 0.0074, 0.0094, 0.0073, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 23:35:30,059 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145422.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:35:30,105 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9840, 1.8568, 1.7191, 1.5290, 2.1055, 1.6536, 2.5812, 1.5682], + device='cuda:3'), covar=tensor([0.3477, 0.1950, 0.5452, 0.3105, 0.1674, 0.2305, 0.1272, 0.4346], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0352, 0.0426, 0.0352, 0.0385, 0.0375, 0.0368, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:35:37,115 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:35:44,061 INFO [finetune.py:976] (3/7) Epoch 26, batch 2250, loss[loss=0.1372, simple_loss=0.2008, pruned_loss=0.03685, over 4328.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2431, pruned_loss=0.04773, over 956080.23 frames. ], batch size: 18, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:35:45,838 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8911, 1.6971, 2.1276, 2.3631, 1.7163, 1.6761, 1.8592, 1.2891], + device='cuda:3'), covar=tensor([0.0481, 0.0678, 0.0386, 0.0502, 0.0654, 0.1044, 0.0612, 0.0586], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0074, 0.0095, 0.0073, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 23:35:51,954 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1545, 1.4407, 1.3001, 1.6914, 1.5619, 1.5456, 1.3808, 2.4320], + device='cuda:3'), covar=tensor([0.0597, 0.0839, 0.0813, 0.1208, 0.0662, 0.0501, 0.0764, 0.0228], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 23:36:03,178 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.640e+02 1.880e+02 2.244e+02 3.153e+02, threshold=3.761e+02, percent-clipped=0.0 +2023-04-27 23:36:17,801 INFO [finetune.py:976] (3/7) Epoch 26, batch 2300, loss[loss=0.1536, simple_loss=0.2209, pruned_loss=0.0431, over 4854.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2436, pruned_loss=0.04799, over 954708.45 frames. ], batch size: 31, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:36:32,306 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2392, 1.7777, 1.5849, 2.1708, 2.3280, 1.8919, 1.9213, 1.5798], + device='cuda:3'), covar=tensor([0.1539, 0.1423, 0.1662, 0.1301, 0.1063, 0.1650, 0.1752, 0.2109], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0308, 0.0349, 0.0285, 0.0328, 0.0304, 0.0296, 0.0373], + device='cuda:3'), out_proj_covar=tensor([6.3794e-05, 6.3299e-05, 7.3237e-05, 5.7000e-05, 6.7134e-05, 6.3471e-05, + 6.1313e-05, 7.9155e-05], device='cuda:3') +2023-04-27 23:36:51,030 INFO [finetune.py:976] (3/7) Epoch 26, batch 2350, loss[loss=0.1623, simple_loss=0.2334, pruned_loss=0.04566, over 4813.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2417, pruned_loss=0.04749, over 955578.56 frames. ], batch size: 41, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:37:10,064 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.548e+02 1.895e+02 2.111e+02 3.968e+02, threshold=3.791e+02, percent-clipped=2.0 +2023-04-27 23:37:40,486 INFO [finetune.py:976] (3/7) Epoch 26, batch 2400, loss[loss=0.1539, simple_loss=0.2223, pruned_loss=0.04276, over 4776.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2388, pruned_loss=0.04646, over 953840.53 frames. ], batch size: 26, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:37:41,821 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145594.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:38:46,679 INFO [finetune.py:976] (3/7) Epoch 26, batch 2450, loss[loss=0.1401, simple_loss=0.222, pruned_loss=0.02915, over 4892.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2363, pruned_loss=0.04556, over 954197.85 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:38:47,932 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:39:00,549 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145652.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:39:08,354 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145655.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:39:30,136 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.495e+02 1.759e+02 2.001e+02 3.629e+02, threshold=3.517e+02, percent-clipped=0.0 +2023-04-27 23:39:40,219 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-27 23:39:43,016 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145683.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:39:48,402 INFO [finetune.py:976] (3/7) Epoch 26, batch 2500, loss[loss=0.1761, simple_loss=0.2444, pruned_loss=0.05392, over 4806.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2363, pruned_loss=0.04552, over 951934.05 frames. ], batch size: 25, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:39:54,791 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145700.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:39:57,808 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:09,646 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145722.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:12,708 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145727.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:15,136 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145731.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:18,227 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.96 vs. limit=5.0 +2023-04-27 23:40:22,235 INFO [finetune.py:976] (3/7) Epoch 26, batch 2550, loss[loss=0.2077, simple_loss=0.2771, pruned_loss=0.06919, over 4822.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2391, pruned_loss=0.04627, over 952391.29 frames. ], batch size: 40, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:40:41,742 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145770.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:42,274 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.201e+01 1.571e+02 1.844e+02 2.174e+02 6.257e+02, threshold=3.689e+02, percent-clipped=2.0 +2023-04-27 23:40:46,768 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 23:40:56,103 INFO [finetune.py:976] (3/7) Epoch 26, batch 2600, loss[loss=0.1482, simple_loss=0.2208, pruned_loss=0.03777, over 4752.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2411, pruned_loss=0.04726, over 951600.71 frames. ], batch size: 26, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:40:57,526 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 23:41:04,950 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6490, 1.3762, 1.2719, 1.4874, 1.8754, 1.5558, 1.4053, 1.2083], + device='cuda:3'), covar=tensor([0.1916, 0.1386, 0.1661, 0.1306, 0.0827, 0.1391, 0.1647, 0.2248], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0308, 0.0349, 0.0285, 0.0327, 0.0303, 0.0297, 0.0373], + device='cuda:3'), out_proj_covar=tensor([6.3911e-05, 6.3335e-05, 7.3265e-05, 5.7051e-05, 6.6994e-05, 6.3240e-05, + 6.1515e-05, 7.9039e-05], device='cuda:3') +2023-04-27 23:41:16,285 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145821.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:41:29,916 INFO [finetune.py:976] (3/7) Epoch 26, batch 2650, loss[loss=0.1772, simple_loss=0.2589, pruned_loss=0.04769, over 4924.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2431, pruned_loss=0.04795, over 951296.97 frames. ], batch size: 42, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:41:40,501 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5868, 1.7571, 1.5835, 1.9384, 1.9342, 2.2630, 1.6349, 3.7006], + device='cuda:3'), covar=tensor([0.0492, 0.0704, 0.0684, 0.1017, 0.0530, 0.0576, 0.0669, 0.0161], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-27 23:41:49,916 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.525e+02 1.770e+02 2.195e+02 3.362e+02, threshold=3.540e+02, percent-clipped=0.0 +2023-04-27 23:41:57,242 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145882.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:42:03,157 INFO [finetune.py:976] (3/7) Epoch 26, batch 2700, loss[loss=0.1444, simple_loss=0.2232, pruned_loss=0.03278, over 4701.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2406, pruned_loss=0.04649, over 951435.31 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:42:36,792 INFO [finetune.py:976] (3/7) Epoch 26, batch 2750, loss[loss=0.1657, simple_loss=0.2427, pruned_loss=0.04435, over 4768.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.239, pruned_loss=0.04631, over 952670.15 frames. ], batch size: 27, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:42:40,535 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-27 23:42:42,209 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145950.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:42:56,787 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.527e+02 1.797e+02 2.320e+02 4.826e+02, threshold=3.594e+02, percent-clipped=4.0 +2023-04-27 23:43:10,064 INFO [finetune.py:976] (3/7) Epoch 26, batch 2800, loss[loss=0.1542, simple_loss=0.2226, pruned_loss=0.04287, over 4919.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2359, pruned_loss=0.04542, over 954911.75 frames. ], batch size: 37, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:43:14,390 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145999.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:43:35,202 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:43:40,212 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 23:43:41,205 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0620, 2.0291, 1.9614, 1.7291, 2.2290, 1.8916, 2.7388, 1.7766], + device='cuda:3'), covar=tensor([0.3540, 0.2119, 0.4474, 0.2891, 0.1723, 0.2381, 0.1430, 0.4403], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0351, 0.0425, 0.0350, 0.0382, 0.0373, 0.0368, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:43:44,744 INFO [finetune.py:976] (3/7) Epoch 26, batch 2850, loss[loss=0.2376, simple_loss=0.2943, pruned_loss=0.09044, over 4892.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2353, pruned_loss=0.04569, over 954613.22 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:44:22,218 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.964e+01 1.598e+02 1.831e+02 2.098e+02 3.575e+02, threshold=3.662e+02, percent-clipped=0.0 +2023-04-27 23:44:25,242 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146075.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:44:43,959 INFO [finetune.py:976] (3/7) Epoch 26, batch 2900, loss[loss=0.172, simple_loss=0.2454, pruned_loss=0.0493, over 4884.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2372, pruned_loss=0.04604, over 954284.09 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:45:01,312 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-04-27 23:45:49,356 INFO [finetune.py:976] (3/7) Epoch 26, batch 2950, loss[loss=0.2052, simple_loss=0.2687, pruned_loss=0.07084, over 4901.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2399, pruned_loss=0.04659, over 954146.00 frames. ], batch size: 35, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:45:59,523 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146149.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:46:13,681 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.553e+02 1.780e+02 2.193e+02 5.128e+02, threshold=3.559e+02, percent-clipped=3.0 +2023-04-27 23:46:18,907 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146177.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:46:28,439 INFO [finetune.py:976] (3/7) Epoch 26, batch 3000, loss[loss=0.1965, simple_loss=0.2754, pruned_loss=0.05879, over 4877.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2419, pruned_loss=0.0474, over 952781.75 frames. ], batch size: 43, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:46:28,439 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-27 23:46:38,927 INFO [finetune.py:1010] (3/7) Epoch 26, validation: loss=0.1526, simple_loss=0.2216, pruned_loss=0.04183, over 2265189.00 frames. +2023-04-27 23:46:38,927 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-27 23:46:41,518 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5148, 1.4808, 1.8974, 1.8955, 1.3363, 1.3051, 1.5348, 0.9543], + device='cuda:3'), covar=tensor([0.0521, 0.0567, 0.0331, 0.0563, 0.0686, 0.1055, 0.0551, 0.0596], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0074, 0.0095, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 23:46:46,405 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4259, 2.9631, 1.1195, 1.6814, 1.6789, 2.2165, 1.6935, 1.0570], + device='cuda:3'), covar=tensor([0.1264, 0.0910, 0.1567, 0.1237, 0.1028, 0.0868, 0.1434, 0.1774], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0236, 0.0135, 0.0120, 0.0130, 0.0151, 0.0116, 0.0117], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 23:46:50,698 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146210.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:47:11,572 INFO [finetune.py:976] (3/7) Epoch 26, batch 3050, loss[loss=0.1178, simple_loss=0.2026, pruned_loss=0.01647, over 4847.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2422, pruned_loss=0.04642, over 953974.24 frames. ], batch size: 47, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:47:17,398 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146250.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:47:26,071 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 23:47:31,021 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.541e+02 1.745e+02 2.066e+02 6.833e+02, threshold=3.490e+02, percent-clipped=2.0 +2023-04-27 23:47:45,140 INFO [finetune.py:976] (3/7) Epoch 26, batch 3100, loss[loss=0.1824, simple_loss=0.2527, pruned_loss=0.05611, over 4842.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2406, pruned_loss=0.04643, over 952723.53 frames. ], batch size: 44, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:47:49,795 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146298.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:47:50,464 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:47:56,513 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5414, 2.0829, 2.4940, 3.1534, 2.3253, 1.9047, 1.9930, 2.3150], + device='cuda:3'), covar=tensor([0.3608, 0.3230, 0.1731, 0.2281, 0.3026, 0.2782, 0.3832, 0.2087], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0246, 0.0228, 0.0313, 0.0221, 0.0235, 0.0227, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-27 23:48:10,927 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4142, 2.2034, 2.3093, 2.7634, 2.7549, 2.1713, 2.0567, 2.5609], + device='cuda:3'), covar=tensor([0.0792, 0.0952, 0.0701, 0.0561, 0.0604, 0.0830, 0.0704, 0.0495], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0183, 0.0170, 0.0176, 0.0177, 0.0149, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:48:18,964 INFO [finetune.py:976] (3/7) Epoch 26, batch 3150, loss[loss=0.1502, simple_loss=0.2162, pruned_loss=0.04208, over 4780.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2388, pruned_loss=0.04639, over 952859.27 frames. ], batch size: 28, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:48:22,567 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146347.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:48:38,469 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.566e+02 1.850e+02 2.179e+02 3.570e+02, threshold=3.699e+02, percent-clipped=2.0 +2023-04-27 23:48:51,064 INFO [finetune.py:976] (3/7) Epoch 26, batch 3200, loss[loss=0.1575, simple_loss=0.2195, pruned_loss=0.04776, over 4437.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2354, pruned_loss=0.04558, over 952455.65 frames. ], batch size: 19, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:48:59,148 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5414, 1.5513, 4.2215, 3.9220, 3.6650, 4.0571, 4.0509, 3.7076], + device='cuda:3'), covar=tensor([0.6891, 0.5789, 0.1093, 0.1883, 0.1225, 0.1693, 0.1197, 0.1568], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0311, 0.0411, 0.0413, 0.0351, 0.0418, 0.0321, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:49:17,616 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([4.8595, 4.7855, 3.2906, 5.5136, 4.9134, 4.7641, 2.2818, 4.6702], + device='cuda:3'), covar=tensor([0.1515, 0.1080, 0.2997, 0.0791, 0.5700, 0.1719, 0.5665, 0.2087], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0219, 0.0252, 0.0305, 0.0300, 0.0247, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:49:24,699 INFO [finetune.py:976] (3/7) Epoch 26, batch 3250, loss[loss=0.1791, simple_loss=0.2611, pruned_loss=0.04854, over 4896.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2355, pruned_loss=0.04543, over 952811.87 frames. ], batch size: 35, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:49:45,355 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.653e+01 1.477e+02 1.784e+02 2.151e+02 4.577e+02, threshold=3.568e+02, percent-clipped=2.0 +2023-04-27 23:49:53,875 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146477.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:50:06,070 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 23:50:13,002 INFO [finetune.py:976] (3/7) Epoch 26, batch 3300, loss[loss=0.1625, simple_loss=0.2527, pruned_loss=0.03611, over 4790.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2393, pruned_loss=0.0469, over 952707.66 frames. ], batch size: 29, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:50:28,263 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146505.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:50:50,087 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6991, 1.4588, 1.6296, 1.9400, 1.9793, 1.5115, 1.3593, 1.7732], + device='cuda:3'), covar=tensor([0.0717, 0.1142, 0.0705, 0.0515, 0.0577, 0.0815, 0.0739, 0.0575], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0201, 0.0184, 0.0170, 0.0176, 0.0177, 0.0149, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:50:57,305 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146525.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:51:19,413 INFO [finetune.py:976] (3/7) Epoch 26, batch 3350, loss[loss=0.2016, simple_loss=0.2668, pruned_loss=0.06818, over 4796.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2416, pruned_loss=0.04713, over 953221.52 frames. ], batch size: 29, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:51:37,954 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-04-27 23:51:42,428 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.17 vs. limit=2.0 +2023-04-27 23:51:42,829 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9411, 2.1494, 2.1911, 2.2688, 2.0390, 2.1961, 2.2953, 2.1885], + device='cuda:3'), covar=tensor([0.3541, 0.6455, 0.4945, 0.4708, 0.6007, 0.7168, 0.5762, 0.5518], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0328, 0.0339, 0.0349, 0.0394, 0.0358, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:51:45,686 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.686e+01 1.611e+02 1.865e+02 2.186e+02 4.215e+02, threshold=3.729e+02, percent-clipped=3.0 +2023-04-27 23:51:54,045 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 23:51:57,913 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146591.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:51:58,395 INFO [finetune.py:976] (3/7) Epoch 26, batch 3400, loss[loss=0.1201, simple_loss=0.194, pruned_loss=0.02315, over 4748.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.242, pruned_loss=0.04749, over 950437.56 frames. ], batch size: 27, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:52:00,896 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2405, 1.1919, 3.8357, 3.5899, 3.3999, 3.7020, 3.7152, 3.3808], + device='cuda:3'), covar=tensor([0.7642, 0.6016, 0.1176, 0.1922, 0.1372, 0.1575, 0.1513, 0.1622], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0310, 0.0410, 0.0411, 0.0350, 0.0416, 0.0320, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:52:05,175 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0152, 2.4568, 2.0352, 2.4210, 1.7083, 2.1041, 2.1178, 1.6390], + device='cuda:3'), covar=tensor([0.2158, 0.1323, 0.0846, 0.1163, 0.3320, 0.1143, 0.2070, 0.2540], + device='cuda:3'), in_proj_covar=tensor([0.0289, 0.0303, 0.0218, 0.0280, 0.0317, 0.0256, 0.0251, 0.0267], + device='cuda:3'), out_proj_covar=tensor([1.1528e-04, 1.1925e-04, 8.5714e-05, 1.1021e-04, 1.2772e-04, 1.0089e-04, + 1.0135e-04, 1.0528e-04], device='cuda:3') +2023-04-27 23:52:27,935 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146626.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:52:43,198 INFO [finetune.py:976] (3/7) Epoch 26, batch 3450, loss[loss=0.1568, simple_loss=0.2355, pruned_loss=0.03899, over 4786.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2416, pruned_loss=0.04736, over 950088.27 frames. ], batch size: 51, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:52:49,805 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146652.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:53:04,800 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.587e+02 1.818e+02 2.131e+02 4.084e+02, threshold=3.635e+02, percent-clipped=1.0 +2023-04-27 23:53:13,981 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146687.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:53:16,856 INFO [finetune.py:976] (3/7) Epoch 26, batch 3500, loss[loss=0.185, simple_loss=0.2478, pruned_loss=0.06108, over 4839.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2409, pruned_loss=0.04794, over 951105.70 frames. ], batch size: 33, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:53:27,682 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6549, 0.6799, 1.5247, 1.9863, 1.6975, 1.5619, 1.5778, 1.5993], + device='cuda:3'), covar=tensor([0.4177, 0.6015, 0.5572, 0.5960, 0.5424, 0.7206, 0.7065, 0.7960], + device='cuda:3'), in_proj_covar=tensor([0.0440, 0.0420, 0.0514, 0.0506, 0.0468, 0.0505, 0.0505, 0.0518], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:53:50,646 INFO [finetune.py:976] (3/7) Epoch 26, batch 3550, loss[loss=0.148, simple_loss=0.2256, pruned_loss=0.03522, over 4861.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2387, pruned_loss=0.0473, over 953354.78 frames. ], batch size: 31, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:53:53,804 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4382, 1.3736, 1.7417, 1.7768, 1.3553, 1.2788, 1.4238, 0.9477], + device='cuda:3'), covar=tensor([0.0485, 0.0593, 0.0331, 0.0472, 0.0720, 0.0935, 0.0587, 0.0507], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0067, 0.0065, 0.0068, 0.0073, 0.0093, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-27 23:54:01,607 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6872, 1.5635, 0.6005, 1.3216, 1.4394, 1.5225, 1.4384, 1.4203], + device='cuda:3'), covar=tensor([0.0515, 0.0380, 0.0383, 0.0574, 0.0291, 0.0508, 0.0519, 0.0586], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 23:54:11,377 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.892e+01 1.463e+02 1.726e+02 2.142e+02 5.887e+02, threshold=3.452e+02, percent-clipped=4.0 +2023-04-27 23:54:24,573 INFO [finetune.py:976] (3/7) Epoch 26, batch 3600, loss[loss=0.1372, simple_loss=0.2069, pruned_loss=0.03371, over 4769.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2363, pruned_loss=0.04681, over 949939.35 frames. ], batch size: 28, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:54:32,778 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146805.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:54:59,552 INFO [finetune.py:976] (3/7) Epoch 26, batch 3650, loss[loss=0.1563, simple_loss=0.2385, pruned_loss=0.03706, over 4782.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2369, pruned_loss=0.04673, over 950175.33 frames. ], batch size: 26, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:55:00,929 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146844.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:55:03,360 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7855, 2.8519, 2.2998, 2.4483, 2.9486, 2.4470, 3.7965, 2.1211], + device='cuda:3'), covar=tensor([0.3436, 0.2041, 0.3923, 0.3602, 0.1700, 0.2655, 0.1224, 0.4120], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0351, 0.0422, 0.0349, 0.0381, 0.0373, 0.0364, 0.0421], + device='cuda:3'), out_proj_covar=tensor([9.9798e-05, 1.0471e-04, 1.2774e-04, 1.0455e-04, 1.1292e-04, 1.1071e-04, + 1.0661e-04, 1.2676e-04], device='cuda:3') +2023-04-27 23:55:11,629 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146853.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:55:30,101 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.832e+01 1.619e+02 1.915e+02 2.507e+02 7.747e+02, threshold=3.830e+02, percent-clipped=2.0 +2023-04-27 23:55:47,895 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9925, 1.6571, 1.8203, 2.2322, 2.2135, 1.8641, 1.7024, 2.0698], + device='cuda:3'), covar=tensor([0.0575, 0.1112, 0.0678, 0.0447, 0.0542, 0.0676, 0.0619, 0.0497], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0202, 0.0184, 0.0170, 0.0176, 0.0178, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-27 23:55:48,367 INFO [finetune.py:976] (3/7) Epoch 26, batch 3700, loss[loss=0.2298, simple_loss=0.2956, pruned_loss=0.082, over 4201.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.241, pruned_loss=0.04781, over 951982.21 frames. ], batch size: 65, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:56:01,844 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146905.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:56:36,151 INFO [finetune.py:976] (3/7) Epoch 26, batch 3750, loss[loss=0.1654, simple_loss=0.245, pruned_loss=0.04292, over 4764.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.243, pruned_loss=0.04818, over 953644.84 frames. ], batch size: 26, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:56:44,462 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146947.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:56:57,504 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-27 23:57:06,090 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.923e+01 1.552e+02 1.944e+02 2.307e+02 4.306e+02, threshold=3.888e+02, percent-clipped=2.0 +2023-04-27 23:57:10,298 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4878, 1.0592, 0.3526, 1.1673, 1.0523, 1.3696, 1.2553, 1.2387], + device='cuda:3'), covar=tensor([0.0509, 0.0407, 0.0408, 0.0563, 0.0310, 0.0511, 0.0486, 0.0591], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-27 23:57:13,818 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146982.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:57:15,607 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3275, 1.5725, 1.9369, 2.0217, 1.9805, 2.0709, 1.9568, 1.9885], + device='cuda:3'), covar=tensor([0.3626, 0.5028, 0.4237, 0.4488, 0.5202, 0.6783, 0.4412, 0.4149], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0375, 0.0329, 0.0340, 0.0350, 0.0395, 0.0359, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-27 23:57:20,816 INFO [finetune.py:976] (3/7) Epoch 26, batch 3800, loss[loss=0.1614, simple_loss=0.2326, pruned_loss=0.04514, over 4834.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2448, pruned_loss=0.04877, over 956106.20 frames. ], batch size: 30, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:58:05,507 INFO [finetune.py:976] (3/7) Epoch 26, batch 3850, loss[loss=0.1654, simple_loss=0.2431, pruned_loss=0.04382, over 4770.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2427, pruned_loss=0.04737, over 956375.29 frames. ], batch size: 51, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:58:23,781 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.597e+02 1.804e+02 2.116e+02 3.825e+02, threshold=3.609e+02, percent-clipped=0.0 +2023-04-27 23:58:39,181 INFO [finetune.py:976] (3/7) Epoch 26, batch 3900, loss[loss=0.17, simple_loss=0.2365, pruned_loss=0.05177, over 4894.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2411, pruned_loss=0.04759, over 956855.91 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:59:02,010 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-27 23:59:12,136 INFO [finetune.py:976] (3/7) Epoch 26, batch 3950, loss[loss=0.1464, simple_loss=0.2128, pruned_loss=0.03996, over 4876.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2376, pruned_loss=0.04654, over 956007.93 frames. ], batch size: 31, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:59:31,437 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.596e+02 1.912e+02 2.259e+02 3.879e+02, threshold=3.825e+02, percent-clipped=1.0 +2023-04-27 23:59:45,578 INFO [finetune.py:976] (3/7) Epoch 26, batch 4000, loss[loss=0.1781, simple_loss=0.2551, pruned_loss=0.05052, over 4822.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2387, pruned_loss=0.04748, over 954730.16 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:59:46,139 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 23:59:51,630 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147200.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:00:18,668 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147241.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:00:19,193 INFO [finetune.py:976] (3/7) Epoch 26, batch 4050, loss[loss=0.1492, simple_loss=0.2372, pruned_loss=0.03065, over 4743.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2418, pruned_loss=0.04829, over 952878.58 frames. ], batch size: 27, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:00:22,075 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4177, 1.4400, 1.7518, 2.6305, 2.7888, 1.9432, 1.7887, 2.2591], + device='cuda:3'), covar=tensor([0.0774, 0.1794, 0.1151, 0.0560, 0.0561, 0.1112, 0.0887, 0.0700], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0201, 0.0184, 0.0170, 0.0176, 0.0177, 0.0150, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:00:23,233 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147247.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:00:44,474 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.656e+02 1.880e+02 2.300e+02 3.890e+02, threshold=3.760e+02, percent-clipped=1.0 +2023-04-28 00:00:55,664 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147282.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:10,422 INFO [finetune.py:976] (3/7) Epoch 26, batch 4100, loss[loss=0.193, simple_loss=0.2749, pruned_loss=0.05552, over 4803.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2437, pruned_loss=0.0485, over 955064.79 frames. ], batch size: 39, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:01:12,318 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147295.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:23,299 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147302.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:34,373 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147310.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:52,850 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3963, 3.0047, 0.8859, 1.6608, 1.6984, 2.1552, 1.6982, 1.0848], + device='cuda:3'), covar=tensor([0.1361, 0.0829, 0.1909, 0.1273, 0.1133, 0.0977, 0.1611, 0.1924], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0239, 0.0136, 0.0121, 0.0132, 0.0153, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:01:54,624 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147330.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:02:04,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9531, 1.6687, 4.4057, 4.1319, 3.8934, 4.1711, 4.1881, 3.9501], + device='cuda:3'), covar=tensor([0.6720, 0.5365, 0.1117, 0.1761, 0.1207, 0.1589, 0.1005, 0.1447], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0306, 0.0404, 0.0406, 0.0345, 0.0411, 0.0316, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:02:13,693 INFO [finetune.py:976] (3/7) Epoch 26, batch 4150, loss[loss=0.1657, simple_loss=0.2412, pruned_loss=0.04507, over 4221.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2443, pruned_loss=0.04865, over 953082.99 frames. ], batch size: 66, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:02:23,930 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7491, 1.9348, 5.5599, 5.2262, 4.8075, 5.2314, 4.8909, 4.9266], + device='cuda:3'), covar=tensor([0.5632, 0.5587, 0.0834, 0.1623, 0.0924, 0.0901, 0.0821, 0.1413], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0306, 0.0404, 0.0406, 0.0345, 0.0411, 0.0316, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:02:56,069 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147371.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:02:56,396 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 00:02:56,542 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.010e+02 1.642e+02 1.885e+02 2.342e+02 4.093e+02, threshold=3.770e+02, percent-clipped=2.0 +2023-04-28 00:02:59,477 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-28 00:03:02,615 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-28 00:03:09,328 INFO [finetune.py:976] (3/7) Epoch 26, batch 4200, loss[loss=0.1755, simple_loss=0.252, pruned_loss=0.04948, over 4170.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2439, pruned_loss=0.04847, over 950418.72 frames. ], batch size: 65, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:03:21,132 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147407.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:03:27,614 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3103, 2.1959, 1.9681, 1.9642, 2.3916, 1.8626, 2.9207, 1.8070], + device='cuda:3'), covar=tensor([0.3390, 0.1871, 0.3864, 0.2721, 0.1630, 0.2402, 0.1387, 0.3695], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0353, 0.0424, 0.0349, 0.0383, 0.0373, 0.0365, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:03:37,883 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147434.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:03:40,959 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5937, 1.8907, 1.5929, 1.8376, 1.3721, 1.5323, 1.5238, 1.2135], + device='cuda:3'), covar=tensor([0.2063, 0.1689, 0.1120, 0.1501, 0.3684, 0.1542, 0.2060, 0.2448], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0301, 0.0217, 0.0279, 0.0317, 0.0256, 0.0251, 0.0265], + device='cuda:3'), out_proj_covar=tensor([1.1429e-04, 1.1855e-04, 8.5525e-05, 1.0976e-04, 1.2777e-04, 1.0061e-04, + 1.0123e-04, 1.0450e-04], device='cuda:3') +2023-04-28 00:03:42,686 INFO [finetune.py:976] (3/7) Epoch 26, batch 4250, loss[loss=0.1915, simple_loss=0.2613, pruned_loss=0.06088, over 4865.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2416, pruned_loss=0.04723, over 950746.64 frames. ], batch size: 31, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:03:53,270 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-28 00:04:01,787 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147468.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:04:04,153 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.178e+01 1.415e+02 1.781e+02 2.171e+02 8.734e+02, threshold=3.562e+02, percent-clipped=6.0 +2023-04-28 00:04:16,305 INFO [finetune.py:976] (3/7) Epoch 26, batch 4300, loss[loss=0.2005, simple_loss=0.2639, pruned_loss=0.06854, over 4868.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2385, pruned_loss=0.04676, over 950965.15 frames. ], batch size: 31, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:04:18,268 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147495.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:04:18,858 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0789, 1.7434, 1.9829, 2.3238, 2.4214, 1.8038, 1.6685, 2.1572], + device='cuda:3'), covar=tensor([0.0802, 0.1162, 0.0714, 0.0532, 0.0557, 0.0896, 0.0747, 0.0533], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0184, 0.0170, 0.0176, 0.0177, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:04:22,177 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147500.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:04:37,651 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.09 vs. limit=2.0 +2023-04-28 00:04:41,643 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3724, 1.9810, 2.2578, 2.6264, 2.6325, 2.0887, 1.8152, 2.4417], + device='cuda:3'), covar=tensor([0.0753, 0.1126, 0.0765, 0.0591, 0.0649, 0.0974, 0.0807, 0.0544], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0184, 0.0170, 0.0177, 0.0177, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:04:49,577 INFO [finetune.py:976] (3/7) Epoch 26, batch 4350, loss[loss=0.1628, simple_loss=0.2238, pruned_loss=0.0509, over 4925.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2352, pruned_loss=0.04567, over 951547.52 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:04:53,338 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147548.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:05:07,385 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9557, 2.1118, 1.2514, 1.7006, 2.2466, 1.7559, 1.7043, 1.8976], + device='cuda:3'), covar=tensor([0.0451, 0.0349, 0.0276, 0.0522, 0.0226, 0.0481, 0.0477, 0.0503], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:3') +2023-04-28 00:05:10,813 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.116e+01 1.521e+02 1.686e+02 2.044e+02 4.229e+02, threshold=3.372e+02, percent-clipped=1.0 +2023-04-28 00:05:23,018 INFO [finetune.py:976] (3/7) Epoch 26, batch 4400, loss[loss=0.1644, simple_loss=0.2528, pruned_loss=0.03798, over 4834.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2367, pruned_loss=0.04609, over 950730.73 frames. ], batch size: 47, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:05:26,139 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147597.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:05:38,848 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6734, 2.5730, 2.1375, 2.3711, 2.7476, 2.2277, 3.4790, 2.0302], + device='cuda:3'), covar=tensor([0.3345, 0.2313, 0.4120, 0.3030, 0.1583, 0.2366, 0.1329, 0.3984], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0353, 0.0424, 0.0349, 0.0381, 0.0373, 0.0365, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:05:56,089 INFO [finetune.py:976] (3/7) Epoch 26, batch 4450, loss[loss=0.1403, simple_loss=0.209, pruned_loss=0.03576, over 4719.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2409, pruned_loss=0.04736, over 951111.52 frames. ], batch size: 23, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:06:01,114 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7637, 1.3282, 1.8174, 2.2668, 1.8680, 1.7072, 1.7750, 1.6906], + device='cuda:3'), covar=tensor([0.4413, 0.7253, 0.6365, 0.5337, 0.5928, 0.8084, 0.8339, 1.0651], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0421, 0.0515, 0.0507, 0.0469, 0.0505, 0.0507, 0.0519], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:06:17,563 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147659.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:06:22,786 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147666.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:06:29,371 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2967, 2.2659, 2.0970, 2.0634, 2.5488, 1.9819, 3.0164, 1.8039], + device='cuda:3'), covar=tensor([0.3573, 0.2081, 0.4317, 0.3097, 0.1405, 0.2372, 0.1218, 0.4309], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0354, 0.0424, 0.0349, 0.0382, 0.0373, 0.0365, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:06:31,478 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.516e+02 1.781e+02 2.208e+02 3.417e+02, threshold=3.563e+02, percent-clipped=1.0 +2023-04-28 00:06:44,118 INFO [finetune.py:976] (3/7) Epoch 26, batch 4500, loss[loss=0.1928, simple_loss=0.2605, pruned_loss=0.06254, over 4804.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2426, pruned_loss=0.04773, over 953198.25 frames. ], batch size: 45, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:06:51,731 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2478, 1.5193, 1.7079, 1.7927, 1.7181, 1.8153, 1.7488, 1.7476], + device='cuda:3'), covar=tensor([0.3746, 0.4760, 0.3965, 0.4174, 0.5475, 0.6864, 0.4658, 0.4307], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0373, 0.0328, 0.0339, 0.0350, 0.0395, 0.0360, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 00:07:17,731 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147720.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:07:37,055 INFO [finetune.py:976] (3/7) Epoch 26, batch 4550, loss[loss=0.13, simple_loss=0.2206, pruned_loss=0.01976, over 4769.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2443, pruned_loss=0.04854, over 953183.46 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:07:59,443 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.73 vs. limit=5.0 +2023-04-28 00:08:01,016 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147763.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:08:14,011 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.503e+02 1.815e+02 2.151e+02 5.912e+02, threshold=3.631e+02, percent-clipped=1.0 +2023-04-28 00:08:21,937 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9049, 1.6767, 1.8805, 2.2644, 2.2435, 1.8705, 1.4174, 2.0627], + device='cuda:3'), covar=tensor([0.0692, 0.1168, 0.0778, 0.0494, 0.0534, 0.0743, 0.0766, 0.0500], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0201, 0.0184, 0.0170, 0.0177, 0.0177, 0.0151, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:08:31,440 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147790.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:08:32,585 INFO [finetune.py:976] (3/7) Epoch 26, batch 4600, loss[loss=0.1964, simple_loss=0.2677, pruned_loss=0.06256, over 4917.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2433, pruned_loss=0.04834, over 951428.76 frames. ], batch size: 38, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:08:47,350 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147815.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:06,110 INFO [finetune.py:976] (3/7) Epoch 26, batch 4650, loss[loss=0.1559, simple_loss=0.2299, pruned_loss=0.04099, over 4790.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2407, pruned_loss=0.04777, over 949687.81 frames. ], batch size: 25, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:09:15,422 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147857.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:25,965 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 7.854e+01 1.470e+02 1.652e+02 1.982e+02 3.804e+02, threshold=3.304e+02, percent-clipped=1.0 +2023-04-28 00:09:29,057 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147876.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:37,838 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-28 00:09:40,051 INFO [finetune.py:976] (3/7) Epoch 26, batch 4700, loss[loss=0.1297, simple_loss=0.2116, pruned_loss=0.02387, over 4765.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2382, pruned_loss=0.04671, over 952282.36 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:09:43,177 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147897.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:51,636 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147911.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:55,892 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147918.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:02,996 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-28 00:10:12,814 INFO [finetune.py:976] (3/7) Epoch 26, batch 4750, loss[loss=0.159, simple_loss=0.2291, pruned_loss=0.04445, over 4752.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2364, pruned_loss=0.04631, over 952231.51 frames. ], batch size: 26, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:10:15,201 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147945.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:15,886 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8828, 2.0602, 1.8475, 1.5700, 1.3783, 1.4210, 1.9397, 1.3493], + device='cuda:3'), covar=tensor([0.1743, 0.1456, 0.1369, 0.1696, 0.2404, 0.1900, 0.0979, 0.2053], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0211, 0.0169, 0.0203, 0.0200, 0.0186, 0.0156, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 00:10:26,622 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 00:10:28,083 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147966.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:31,603 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.547e+02 1.796e+02 2.219e+02 3.890e+02, threshold=3.592e+02, percent-clipped=1.0 +2023-04-28 00:10:31,735 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147972.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:46,578 INFO [finetune.py:976] (3/7) Epoch 26, batch 4800, loss[loss=0.1383, simple_loss=0.2234, pruned_loss=0.02658, over 4877.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2409, pruned_loss=0.04783, over 952573.38 frames. ], batch size: 34, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:11:01,950 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148014.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:11:02,582 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148015.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:11:05,681 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7994, 1.3598, 1.9028, 2.3222, 1.8574, 1.7802, 1.8473, 1.7883], + device='cuda:3'), covar=tensor([0.4809, 0.7161, 0.6724, 0.5555, 0.6238, 0.8730, 0.8414, 0.9944], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0422, 0.0515, 0.0508, 0.0470, 0.0506, 0.0507, 0.0520], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:11:21,741 INFO [finetune.py:976] (3/7) Epoch 26, batch 4850, loss[loss=0.1775, simple_loss=0.2662, pruned_loss=0.0444, over 4825.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.243, pruned_loss=0.04806, over 953571.35 frames. ], batch size: 47, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:11:40,950 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8595, 1.0604, 1.6493, 1.7556, 1.7183, 1.7316, 1.6093, 1.6601], + device='cuda:3'), covar=tensor([0.3793, 0.5387, 0.4290, 0.4223, 0.5496, 0.7058, 0.4776, 0.4264], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0374, 0.0328, 0.0339, 0.0350, 0.0395, 0.0360, 0.0331], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 00:11:44,031 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2101, 1.1580, 1.2155, 1.5055, 1.4968, 1.2339, 0.9444, 1.4030], + device='cuda:3'), covar=tensor([0.0758, 0.1261, 0.0793, 0.0540, 0.0591, 0.0803, 0.0825, 0.0521], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0202, 0.0184, 0.0171, 0.0177, 0.0177, 0.0152, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:11:44,618 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148063.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:11:55,267 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.622e+02 1.940e+02 2.340e+02 4.108e+02, threshold=3.881e+02, percent-clipped=3.0 +2023-04-28 00:11:56,002 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8723, 1.7550, 2.2190, 2.2523, 1.5725, 1.4976, 1.9191, 1.1263], + device='cuda:3'), covar=tensor([0.0573, 0.0651, 0.0420, 0.0861, 0.0798, 0.1122, 0.0649, 0.0729], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0067, 0.0067, 0.0069, 0.0074, 0.0094, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:12:18,046 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148090.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:12:24,897 INFO [finetune.py:976] (3/7) Epoch 26, batch 4900, loss[loss=0.2164, simple_loss=0.2845, pruned_loss=0.0741, over 4879.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2441, pruned_loss=0.04815, over 953816.72 frames. ], batch size: 43, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:12:47,897 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148111.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:13:01,809 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2595, 3.2547, 0.7395, 1.5778, 1.6676, 2.2594, 1.7926, 1.0557], + device='cuda:3'), covar=tensor([0.2050, 0.1564, 0.2746, 0.1946, 0.1523, 0.1501, 0.2013, 0.2443], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0135, 0.0121, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:13:09,208 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 00:13:22,192 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148138.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:13:30,215 INFO [finetune.py:976] (3/7) Epoch 26, batch 4950, loss[loss=0.1316, simple_loss=0.2177, pruned_loss=0.02275, over 4752.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2449, pruned_loss=0.04831, over 953909.00 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:13:35,107 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0353, 1.0578, 1.1500, 1.1778, 1.0112, 0.9573, 1.0270, 0.4203], + device='cuda:3'), covar=tensor([0.0539, 0.0496, 0.0455, 0.0474, 0.0668, 0.1135, 0.0415, 0.0668], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0068, 0.0067, 0.0069, 0.0074, 0.0094, 0.0073, 0.0064], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:13:52,207 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148171.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:13:52,736 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.839e+01 1.452e+02 1.801e+02 2.197e+02 4.892e+02, threshold=3.601e+02, percent-clipped=1.0 +2023-04-28 00:14:00,245 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5518, 2.9920, 2.8003, 2.9965, 2.2720, 2.6303, 2.7792, 2.1667], + device='cuda:3'), covar=tensor([0.1897, 0.1117, 0.0603, 0.0951, 0.2875, 0.0873, 0.1756, 0.2488], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0298, 0.0214, 0.0276, 0.0313, 0.0253, 0.0248, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1306e-04, 1.1742e-04, 8.4271e-05, 1.0843e-04, 1.2604e-04, 9.9527e-05, + 1.0021e-04, 1.0339e-04], device='cuda:3') +2023-04-28 00:14:06,564 INFO [finetune.py:976] (3/7) Epoch 26, batch 5000, loss[loss=0.1579, simple_loss=0.2271, pruned_loss=0.04435, over 4815.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2428, pruned_loss=0.0477, over 954238.04 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:14:21,042 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148213.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:14:24,159 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6022, 1.3397, 1.3790, 1.3980, 1.8427, 1.4509, 1.2847, 1.2523], + device='cuda:3'), covar=tensor([0.1980, 0.1466, 0.1953, 0.1497, 0.0851, 0.2040, 0.2301, 0.2728], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0310, 0.0353, 0.0287, 0.0328, 0.0305, 0.0299, 0.0376], + device='cuda:3'), out_proj_covar=tensor([6.4260e-05, 6.3846e-05, 7.4036e-05, 5.7512e-05, 6.7141e-05, 6.3737e-05, + 6.1851e-05, 7.9771e-05], device='cuda:3') +2023-04-28 00:14:39,768 INFO [finetune.py:976] (3/7) Epoch 26, batch 5050, loss[loss=0.1773, simple_loss=0.2474, pruned_loss=0.05364, over 4815.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2394, pruned_loss=0.04647, over 953011.38 frames. ], batch size: 41, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:14:56,595 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148267.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:14:59,523 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.057e+01 1.422e+02 1.750e+02 2.059e+02 5.482e+02, threshold=3.501e+02, percent-clipped=2.0 +2023-04-28 00:15:08,807 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148287.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:15:11,736 INFO [finetune.py:976] (3/7) Epoch 26, batch 5100, loss[loss=0.1712, simple_loss=0.231, pruned_loss=0.05574, over 4818.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.237, pruned_loss=0.04603, over 954474.91 frames. ], batch size: 38, lr: 2.95e-03, grad_scale: 64.0 +2023-04-28 00:15:28,678 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148315.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:15:45,037 INFO [finetune.py:976] (3/7) Epoch 26, batch 5150, loss[loss=0.1773, simple_loss=0.2556, pruned_loss=0.04955, over 4745.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2371, pruned_loss=0.04631, over 955900.69 frames. ], batch size: 54, lr: 2.95e-03, grad_scale: 64.0 +2023-04-28 00:15:49,291 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148348.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:16:00,256 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148363.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:16:06,186 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.267e+01 1.646e+02 1.916e+02 2.227e+02 4.429e+02, threshold=3.832e+02, percent-clipped=2.0 +2023-04-28 00:16:18,411 INFO [finetune.py:976] (3/7) Epoch 26, batch 5200, loss[loss=0.1885, simple_loss=0.2622, pruned_loss=0.05739, over 4931.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.242, pruned_loss=0.04772, over 956789.21 frames. ], batch size: 36, lr: 2.95e-03, grad_scale: 64.0 +2023-04-28 00:16:20,941 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6963, 1.3907, 1.3046, 1.5008, 1.8810, 1.5197, 1.3516, 1.2290], + device='cuda:3'), covar=tensor([0.1592, 0.1299, 0.1729, 0.1107, 0.0779, 0.1652, 0.1822, 0.2226], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0311, 0.0354, 0.0288, 0.0329, 0.0307, 0.0301, 0.0378], + device='cuda:3'), out_proj_covar=tensor([6.4717e-05, 6.4015e-05, 7.4381e-05, 5.7673e-05, 6.7301e-05, 6.4032e-05, + 6.2207e-05, 8.0131e-05], device='cuda:3') +2023-04-28 00:16:40,379 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1621, 1.6275, 2.0120, 2.5572, 2.1502, 1.6818, 1.6607, 1.8890], + device='cuda:3'), covar=tensor([0.2779, 0.3564, 0.1731, 0.2273, 0.2458, 0.2522, 0.4166, 0.2197], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0245, 0.0227, 0.0313, 0.0221, 0.0235, 0.0228, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 00:16:44,020 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148429.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:16:51,852 INFO [finetune.py:976] (3/7) Epoch 26, batch 5250, loss[loss=0.2039, simple_loss=0.2803, pruned_loss=0.06379, over 4823.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2427, pruned_loss=0.04796, over 956979.77 frames. ], batch size: 39, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:17:02,618 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2669, 1.5311, 1.3639, 1.4824, 1.2692, 1.2586, 1.2356, 1.0366], + device='cuda:3'), covar=tensor([0.1777, 0.1325, 0.0944, 0.1313, 0.3870, 0.1311, 0.1965, 0.2383], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0300, 0.0216, 0.0277, 0.0315, 0.0254, 0.0250, 0.0264], + device='cuda:3'), out_proj_covar=tensor([1.1391e-04, 1.1795e-04, 8.4854e-05, 1.0909e-04, 1.2684e-04, 1.0017e-04, + 1.0086e-04, 1.0406e-04], device='cuda:3') +2023-04-28 00:17:12,108 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148471.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:17:12,616 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.667e+02 1.923e+02 2.369e+02 3.667e+02, threshold=3.847e+02, percent-clipped=0.0 +2023-04-28 00:17:13,383 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5615, 1.9412, 2.0222, 2.0630, 1.9572, 1.9477, 2.0843, 2.0366], + device='cuda:3'), covar=tensor([0.3899, 0.5325, 0.4361, 0.4462, 0.5187, 0.7007, 0.5150, 0.4561], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0374, 0.0328, 0.0339, 0.0350, 0.0394, 0.0359, 0.0332], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 00:17:23,730 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148490.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:17:24,853 INFO [finetune.py:976] (3/7) Epoch 26, batch 5300, loss[loss=0.1748, simple_loss=0.25, pruned_loss=0.04977, over 4808.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2435, pruned_loss=0.04787, over 954383.06 frames. ], batch size: 40, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:17:28,554 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148498.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:17:32,244 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7550, 1.3854, 1.3000, 1.0623, 1.3527, 1.1880, 1.6272, 1.2307], + device='cuda:3'), covar=tensor([0.2760, 0.1555, 0.3808, 0.2220, 0.1335, 0.1804, 0.1518, 0.3820], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0354, 0.0425, 0.0349, 0.0382, 0.0375, 0.0365, 0.0423], + device='cuda:3'), out_proj_covar=tensor([9.9875e-05, 1.0564e-04, 1.2875e-04, 1.0479e-04, 1.1340e-04, 1.1140e-04, + 1.0659e-04, 1.2711e-04], device='cuda:3') +2023-04-28 00:17:52,364 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148513.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:17:54,331 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-28 00:17:56,053 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148519.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:17:57,349 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7120, 1.0606, 1.2844, 1.3445, 1.7949, 1.3858, 1.2412, 1.2096], + device='cuda:3'), covar=tensor([0.1699, 0.1840, 0.2228, 0.1593, 0.1129, 0.1866, 0.1991, 0.2621], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0312, 0.0354, 0.0288, 0.0329, 0.0307, 0.0301, 0.0377], + device='cuda:3'), out_proj_covar=tensor([6.4748e-05, 6.4081e-05, 7.4337e-05, 5.7771e-05, 6.7224e-05, 6.4023e-05, + 6.2296e-05, 7.9878e-05], device='cuda:3') +2023-04-28 00:18:24,825 INFO [finetune.py:976] (3/7) Epoch 26, batch 5350, loss[loss=0.1936, simple_loss=0.2441, pruned_loss=0.07154, over 4389.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2446, pruned_loss=0.04821, over 954791.33 frames. ], batch size: 19, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:18:46,001 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148559.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:18:47,670 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148561.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:18:48,330 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2563, 2.1038, 2.3605, 2.5226, 2.7216, 2.2043, 1.8662, 2.3822], + device='cuda:3'), covar=tensor([0.0744, 0.1008, 0.0579, 0.0640, 0.0560, 0.0800, 0.0794, 0.0522], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0200, 0.0183, 0.0170, 0.0176, 0.0175, 0.0151, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:18:56,338 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148567.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:18:59,259 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.519e+02 1.727e+02 2.044e+02 3.466e+02, threshold=3.453e+02, percent-clipped=0.0 +2023-04-28 00:19:22,270 INFO [finetune.py:976] (3/7) Epoch 26, batch 5400, loss[loss=0.1229, simple_loss=0.1954, pruned_loss=0.02526, over 4819.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2413, pruned_loss=0.04742, over 953696.44 frames. ], batch size: 25, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:19:43,240 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2788, 1.6486, 1.4492, 1.8233, 1.7318, 1.9063, 1.5088, 3.5729], + device='cuda:3'), covar=tensor([0.0557, 0.0731, 0.0725, 0.1102, 0.0563, 0.0465, 0.0666, 0.0124], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 00:19:54,124 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148615.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:20:11,715 INFO [finetune.py:976] (3/7) Epoch 26, batch 5450, loss[loss=0.154, simple_loss=0.2202, pruned_loss=0.04388, over 4829.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2385, pruned_loss=0.04696, over 953837.30 frames. ], batch size: 25, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:20:12,386 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148643.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:20:17,244 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148651.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:20:22,789 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3939, 1.8974, 2.3086, 2.8176, 2.2726, 1.7730, 1.6008, 2.0884], + device='cuda:3'), covar=tensor([0.2998, 0.3008, 0.1549, 0.2140, 0.2595, 0.2543, 0.3926, 0.1915], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0244, 0.0227, 0.0313, 0.0220, 0.0234, 0.0227, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 00:20:31,291 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.574e+02 1.926e+02 2.349e+02 4.613e+02, threshold=3.853e+02, percent-clipped=6.0 +2023-04-28 00:20:45,355 INFO [finetune.py:976] (3/7) Epoch 26, batch 5500, loss[loss=0.1517, simple_loss=0.2198, pruned_loss=0.04183, over 4158.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2351, pruned_loss=0.04587, over 953315.40 frames. ], batch size: 18, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:20:57,698 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148712.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:00,132 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148716.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:18,756 INFO [finetune.py:976] (3/7) Epoch 26, batch 5550, loss[loss=0.2019, simple_loss=0.2801, pruned_loss=0.06184, over 4850.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2363, pruned_loss=0.04668, over 952410.12 frames. ], batch size: 47, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:21:26,263 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9560, 2.2949, 2.2002, 2.3117, 2.1568, 2.1687, 2.3288, 2.1997], + device='cuda:3'), covar=tensor([0.3808, 0.5437, 0.4532, 0.4105, 0.5429, 0.6498, 0.5465, 0.5320], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0375, 0.0330, 0.0340, 0.0351, 0.0396, 0.0361, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 00:21:38,040 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.670e+01 1.487e+02 1.767e+02 2.141e+02 3.989e+02, threshold=3.535e+02, percent-clipped=1.0 +2023-04-28 00:21:41,044 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148777.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:45,567 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148785.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:49,600 INFO [finetune.py:976] (3/7) Epoch 26, batch 5600, loss[loss=0.1741, simple_loss=0.2466, pruned_loss=0.05081, over 4816.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2401, pruned_loss=0.0473, over 950538.15 frames. ], batch size: 33, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:22:19,037 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7612, 1.2306, 1.8409, 2.1757, 1.7893, 1.7480, 1.7939, 1.7410], + device='cuda:3'), covar=tensor([0.4178, 0.6240, 0.5826, 0.5374, 0.5772, 0.7056, 0.7213, 0.8314], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0421, 0.0515, 0.0504, 0.0468, 0.0506, 0.0507, 0.0518], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:22:20,339 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-28 00:22:20,666 INFO [finetune.py:976] (3/7) Epoch 26, batch 5650, loss[loss=0.1479, simple_loss=0.2249, pruned_loss=0.03548, over 4815.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2422, pruned_loss=0.04724, over 949598.86 frames. ], batch size: 25, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:22:27,775 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148854.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:22:36,232 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0275, 1.5106, 1.6945, 1.7758, 2.1389, 1.7415, 1.5839, 1.6044], + device='cuda:3'), covar=tensor([0.1536, 0.1545, 0.1869, 0.1386, 0.0891, 0.1606, 0.1661, 0.2283], + device='cuda:3'), in_proj_covar=tensor([0.0318, 0.0312, 0.0356, 0.0290, 0.0330, 0.0309, 0.0303, 0.0379], + device='cuda:3'), out_proj_covar=tensor([6.5019e-05, 6.4133e-05, 7.4649e-05, 5.8123e-05, 6.7471e-05, 6.4408e-05, + 6.2650e-05, 8.0190e-05], device='cuda:3') +2023-04-28 00:22:39,028 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.517e+02 1.815e+02 2.254e+02 4.751e+02, threshold=3.630e+02, percent-clipped=7.0 +2023-04-28 00:22:50,265 INFO [finetune.py:976] (3/7) Epoch 26, batch 5700, loss[loss=0.1767, simple_loss=0.2466, pruned_loss=0.05342, over 4035.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2373, pruned_loss=0.04655, over 931043.37 frames. ], batch size: 17, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:23:00,538 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3391, 1.4730, 1.3992, 1.7873, 1.5736, 1.7013, 1.4062, 2.8773], + device='cuda:3'), covar=tensor([0.0595, 0.0787, 0.0804, 0.1085, 0.0666, 0.0490, 0.0706, 0.0207], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 00:23:20,150 INFO [finetune.py:976] (3/7) Epoch 27, batch 0, loss[loss=0.1707, simple_loss=0.2444, pruned_loss=0.04851, over 4855.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2444, pruned_loss=0.04851, over 4855.00 frames. ], batch size: 31, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:23:20,150 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 00:23:31,642 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1252, 2.5490, 1.1058, 1.4283, 2.0252, 1.2609, 2.9999, 1.7145], + device='cuda:3'), covar=tensor([0.0604, 0.0575, 0.0650, 0.1096, 0.0393, 0.0888, 0.0232, 0.0554], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0050, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 00:23:32,134 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4447, 1.3691, 1.6710, 1.6826, 1.3323, 1.2885, 1.4002, 0.8222], + device='cuda:3'), covar=tensor([0.0518, 0.0539, 0.0431, 0.0495, 0.0805, 0.1133, 0.0514, 0.0559], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:23:41,724 INFO [finetune.py:1010] (3/7) Epoch 27, validation: loss=0.1548, simple_loss=0.2237, pruned_loss=0.04298, over 2265189.00 frames. +2023-04-28 00:23:41,724 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 00:24:11,145 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148943.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:24:41,616 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-28 00:24:43,144 INFO [finetune.py:976] (3/7) Epoch 27, batch 50, loss[loss=0.1651, simple_loss=0.2367, pruned_loss=0.04671, over 4783.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2434, pruned_loss=0.0472, over 216999.82 frames. ], batch size: 29, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:24:45,492 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.880e+01 1.482e+02 1.782e+02 2.132e+02 4.593e+02, threshold=3.564e+02, percent-clipped=2.0 +2023-04-28 00:24:53,727 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4041, 1.2836, 4.1037, 3.8905, 3.6332, 3.9541, 3.8737, 3.5890], + device='cuda:3'), covar=tensor([0.7492, 0.5837, 0.1101, 0.1744, 0.1154, 0.1593, 0.1621, 0.1745], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0310, 0.0410, 0.0411, 0.0349, 0.0415, 0.0320, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:25:07,487 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148991.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:25:28,981 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149007.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:25:37,829 INFO [finetune.py:976] (3/7) Epoch 27, batch 100, loss[loss=0.1618, simple_loss=0.2277, pruned_loss=0.04791, over 4831.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2416, pruned_loss=0.04843, over 382795.89 frames. ], batch size: 38, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:25:48,517 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5286, 2.2893, 2.6912, 2.8060, 2.9102, 2.3252, 2.2197, 2.7663], + device='cuda:3'), covar=tensor([0.0773, 0.0970, 0.0573, 0.0533, 0.0572, 0.0753, 0.0649, 0.0495], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0200, 0.0182, 0.0169, 0.0176, 0.0175, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:26:04,852 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7428, 1.4800, 4.5663, 4.3009, 3.9763, 4.3132, 4.2292, 4.1162], + device='cuda:3'), covar=tensor([0.6584, 0.5550, 0.1016, 0.1553, 0.1026, 0.1651, 0.1256, 0.1303], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0310, 0.0409, 0.0410, 0.0348, 0.0414, 0.0320, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:26:11,901 INFO [finetune.py:976] (3/7) Epoch 27, batch 150, loss[loss=0.1404, simple_loss=0.2141, pruned_loss=0.03334, over 4821.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2366, pruned_loss=0.04703, over 509845.14 frames. ], batch size: 39, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:26:13,183 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149072.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:26:13,727 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.439e+02 1.708e+02 1.990e+02 3.271e+02, threshold=3.417e+02, percent-clipped=0.0 +2023-04-28 00:26:22,647 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149085.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:26:27,974 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 00:26:45,180 INFO [finetune.py:976] (3/7) Epoch 27, batch 200, loss[loss=0.177, simple_loss=0.2494, pruned_loss=0.05227, over 4916.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2335, pruned_loss=0.04561, over 610264.92 frames. ], batch size: 37, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:26:45,420 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-28 00:26:55,612 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149133.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:27:08,548 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149154.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:27:18,795 INFO [finetune.py:976] (3/7) Epoch 27, batch 250, loss[loss=0.1361, simple_loss=0.2011, pruned_loss=0.03557, over 4523.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2372, pruned_loss=0.04723, over 686674.23 frames. ], batch size: 19, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:27:21,648 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.484e+02 1.804e+02 2.162e+02 4.404e+02, threshold=3.609e+02, percent-clipped=2.0 +2023-04-28 00:27:41,049 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149202.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:27:52,340 INFO [finetune.py:976] (3/7) Epoch 27, batch 300, loss[loss=0.1529, simple_loss=0.2351, pruned_loss=0.03532, over 4902.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2432, pruned_loss=0.04884, over 748449.83 frames. ], batch size: 32, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:28:25,827 INFO [finetune.py:976] (3/7) Epoch 27, batch 350, loss[loss=0.191, simple_loss=0.2655, pruned_loss=0.05831, over 4749.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2448, pruned_loss=0.04921, over 794236.25 frames. ], batch size: 27, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:28:28,122 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.612e+02 1.887e+02 2.251e+02 5.949e+02, threshold=3.774e+02, percent-clipped=2.0 +2023-04-28 00:28:38,469 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-04-28 00:28:51,989 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149307.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:28:59,853 INFO [finetune.py:976] (3/7) Epoch 27, batch 400, loss[loss=0.1561, simple_loss=0.2326, pruned_loss=0.0398, over 4762.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2452, pruned_loss=0.04935, over 829594.38 frames. ], batch size: 26, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:29:39,549 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149355.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:29:48,682 INFO [finetune.py:976] (3/7) Epoch 27, batch 450, loss[loss=0.1408, simple_loss=0.2053, pruned_loss=0.03817, over 4822.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2426, pruned_loss=0.04808, over 857944.85 frames. ], batch size: 25, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:29:50,017 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149372.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:29:50,984 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.462e+01 1.496e+02 1.752e+02 2.050e+02 4.938e+02, threshold=3.505e+02, percent-clipped=1.0 +2023-04-28 00:29:55,960 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8291, 3.8002, 2.9080, 4.4338, 3.7741, 3.8229, 1.6341, 3.7335], + device='cuda:3'), covar=tensor([0.1657, 0.1277, 0.3243, 0.1471, 0.3182, 0.1893, 0.5647, 0.2550], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0222, 0.0254, 0.0305, 0.0302, 0.0250, 0.0276, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 00:30:41,613 INFO [finetune.py:976] (3/7) Epoch 27, batch 500, loss[loss=0.1651, simple_loss=0.2339, pruned_loss=0.0482, over 4822.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2388, pruned_loss=0.04661, over 878594.15 frames. ], batch size: 41, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:30:41,677 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149420.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:31:43,468 INFO [finetune.py:976] (3/7) Epoch 27, batch 550, loss[loss=0.1659, simple_loss=0.2394, pruned_loss=0.04617, over 4774.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2366, pruned_loss=0.04605, over 895689.23 frames. ], batch size: 26, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:31:45,298 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.482e+02 1.685e+02 2.089e+02 3.058e+02, threshold=3.371e+02, percent-clipped=0.0 +2023-04-28 00:32:06,929 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0201, 1.3026, 4.8882, 4.5742, 4.2595, 4.6414, 4.3495, 4.2987], + device='cuda:3'), covar=tensor([0.7136, 0.6052, 0.0958, 0.1820, 0.1096, 0.1383, 0.1855, 0.1517], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0310, 0.0410, 0.0411, 0.0348, 0.0415, 0.0319, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:32:08,160 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3001, 1.4753, 1.3722, 1.7223, 1.5754, 1.9786, 1.4489, 3.5549], + device='cuda:3'), covar=tensor([0.0560, 0.0843, 0.0810, 0.1189, 0.0643, 0.0518, 0.0732, 0.0127], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 00:32:13,107 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149514.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:32:16,577 INFO [finetune.py:976] (3/7) Epoch 27, batch 600, loss[loss=0.1169, simple_loss=0.1957, pruned_loss=0.01903, over 4768.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2376, pruned_loss=0.04647, over 908331.93 frames. ], batch size: 27, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:32:47,510 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149565.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:32:50,406 INFO [finetune.py:976] (3/7) Epoch 27, batch 650, loss[loss=0.2039, simple_loss=0.2704, pruned_loss=0.06876, over 4904.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2406, pruned_loss=0.04727, over 918562.92 frames. ], batch size: 37, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:32:52,258 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.546e+02 1.916e+02 2.340e+02 4.465e+02, threshold=3.832e+02, percent-clipped=5.0 +2023-04-28 00:32:53,614 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149575.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:33:23,679 INFO [finetune.py:976] (3/7) Epoch 27, batch 700, loss[loss=0.1446, simple_loss=0.2211, pruned_loss=0.03402, over 4853.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2409, pruned_loss=0.04615, over 927748.19 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:33:27,479 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149626.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 00:33:56,990 INFO [finetune.py:976] (3/7) Epoch 27, batch 750, loss[loss=0.1482, simple_loss=0.2152, pruned_loss=0.04063, over 4737.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2406, pruned_loss=0.04587, over 932511.52 frames. ], batch size: 23, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:33:58,787 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.603e+02 1.862e+02 2.208e+02 4.099e+02, threshold=3.723e+02, percent-clipped=2.0 +2023-04-28 00:34:42,995 INFO [finetune.py:976] (3/7) Epoch 27, batch 800, loss[loss=0.1908, simple_loss=0.2586, pruned_loss=0.06149, over 4810.00 frames. ], tot_loss[loss=0.168, simple_loss=0.242, pruned_loss=0.04706, over 937244.67 frames. ], batch size: 39, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:35:14,861 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149746.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:35:32,504 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6904, 1.7078, 0.7541, 1.3051, 1.9570, 1.5553, 1.4124, 1.5264], + device='cuda:3'), covar=tensor([0.0508, 0.0384, 0.0348, 0.0572, 0.0251, 0.0509, 0.0549, 0.0589], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 00:35:34,225 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5795, 1.5163, 1.6931, 1.9301, 1.9861, 1.5620, 1.3930, 1.8568], + device='cuda:3'), covar=tensor([0.0816, 0.1183, 0.0815, 0.0538, 0.0580, 0.0776, 0.0711, 0.0521], + device='cuda:3'), in_proj_covar=tensor([0.0182, 0.0199, 0.0182, 0.0168, 0.0175, 0.0175, 0.0150, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:35:36,722 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0417, 1.7169, 2.2724, 2.3825, 2.1172, 1.9839, 2.0708, 2.0294], + device='cuda:3'), covar=tensor([0.4894, 0.7561, 0.6835, 0.5900, 0.6218, 0.9200, 0.9117, 1.0035], + device='cuda:3'), in_proj_covar=tensor([0.0444, 0.0423, 0.0518, 0.0508, 0.0471, 0.0508, 0.0510, 0.0523], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:35:48,631 INFO [finetune.py:976] (3/7) Epoch 27, batch 850, loss[loss=0.1475, simple_loss=0.2241, pruned_loss=0.03545, over 4819.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2408, pruned_loss=0.04732, over 940992.59 frames. ], batch size: 38, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:35:55,555 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.475e+02 1.749e+02 2.076e+02 4.110e+02, threshold=3.498e+02, percent-clipped=2.0 +2023-04-28 00:36:06,636 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5662, 2.3256, 2.5937, 2.9678, 2.8818, 2.3834, 2.3299, 2.7902], + device='cuda:3'), covar=tensor([0.0817, 0.0939, 0.0618, 0.0515, 0.0552, 0.0796, 0.0594, 0.0485], + device='cuda:3'), in_proj_covar=tensor([0.0182, 0.0198, 0.0181, 0.0167, 0.0175, 0.0174, 0.0150, 0.0175], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:36:23,787 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1605, 1.3807, 4.3702, 3.9084, 3.8760, 3.7963, 3.7565, 3.7018], + device='cuda:3'), covar=tensor([0.9295, 0.7712, 0.1400, 0.2756, 0.1850, 0.3343, 0.7333, 0.2802], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0310, 0.0411, 0.0410, 0.0349, 0.0416, 0.0321, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:36:24,413 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149807.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 00:36:33,260 INFO [finetune.py:976] (3/7) Epoch 27, batch 900, loss[loss=0.1454, simple_loss=0.2242, pruned_loss=0.03334, over 4789.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2381, pruned_loss=0.04649, over 944503.51 frames. ], batch size: 29, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:36:41,325 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-04-28 00:36:42,489 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5151, 2.3681, 2.6684, 2.9827, 2.7976, 2.4227, 2.2967, 2.7776], + device='cuda:3'), covar=tensor([0.0902, 0.0963, 0.0615, 0.0598, 0.0682, 0.0834, 0.0684, 0.0569], + device='cuda:3'), in_proj_covar=tensor([0.0183, 0.0199, 0.0182, 0.0168, 0.0175, 0.0175, 0.0151, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:37:06,214 INFO [finetune.py:976] (3/7) Epoch 27, batch 950, loss[loss=0.2377, simple_loss=0.3008, pruned_loss=0.08734, over 4913.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2365, pruned_loss=0.04631, over 947098.59 frames. ], batch size: 37, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:37:06,275 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149870.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:37:08,535 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.514e+02 1.776e+02 2.110e+02 4.018e+02, threshold=3.552e+02, percent-clipped=2.0 +2023-04-28 00:37:37,880 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-04-28 00:37:40,041 INFO [finetune.py:976] (3/7) Epoch 27, batch 1000, loss[loss=0.16, simple_loss=0.2137, pruned_loss=0.05311, over 4259.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2391, pruned_loss=0.04715, over 950115.17 frames. ], batch size: 18, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:37:40,686 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149921.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 00:37:51,045 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 00:38:13,688 INFO [finetune.py:976] (3/7) Epoch 27, batch 1050, loss[loss=0.2295, simple_loss=0.2903, pruned_loss=0.08435, over 4866.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2403, pruned_loss=0.04684, over 952263.24 frames. ], batch size: 44, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:38:15,491 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.212e+02 1.681e+02 2.029e+02 2.486e+02 4.683e+02, threshold=4.057e+02, percent-clipped=3.0 +2023-04-28 00:38:26,900 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-28 00:38:44,646 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-28 00:38:48,335 INFO [finetune.py:976] (3/7) Epoch 27, batch 1100, loss[loss=0.2061, simple_loss=0.2696, pruned_loss=0.07134, over 4803.00 frames. ], tot_loss[loss=0.167, simple_loss=0.241, pruned_loss=0.0465, over 955408.95 frames. ], batch size: 40, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:38:59,847 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-28 00:39:22,077 INFO [finetune.py:976] (3/7) Epoch 27, batch 1150, loss[loss=0.1917, simple_loss=0.2628, pruned_loss=0.0603, over 4736.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2423, pruned_loss=0.04696, over 955891.24 frames. ], batch size: 54, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:39:23,886 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.471e+02 1.817e+02 2.070e+02 3.709e+02, threshold=3.635e+02, percent-clipped=0.0 +2023-04-28 00:39:28,093 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150079.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:39:43,680 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150102.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 00:39:56,128 INFO [finetune.py:976] (3/7) Epoch 27, batch 1200, loss[loss=0.1744, simple_loss=0.2348, pruned_loss=0.05696, over 4709.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2411, pruned_loss=0.04675, over 955833.27 frames. ], batch size: 23, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:40:10,320 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150140.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:40:45,076 INFO [finetune.py:976] (3/7) Epoch 27, batch 1250, loss[loss=0.1827, simple_loss=0.2512, pruned_loss=0.0571, over 4829.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2389, pruned_loss=0.04624, over 956016.48 frames. ], batch size: 39, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:40:45,191 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150170.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:40:47,940 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.413e+01 1.487e+02 1.765e+02 2.046e+02 3.394e+02, threshold=3.529e+02, percent-clipped=0.0 +2023-04-28 00:41:41,402 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5463, 1.9549, 1.9365, 2.0269, 1.9418, 1.9626, 1.9517, 1.9122], + device='cuda:3'), covar=tensor([0.4000, 0.5459, 0.4428, 0.4093, 0.5328, 0.6696, 0.5392, 0.4768], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0373, 0.0329, 0.0339, 0.0349, 0.0392, 0.0358, 0.0330], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 00:41:49,586 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150218.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:41:50,798 INFO [finetune.py:976] (3/7) Epoch 27, batch 1300, loss[loss=0.1386, simple_loss=0.2156, pruned_loss=0.03084, over 4858.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2363, pruned_loss=0.04506, over 956169.08 frames. ], batch size: 31, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:41:51,514 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150221.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:42:56,733 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150269.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:42:57,284 INFO [finetune.py:976] (3/7) Epoch 27, batch 1350, loss[loss=0.1258, simple_loss=0.1987, pruned_loss=0.02645, over 4763.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2357, pruned_loss=0.04454, over 957306.52 frames. ], batch size: 26, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:42:59,132 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.474e+02 1.801e+02 2.070e+02 3.416e+02, threshold=3.602e+02, percent-clipped=0.0 +2023-04-28 00:43:05,182 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150274.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:43:19,078 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4542, 1.4461, 1.7732, 1.8029, 1.3811, 1.2136, 1.5205, 1.0025], + device='cuda:3'), covar=tensor([0.0550, 0.0530, 0.0365, 0.0513, 0.0684, 0.1056, 0.0542, 0.0615], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0074, 0.0094, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:44:01,100 INFO [finetune.py:976] (3/7) Epoch 27, batch 1400, loss[loss=0.1765, simple_loss=0.2602, pruned_loss=0.04637, over 4792.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2386, pruned_loss=0.04554, over 955896.67 frames. ], batch size: 54, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:44:22,624 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150335.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:45:00,624 INFO [finetune.py:976] (3/7) Epoch 27, batch 1450, loss[loss=0.1598, simple_loss=0.2349, pruned_loss=0.04235, over 4796.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2404, pruned_loss=0.04567, over 956471.53 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:45:03,059 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.594e+02 1.851e+02 2.331e+02 4.105e+02, threshold=3.701e+02, percent-clipped=2.0 +2023-04-28 00:45:22,904 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150402.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:45:26,590 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9536, 1.7548, 2.1747, 2.3480, 2.0038, 1.8895, 1.9840, 1.9805], + device='cuda:3'), covar=tensor([0.4523, 0.6783, 0.7109, 0.5497, 0.6118, 0.9012, 0.9148, 1.0384], + device='cuda:3'), in_proj_covar=tensor([0.0441, 0.0421, 0.0515, 0.0505, 0.0469, 0.0506, 0.0507, 0.0520], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:45:29,025 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1840, 1.8421, 2.3408, 2.5804, 2.1966, 2.0890, 2.1960, 2.1968], + device='cuda:3'), covar=tensor([0.4808, 0.7810, 0.7263, 0.5613, 0.6258, 0.9023, 0.9367, 0.9677], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0422, 0.0516, 0.0506, 0.0469, 0.0506, 0.0507, 0.0520], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:45:33,783 INFO [finetune.py:976] (3/7) Epoch 27, batch 1500, loss[loss=0.1509, simple_loss=0.2225, pruned_loss=0.03969, over 4671.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2399, pruned_loss=0.04521, over 955288.24 frames. ], batch size: 23, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:45:44,392 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150435.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:45:54,543 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150450.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:46:06,958 INFO [finetune.py:976] (3/7) Epoch 27, batch 1550, loss[loss=0.1466, simple_loss=0.2327, pruned_loss=0.0303, over 4785.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2414, pruned_loss=0.04602, over 955456.36 frames. ], batch size: 29, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:46:09,361 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.486e+02 1.780e+02 2.149e+02 3.630e+02, threshold=3.561e+02, percent-clipped=0.0 +2023-04-28 00:46:19,678 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-28 00:46:22,365 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3214, 1.5431, 1.4847, 1.8872, 1.7220, 1.8894, 1.4752, 3.6419], + device='cuda:3'), covar=tensor([0.0675, 0.0943, 0.0934, 0.1340, 0.0728, 0.0567, 0.0879, 0.0185], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 00:46:40,774 INFO [finetune.py:976] (3/7) Epoch 27, batch 1600, loss[loss=0.1544, simple_loss=0.2171, pruned_loss=0.04584, over 4826.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2392, pruned_loss=0.04573, over 956113.15 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:46:59,330 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1081, 0.7951, 0.9007, 0.7542, 1.2454, 0.9818, 0.9410, 0.9507], + device='cuda:3'), covar=tensor([0.1731, 0.1375, 0.1724, 0.1643, 0.0999, 0.1343, 0.1562, 0.2097], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0306, 0.0348, 0.0285, 0.0325, 0.0305, 0.0297, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.3604e-05, 6.2921e-05, 7.3116e-05, 5.6946e-05, 6.6419e-05, 6.3567e-05, + 6.1442e-05, 7.8719e-05], device='cuda:3') +2023-04-28 00:47:01,171 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7093, 1.7158, 0.6950, 1.3591, 1.7810, 1.5595, 1.4452, 1.5246], + device='cuda:3'), covar=tensor([0.0472, 0.0380, 0.0348, 0.0553, 0.0275, 0.0518, 0.0517, 0.0574], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 00:47:08,133 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150537.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:47:41,253 INFO [finetune.py:976] (3/7) Epoch 27, batch 1650, loss[loss=0.1646, simple_loss=0.2341, pruned_loss=0.04757, over 4820.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2376, pruned_loss=0.04571, over 953912.35 frames. ], batch size: 33, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:47:41,386 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2217, 3.0012, 2.7185, 2.7612, 2.2095, 2.4667, 2.7224, 1.9942], + device='cuda:3'), covar=tensor([0.2300, 0.1242, 0.0694, 0.1317, 0.2866, 0.1161, 0.2001, 0.2659], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0302, 0.0217, 0.0279, 0.0316, 0.0255, 0.0251, 0.0266], + device='cuda:3'), out_proj_covar=tensor([1.1432e-04, 1.1903e-04, 8.5539e-05, 1.0992e-04, 1.2739e-04, 1.0027e-04, + 1.0115e-04, 1.0506e-04], device='cuda:3') +2023-04-28 00:47:43,700 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.594e+02 1.886e+02 2.369e+02 4.479e+02, threshold=3.773e+02, percent-clipped=3.0 +2023-04-28 00:48:00,527 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150598.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:48:14,790 INFO [finetune.py:976] (3/7) Epoch 27, batch 1700, loss[loss=0.1679, simple_loss=0.24, pruned_loss=0.04789, over 4808.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2369, pruned_loss=0.04586, over 955278.17 frames. ], batch size: 45, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:48:20,948 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150630.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:48:36,003 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 00:48:48,160 INFO [finetune.py:976] (3/7) Epoch 27, batch 1750, loss[loss=0.2213, simple_loss=0.2912, pruned_loss=0.07573, over 4923.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2379, pruned_loss=0.04631, over 956170.67 frames. ], batch size: 38, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:48:50,603 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.571e+02 1.900e+02 2.319e+02 4.181e+02, threshold=3.799e+02, percent-clipped=4.0 +2023-04-28 00:49:14,390 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 00:49:21,909 INFO [finetune.py:976] (3/7) Epoch 27, batch 1800, loss[loss=0.2052, simple_loss=0.2699, pruned_loss=0.07027, over 4918.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.241, pruned_loss=0.04729, over 955457.79 frames. ], batch size: 36, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:49:25,605 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6528, 1.2262, 4.4105, 4.1234, 3.8303, 4.2130, 4.0673, 3.9598], + device='cuda:3'), covar=tensor([0.7412, 0.6283, 0.1143, 0.1683, 0.1186, 0.1721, 0.1612, 0.1564], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0310, 0.0413, 0.0412, 0.0351, 0.0418, 0.0321, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:49:29,206 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150732.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:49:31,007 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150735.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:49:45,681 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8974, 0.9028, 1.0320, 1.0267, 0.8637, 0.8017, 0.8201, 0.4833], + device='cuda:3'), covar=tensor([0.0495, 0.0435, 0.0467, 0.0425, 0.0630, 0.0935, 0.0499, 0.0587], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0066, 0.0065, 0.0068, 0.0074, 0.0094, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:49:54,541 INFO [finetune.py:976] (3/7) Epoch 27, batch 1850, loss[loss=0.1742, simple_loss=0.242, pruned_loss=0.05326, over 4747.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2434, pruned_loss=0.04805, over 955633.72 frames. ], batch size: 28, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:50:02,278 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.541e+02 1.934e+02 2.290e+02 4.009e+02, threshold=3.867e+02, percent-clipped=2.0 +2023-04-28 00:50:13,364 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150783.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:50:25,728 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150793.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 00:50:49,055 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7955, 3.5812, 1.0568, 2.0889, 2.0159, 2.5228, 2.1895, 1.1947], + device='cuda:3'), covar=tensor([0.1220, 0.0851, 0.1850, 0.1167, 0.1076, 0.1030, 0.1355, 0.2206], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0121, 0.0131, 0.0152, 0.0118, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:50:56,513 INFO [finetune.py:976] (3/7) Epoch 27, batch 1900, loss[loss=0.1909, simple_loss=0.2905, pruned_loss=0.04562, over 4880.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2454, pruned_loss=0.04856, over 954980.81 frames. ], batch size: 43, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:50:57,607 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-28 00:51:26,379 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8153, 1.3986, 1.9904, 2.3303, 1.9536, 1.8522, 1.8983, 1.8771], + device='cuda:3'), covar=tensor([0.4325, 0.6375, 0.5740, 0.5045, 0.5232, 0.6981, 0.7450, 0.8532], + device='cuda:3'), in_proj_covar=tensor([0.0443, 0.0421, 0.0518, 0.0506, 0.0469, 0.0506, 0.0507, 0.0522], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:51:29,793 INFO [finetune.py:976] (3/7) Epoch 27, batch 1950, loss[loss=0.151, simple_loss=0.2332, pruned_loss=0.03441, over 4848.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2432, pruned_loss=0.04775, over 954168.23 frames. ], batch size: 44, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:51:32,194 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.535e+02 1.747e+02 2.093e+02 5.445e+02, threshold=3.494e+02, percent-clipped=3.0 +2023-04-28 00:51:44,420 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150893.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:51:49,236 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150900.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:51:54,029 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-04-28 00:52:03,055 INFO [finetune.py:976] (3/7) Epoch 27, batch 2000, loss[loss=0.1999, simple_loss=0.2632, pruned_loss=0.06828, over 4918.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2415, pruned_loss=0.04775, over 955897.99 frames. ], batch size: 43, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:52:07,360 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7920, 1.2357, 4.2429, 3.9707, 3.6853, 3.9810, 3.9452, 3.7930], + device='cuda:3'), covar=tensor([0.6847, 0.6017, 0.1064, 0.1726, 0.1088, 0.1744, 0.2115, 0.1363], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0309, 0.0411, 0.0410, 0.0349, 0.0416, 0.0320, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:52:09,165 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150930.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:52:51,921 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150961.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:53:02,952 INFO [finetune.py:976] (3/7) Epoch 27, batch 2050, loss[loss=0.1278, simple_loss=0.1977, pruned_loss=0.02896, over 4805.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2375, pruned_loss=0.04654, over 954961.79 frames. ], batch size: 25, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:53:05,397 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.536e+02 1.853e+02 2.161e+02 4.480e+02, threshold=3.705e+02, percent-clipped=3.0 +2023-04-28 00:53:12,615 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150978.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:53:27,874 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4248, 1.2723, 1.5674, 1.6262, 1.3327, 1.2167, 1.3523, 0.8744], + device='cuda:3'), covar=tensor([0.0458, 0.0663, 0.0400, 0.0463, 0.0631, 0.1086, 0.0456, 0.0566], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0066, 0.0065, 0.0068, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:53:49,033 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-28 00:54:07,734 INFO [finetune.py:976] (3/7) Epoch 27, batch 2100, loss[loss=0.1691, simple_loss=0.2459, pruned_loss=0.04619, over 4820.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2373, pruned_loss=0.04628, over 957257.62 frames. ], batch size: 45, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:55:11,727 INFO [finetune.py:976] (3/7) Epoch 27, batch 2150, loss[loss=0.1757, simple_loss=0.2546, pruned_loss=0.04839, over 4826.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2389, pruned_loss=0.04628, over 956590.46 frames. ], batch size: 30, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:55:19,340 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.480e+02 1.779e+02 2.139e+02 3.586e+02, threshold=3.558e+02, percent-clipped=0.0 +2023-04-28 00:55:33,247 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151088.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:56:16,163 INFO [finetune.py:976] (3/7) Epoch 27, batch 2200, loss[loss=0.1559, simple_loss=0.2351, pruned_loss=0.03829, over 4759.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2407, pruned_loss=0.0469, over 955777.22 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:57:18,615 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-28 00:57:19,653 INFO [finetune.py:976] (3/7) Epoch 27, batch 2250, loss[loss=0.175, simple_loss=0.2613, pruned_loss=0.04439, over 4891.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2433, pruned_loss=0.04818, over 956568.45 frames. ], batch size: 37, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:57:27,225 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.948e+01 1.668e+02 1.911e+02 2.463e+02 4.700e+02, threshold=3.822e+02, percent-clipped=3.0 +2023-04-28 00:57:27,954 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3487, 1.5741, 1.5847, 1.8427, 1.8423, 2.0609, 1.5310, 3.7392], + device='cuda:3'), covar=tensor([0.0566, 0.0766, 0.0739, 0.1132, 0.0569, 0.0453, 0.0708, 0.0120], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 00:57:37,936 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8640, 2.8274, 2.3002, 3.3043, 2.8414, 2.8301, 1.2047, 2.8656], + device='cuda:3'), covar=tensor([0.2451, 0.1655, 0.3163, 0.3004, 0.2825, 0.2291, 0.6194, 0.2839], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0217, 0.0250, 0.0300, 0.0296, 0.0247, 0.0273, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 00:57:38,568 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151184.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:57:49,166 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151193.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:58:22,086 INFO [finetune.py:976] (3/7) Epoch 27, batch 2300, loss[loss=0.1722, simple_loss=0.247, pruned_loss=0.04871, over 4837.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2436, pruned_loss=0.04826, over 956203.66 frames. ], batch size: 49, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:58:28,306 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3203, 2.2394, 2.0846, 1.9474, 2.3879, 2.0793, 2.9435, 1.8622], + device='cuda:3'), covar=tensor([0.3394, 0.1946, 0.4017, 0.2684, 0.1720, 0.2104, 0.1279, 0.3896], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0353, 0.0424, 0.0350, 0.0380, 0.0374, 0.0366, 0.0419], + device='cuda:3'), out_proj_covar=tensor([9.9833e-05, 1.0526e-04, 1.2824e-04, 1.0473e-04, 1.1249e-04, 1.1124e-04, + 1.0690e-04, 1.2602e-04], device='cuda:3') +2023-04-28 00:58:42,722 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151241.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:58:42,789 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9187, 2.3331, 2.0547, 2.2471, 1.8264, 1.9639, 1.9465, 1.6437], + device='cuda:3'), covar=tensor([0.1364, 0.0902, 0.0672, 0.0798, 0.2796, 0.0947, 0.1397, 0.1788], + device='cuda:3'), in_proj_covar=tensor([0.0285, 0.0303, 0.0217, 0.0279, 0.0318, 0.0255, 0.0250, 0.0266], + device='cuda:3'), out_proj_covar=tensor([1.1395e-04, 1.1937e-04, 8.5377e-05, 1.1000e-04, 1.2822e-04, 1.0037e-04, + 1.0080e-04, 1.0495e-04], device='cuda:3') +2023-04-28 00:58:45,168 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151245.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:58:52,341 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151256.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:58:59,486 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6613, 1.5430, 1.7641, 2.0732, 2.0801, 1.6234, 1.2979, 1.7883], + device='cuda:3'), covar=tensor([0.0758, 0.1155, 0.0669, 0.0476, 0.0533, 0.0741, 0.0726, 0.0548], + device='cuda:3'), in_proj_covar=tensor([0.0181, 0.0198, 0.0181, 0.0168, 0.0175, 0.0175, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:59:01,205 INFO [finetune.py:976] (3/7) Epoch 27, batch 2350, loss[loss=0.1421, simple_loss=0.2258, pruned_loss=0.02925, over 4789.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2408, pruned_loss=0.04733, over 954810.73 frames. ], batch size: 51, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:59:03,615 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8622, 2.1456, 1.2472, 1.5391, 2.4559, 1.7272, 1.6046, 1.7678], + device='cuda:3'), covar=tensor([0.0442, 0.0318, 0.0248, 0.0507, 0.0203, 0.0446, 0.0448, 0.0504], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0051, 0.0052], + device='cuda:3') +2023-04-28 00:59:04,082 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.580e+02 1.872e+02 2.240e+02 4.120e+02, threshold=3.744e+02, percent-clipped=1.0 +2023-04-28 00:59:10,166 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151282.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:59:14,374 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9352, 1.7716, 2.2493, 2.1062, 1.9575, 1.9077, 2.0502, 2.0945], + device='cuda:3'), covar=tensor([0.6153, 0.9207, 0.8966, 0.9112, 0.7939, 1.2040, 1.1784, 1.2013], + device='cuda:3'), in_proj_covar=tensor([0.0443, 0.0422, 0.0519, 0.0507, 0.0471, 0.0507, 0.0507, 0.0521], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 00:59:20,456 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151298.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:59:34,699 INFO [finetune.py:976] (3/7) Epoch 27, batch 2400, loss[loss=0.1683, simple_loss=0.2309, pruned_loss=0.05281, over 4825.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2384, pruned_loss=0.04666, over 956932.15 frames. ], batch size: 39, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:59:36,039 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7236, 1.5495, 1.9710, 2.1147, 1.5113, 1.4085, 1.6996, 0.9615], + device='cuda:3'), covar=tensor([0.0438, 0.0615, 0.0361, 0.0541, 0.0657, 0.1043, 0.0508, 0.0625], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 00:59:50,899 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151343.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:00:01,125 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151359.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:00:07,717 INFO [finetune.py:976] (3/7) Epoch 27, batch 2450, loss[loss=0.1341, simple_loss=0.2106, pruned_loss=0.02876, over 4800.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2357, pruned_loss=0.04607, over 956063.94 frames. ], batch size: 29, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:00:10,600 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.893e+01 1.571e+02 1.897e+02 2.275e+02 8.002e+02, threshold=3.795e+02, percent-clipped=4.0 +2023-04-28 01:00:21,185 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151388.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:00:41,810 INFO [finetune.py:976] (3/7) Epoch 27, batch 2500, loss[loss=0.2083, simple_loss=0.2911, pruned_loss=0.06272, over 4735.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.238, pruned_loss=0.0473, over 953580.38 frames. ], batch size: 59, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:00:45,416 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7794, 1.6982, 1.8067, 1.4095, 1.8093, 1.5988, 2.3245, 1.6416], + device='cuda:3'), covar=tensor([0.3622, 0.1882, 0.5129, 0.2656, 0.1574, 0.2159, 0.1497, 0.4372], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0352, 0.0421, 0.0348, 0.0377, 0.0372, 0.0364, 0.0417], + device='cuda:3'), out_proj_covar=tensor([9.9220e-05, 1.0472e-04, 1.2763e-04, 1.0416e-04, 1.1177e-04, 1.1059e-04, + 1.0635e-04, 1.2545e-04], device='cuda:3') +2023-04-28 01:00:53,957 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151436.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:01:15,577 INFO [finetune.py:976] (3/7) Epoch 27, batch 2550, loss[loss=0.1707, simple_loss=0.2439, pruned_loss=0.04874, over 4833.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2406, pruned_loss=0.04749, over 953986.74 frames. ], batch size: 39, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:01:17,942 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.630e+02 1.877e+02 2.163e+02 4.851e+02, threshold=3.753e+02, percent-clipped=1.0 +2023-04-28 01:01:37,867 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.76 vs. limit=5.0 +2023-04-28 01:01:48,927 INFO [finetune.py:976] (3/7) Epoch 27, batch 2600, loss[loss=0.1343, simple_loss=0.1872, pruned_loss=0.04064, over 3605.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2422, pruned_loss=0.04853, over 953256.30 frames. ], batch size: 15, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:02:02,557 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151540.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:02:19,375 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151556.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:02:28,941 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6061, 1.4569, 0.4840, 1.2629, 1.5040, 1.4903, 1.3635, 1.4036], + device='cuda:3'), covar=tensor([0.0516, 0.0398, 0.0402, 0.0567, 0.0291, 0.0504, 0.0510, 0.0567], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0052], + device='cuda:3') +2023-04-28 01:02:31,215 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-04-28 01:02:33,622 INFO [finetune.py:976] (3/7) Epoch 27, batch 2650, loss[loss=0.1641, simple_loss=0.2363, pruned_loss=0.04595, over 4799.00 frames. ], tot_loss[loss=0.169, simple_loss=0.242, pruned_loss=0.04796, over 951693.33 frames. ], batch size: 51, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:02:33,755 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8047, 1.3968, 1.3213, 1.4731, 1.8897, 1.4847, 1.3043, 1.2685], + device='cuda:3'), covar=tensor([0.1583, 0.1655, 0.2272, 0.1359, 0.0934, 0.2029, 0.2164, 0.2454], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0309, 0.0353, 0.0288, 0.0329, 0.0307, 0.0301, 0.0377], + device='cuda:3'), out_proj_covar=tensor([6.4664e-05, 6.3464e-05, 7.4210e-05, 5.7724e-05, 6.7184e-05, 6.4133e-05, + 6.2346e-05, 7.9734e-05], device='cuda:3') +2023-04-28 01:02:41,499 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.560e+02 1.768e+02 2.112e+02 4.460e+02, threshold=3.536e+02, percent-clipped=1.0 +2023-04-28 01:03:17,954 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.92 vs. limit=5.0 +2023-04-28 01:03:22,902 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2736, 2.8617, 2.5185, 2.7269, 2.0371, 2.5552, 2.6453, 1.9201], + device='cuda:3'), covar=tensor([0.2044, 0.1131, 0.0709, 0.1110, 0.3187, 0.0942, 0.1705, 0.2579], + device='cuda:3'), in_proj_covar=tensor([0.0286, 0.0302, 0.0217, 0.0279, 0.0317, 0.0254, 0.0250, 0.0266], + device='cuda:3'), out_proj_covar=tensor([1.1412e-04, 1.1877e-04, 8.5207e-05, 1.0980e-04, 1.2787e-04, 1.0008e-04, + 1.0068e-04, 1.0493e-04], device='cuda:3') +2023-04-28 01:03:23,448 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151604.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:03:39,057 INFO [finetune.py:976] (3/7) Epoch 27, batch 2700, loss[loss=0.1803, simple_loss=0.2479, pruned_loss=0.05635, over 4893.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2412, pruned_loss=0.04728, over 952648.25 frames. ], batch size: 43, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:04:05,039 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151638.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:04:05,743 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9449, 1.2273, 1.7713, 1.8575, 1.8430, 1.8382, 1.7346, 1.7307], + device='cuda:3'), covar=tensor([0.3784, 0.4692, 0.3943, 0.4007, 0.4857, 0.6556, 0.4006, 0.4043], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0376, 0.0330, 0.0340, 0.0349, 0.0394, 0.0359, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 01:04:21,257 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151654.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:04:40,169 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-28 01:04:40,569 INFO [finetune.py:976] (3/7) Epoch 27, batch 2750, loss[loss=0.1794, simple_loss=0.2446, pruned_loss=0.05712, over 4807.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2394, pruned_loss=0.04739, over 952889.49 frames. ], batch size: 41, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:04:48,521 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.458e+01 1.381e+02 1.737e+02 2.155e+02 3.699e+02, threshold=3.473e+02, percent-clipped=1.0 +2023-04-28 01:05:07,216 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151688.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:05:44,336 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-28 01:05:44,741 INFO [finetune.py:976] (3/7) Epoch 27, batch 2800, loss[loss=0.1512, simple_loss=0.233, pruned_loss=0.03469, over 4755.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2381, pruned_loss=0.04725, over 952392.79 frames. ], batch size: 26, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:06:24,902 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151749.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:06:53,982 INFO [finetune.py:976] (3/7) Epoch 27, batch 2850, loss[loss=0.2029, simple_loss=0.2786, pruned_loss=0.0636, over 4905.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2363, pruned_loss=0.04644, over 952882.41 frames. ], batch size: 37, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:06:56,485 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.506e+02 1.770e+02 2.075e+02 3.794e+02, threshold=3.540e+02, percent-clipped=1.0 +2023-04-28 01:07:29,480 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151801.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:07:58,280 INFO [finetune.py:976] (3/7) Epoch 27, batch 2900, loss[loss=0.1936, simple_loss=0.2664, pruned_loss=0.06034, over 4824.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2385, pruned_loss=0.04718, over 954051.72 frames. ], batch size: 39, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:08:11,750 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151840.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:08:32,203 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151862.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:08:42,905 INFO [finetune.py:976] (3/7) Epoch 27, batch 2950, loss[loss=0.1438, simple_loss=0.2342, pruned_loss=0.02676, over 4810.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2413, pruned_loss=0.04738, over 954726.20 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:08:50,553 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.751e+01 1.623e+02 1.878e+02 2.443e+02 4.733e+02, threshold=3.756e+02, percent-clipped=2.0 +2023-04-28 01:09:04,984 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151888.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:09:34,740 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5925, 1.6174, 0.8814, 1.2820, 1.7188, 1.4822, 1.4019, 1.4528], + device='cuda:3'), covar=tensor([0.0470, 0.0331, 0.0314, 0.0497, 0.0267, 0.0451, 0.0433, 0.0532], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0050, 0.0052], + device='cuda:3') +2023-04-28 01:09:44,046 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-04-28 01:09:46,746 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 01:09:47,998 INFO [finetune.py:976] (3/7) Epoch 27, batch 3000, loss[loss=0.1527, simple_loss=0.2404, pruned_loss=0.03246, over 4893.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2433, pruned_loss=0.04771, over 956382.91 frames. ], batch size: 43, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:09:47,998 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 01:10:08,675 INFO [finetune.py:1010] (3/7) Epoch 27, validation: loss=0.1539, simple_loss=0.2224, pruned_loss=0.04268, over 2265189.00 frames. +2023-04-28 01:10:08,676 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 01:10:09,427 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9784, 1.9293, 2.5714, 2.6500, 1.7284, 1.6519, 2.0170, 1.2023], + device='cuda:3'), covar=tensor([0.0514, 0.0658, 0.0328, 0.0568, 0.0661, 0.0998, 0.0554, 0.0644], + device='cuda:3'), in_proj_covar=tensor([0.0069, 0.0067, 0.0065, 0.0068, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 01:10:24,778 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151938.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:10:28,097 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-04-28 01:10:35,083 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151954.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:10:45,721 INFO [finetune.py:976] (3/7) Epoch 27, batch 3050, loss[loss=0.1781, simple_loss=0.235, pruned_loss=0.06058, over 4756.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2438, pruned_loss=0.04773, over 956211.85 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:10:48,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.615e+02 1.912e+02 2.194e+02 5.604e+02, threshold=3.825e+02, percent-clipped=1.0 +2023-04-28 01:10:52,483 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-28 01:10:57,067 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151986.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:11:01,364 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.5831, 3.5743, 2.8630, 4.1325, 3.3903, 3.5668, 1.8288, 3.5273], + device='cuda:3'), covar=tensor([0.1560, 0.1147, 0.4290, 0.1372, 0.3092, 0.1736, 0.5095, 0.2285], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0218, 0.0251, 0.0302, 0.0298, 0.0247, 0.0274, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 01:11:08,925 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152002.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:11:12,698 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0544, 4.0698, 0.6636, 2.2667, 2.3557, 2.5880, 2.4455, 0.9805], + device='cuda:3'), covar=tensor([0.1138, 0.0865, 0.2125, 0.1076, 0.0926, 0.1072, 0.1385, 0.2158], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0135, 0.0121, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 01:11:20,459 INFO [finetune.py:976] (3/7) Epoch 27, batch 3100, loss[loss=0.1328, simple_loss=0.2023, pruned_loss=0.03166, over 4937.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2417, pruned_loss=0.04681, over 956533.43 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:11:31,567 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0412, 1.2467, 5.3649, 4.9972, 4.5986, 5.0943, 4.7453, 4.7588], + device='cuda:3'), covar=tensor([0.7027, 0.6389, 0.1009, 0.1806, 0.1026, 0.1395, 0.1097, 0.1603], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0313, 0.0415, 0.0413, 0.0354, 0.0421, 0.0324, 0.0369], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:11:37,593 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152044.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:11:54,337 INFO [finetune.py:976] (3/7) Epoch 27, batch 3150, loss[loss=0.2046, simple_loss=0.2688, pruned_loss=0.07018, over 4826.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2393, pruned_loss=0.04653, over 952055.76 frames. ], batch size: 41, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:11:56,743 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.740e+01 1.539e+02 1.884e+02 2.286e+02 5.253e+02, threshold=3.767e+02, percent-clipped=2.0 +2023-04-28 01:12:27,069 INFO [finetune.py:976] (3/7) Epoch 27, batch 3200, loss[loss=0.1633, simple_loss=0.2399, pruned_loss=0.04331, over 4912.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2365, pruned_loss=0.0455, over 954008.10 frames. ], batch size: 36, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:12:52,024 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152157.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:12:55,835 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-28 01:13:00,433 INFO [finetune.py:976] (3/7) Epoch 27, batch 3250, loss[loss=0.196, simple_loss=0.2609, pruned_loss=0.06552, over 4766.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2378, pruned_loss=0.04622, over 954515.35 frames. ], batch size: 54, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:13:02,810 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.216e+01 1.479e+02 1.800e+02 2.164e+02 4.753e+02, threshold=3.600e+02, percent-clipped=3.0 +2023-04-28 01:13:33,550 INFO [finetune.py:976] (3/7) Epoch 27, batch 3300, loss[loss=0.1937, simple_loss=0.2711, pruned_loss=0.05815, over 4768.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2424, pruned_loss=0.04752, over 956282.82 frames. ], batch size: 54, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:13:36,714 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152225.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:13:40,965 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4037, 1.7522, 1.8598, 1.9580, 1.8098, 1.8278, 1.9123, 1.8885], + device='cuda:3'), covar=tensor([0.4394, 0.5319, 0.4577, 0.4424, 0.5527, 0.6843, 0.5254, 0.4609], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0376, 0.0332, 0.0341, 0.0351, 0.0393, 0.0362, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 01:14:13,061 INFO [finetune.py:976] (3/7) Epoch 27, batch 3350, loss[loss=0.1881, simple_loss=0.2501, pruned_loss=0.06308, over 4893.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2436, pruned_loss=0.04806, over 954764.76 frames. ], batch size: 35, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:14:14,902 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152273.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:14:15,406 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.492e+02 1.749e+02 2.149e+02 5.486e+02, threshold=3.498e+02, percent-clipped=3.0 +2023-04-28 01:14:34,621 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152286.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:15:17,723 INFO [finetune.py:976] (3/7) Epoch 27, batch 3400, loss[loss=0.1575, simple_loss=0.2382, pruned_loss=0.0384, over 4756.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2444, pruned_loss=0.04806, over 955417.81 frames. ], batch size: 27, lr: 2.92e-03, grad_scale: 32.0 +2023-04-28 01:15:36,876 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152334.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:15:50,205 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152344.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:16:22,890 INFO [finetune.py:976] (3/7) Epoch 27, batch 3450, loss[loss=0.1301, simple_loss=0.2131, pruned_loss=0.02352, over 4759.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2448, pruned_loss=0.04799, over 956512.51 frames. ], batch size: 28, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:16:31,451 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.589e+02 1.871e+02 2.255e+02 4.038e+02, threshold=3.742e+02, percent-clipped=2.0 +2023-04-28 01:16:53,822 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152392.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:17:26,788 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5143, 3.0022, 1.0617, 1.8665, 2.4480, 1.5577, 4.1395, 2.2956], + device='cuda:3'), covar=tensor([0.0621, 0.0812, 0.0947, 0.1186, 0.0503, 0.0932, 0.0160, 0.0555], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 01:17:27,970 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3770, 1.2661, 4.1222, 3.8744, 3.6129, 3.9569, 3.8827, 3.6082], + device='cuda:3'), covar=tensor([0.7147, 0.5761, 0.1118, 0.1671, 0.1255, 0.1319, 0.1585, 0.1467], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0309, 0.0411, 0.0408, 0.0350, 0.0417, 0.0320, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:17:28,479 INFO [finetune.py:976] (3/7) Epoch 27, batch 3500, loss[loss=0.1881, simple_loss=0.2592, pruned_loss=0.05847, over 4832.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2425, pruned_loss=0.04775, over 957595.59 frames. ], batch size: 33, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:17:29,217 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4544, 1.1358, 1.2213, 1.1672, 1.6115, 1.3284, 1.1221, 1.1705], + device='cuda:3'), covar=tensor([0.1323, 0.1139, 0.1520, 0.1239, 0.0667, 0.1363, 0.1483, 0.2063], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0309, 0.0353, 0.0287, 0.0328, 0.0307, 0.0301, 0.0376], + device='cuda:3'), out_proj_covar=tensor([6.4551e-05, 6.3538e-05, 7.4103e-05, 5.7530e-05, 6.7135e-05, 6.3987e-05, + 6.2339e-05, 7.9566e-05], device='cuda:3') +2023-04-28 01:18:00,042 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152445.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:18:01,799 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-28 01:18:18,960 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152457.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:18:32,264 INFO [finetune.py:976] (3/7) Epoch 27, batch 3550, loss[loss=0.1599, simple_loss=0.2212, pruned_loss=0.04928, over 4072.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.24, pruned_loss=0.04727, over 957393.97 frames. ], batch size: 17, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:18:34,714 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.620e+01 1.515e+02 1.760e+02 2.193e+02 3.921e+02, threshold=3.521e+02, percent-clipped=1.0 +2023-04-28 01:18:42,749 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4676, 1.8685, 1.6627, 2.1535, 2.0146, 2.1563, 1.7516, 4.4403], + device='cuda:3'), covar=tensor([0.0507, 0.0725, 0.0714, 0.1058, 0.0557, 0.0482, 0.0649, 0.0089], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 01:19:25,050 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152505.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:19:25,762 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152506.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:19:39,438 INFO [finetune.py:976] (3/7) Epoch 27, batch 3600, loss[loss=0.1333, simple_loss=0.1981, pruned_loss=0.03422, over 4186.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2371, pruned_loss=0.04658, over 957210.30 frames. ], batch size: 18, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:19:49,563 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-28 01:20:33,907 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1470, 1.7226, 2.0627, 2.1364, 2.0620, 1.7295, 1.2861, 1.7773], + device='cuda:3'), covar=tensor([0.3207, 0.2931, 0.1600, 0.2019, 0.2416, 0.2595, 0.3953, 0.1756], + device='cuda:3'), in_proj_covar=tensor([0.0297, 0.0249, 0.0230, 0.0318, 0.0224, 0.0236, 0.0230, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 01:20:44,727 INFO [finetune.py:976] (3/7) Epoch 27, batch 3650, loss[loss=0.1804, simple_loss=0.2564, pruned_loss=0.05217, over 4819.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2397, pruned_loss=0.04822, over 954115.85 frames. ], batch size: 38, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:20:46,018 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0206, 1.2838, 5.1220, 4.7876, 4.3528, 4.8751, 4.4952, 4.5031], + device='cuda:3'), covar=tensor([0.6657, 0.6202, 0.0866, 0.1626, 0.1098, 0.1330, 0.1428, 0.1745], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0309, 0.0411, 0.0409, 0.0350, 0.0418, 0.0321, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:20:51,634 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.532e+02 1.875e+02 2.206e+02 4.612e+02, threshold=3.749e+02, percent-clipped=4.0 +2023-04-28 01:20:56,028 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152581.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:21:02,385 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152582.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:21:12,801 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152589.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:21:49,538 INFO [finetune.py:976] (3/7) Epoch 27, batch 3700, loss[loss=0.1874, simple_loss=0.2739, pruned_loss=0.05038, over 4820.00 frames. ], tot_loss[loss=0.17, simple_loss=0.243, pruned_loss=0.04854, over 954079.88 frames. ], batch size: 47, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:22:00,773 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152629.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:22:21,235 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152643.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:22:31,085 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152650.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:22:56,764 INFO [finetune.py:976] (3/7) Epoch 27, batch 3750, loss[loss=0.2112, simple_loss=0.2713, pruned_loss=0.07552, over 4832.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2433, pruned_loss=0.04814, over 953566.86 frames. ], batch size: 30, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:23:04,992 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.529e+02 1.747e+02 2.097e+02 4.287e+02, threshold=3.495e+02, percent-clipped=2.0 +2023-04-28 01:23:37,927 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1127, 1.7713, 2.1670, 2.5560, 2.4973, 2.1167, 1.8292, 2.2410], + device='cuda:3'), covar=tensor([0.0783, 0.1050, 0.0622, 0.0467, 0.0563, 0.0745, 0.0690, 0.0509], + device='cuda:3'), in_proj_covar=tensor([0.0182, 0.0199, 0.0182, 0.0168, 0.0176, 0.0176, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:24:08,007 INFO [finetune.py:976] (3/7) Epoch 27, batch 3800, loss[loss=0.1708, simple_loss=0.236, pruned_loss=0.05276, over 4694.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2441, pruned_loss=0.04814, over 953217.18 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:24:11,816 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152726.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:24:29,225 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1031, 1.2174, 5.3053, 5.0107, 4.5691, 5.0395, 4.6682, 4.6705], + device='cuda:3'), covar=tensor([0.6788, 0.6524, 0.0958, 0.1723, 0.0989, 0.1296, 0.1111, 0.1660], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0310, 0.0411, 0.0409, 0.0350, 0.0418, 0.0321, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:24:31,130 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7707, 1.2464, 1.8543, 2.2254, 1.8368, 1.7797, 1.8060, 1.7613], + device='cuda:3'), covar=tensor([0.4442, 0.6806, 0.5974, 0.5704, 0.6012, 0.8114, 0.7653, 0.9224], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0421, 0.0517, 0.0507, 0.0469, 0.0506, 0.0505, 0.0522], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:25:13,181 INFO [finetune.py:976] (3/7) Epoch 27, batch 3850, loss[loss=0.1439, simple_loss=0.2233, pruned_loss=0.03224, over 4927.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2424, pruned_loss=0.0471, over 955670.14 frames. ], batch size: 33, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:25:15,590 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.420e+02 1.784e+02 2.045e+02 3.500e+02, threshold=3.567e+02, percent-clipped=1.0 +2023-04-28 01:25:24,522 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3561, 1.6673, 1.8447, 1.9877, 1.8427, 1.8836, 1.8592, 1.8841], + device='cuda:3'), covar=tensor([0.3930, 0.4969, 0.4291, 0.4072, 0.5272, 0.6705, 0.4837, 0.4626], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0375, 0.0330, 0.0341, 0.0350, 0.0394, 0.0361, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 01:25:33,560 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152787.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:25:47,900 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152801.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:26:06,635 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6693, 3.3134, 3.0245, 3.1239, 2.3109, 2.7841, 2.9979, 2.0035], + device='cuda:3'), covar=tensor([0.1974, 0.1147, 0.0697, 0.1168, 0.3172, 0.1142, 0.1851, 0.2897], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0300, 0.0215, 0.0275, 0.0314, 0.0253, 0.0246, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1261e-04, 1.1805e-04, 8.4708e-05, 1.0846e-04, 1.2637e-04, 9.9421e-05, + 9.9246e-05, 1.0392e-04], device='cuda:3') +2023-04-28 01:26:08,592 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-28 01:26:17,504 INFO [finetune.py:976] (3/7) Epoch 27, batch 3900, loss[loss=0.1344, simple_loss=0.2126, pruned_loss=0.02804, over 4857.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2396, pruned_loss=0.04629, over 957684.44 frames. ], batch size: 31, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:26:53,011 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-28 01:27:01,809 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 01:27:12,213 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7078, 3.6632, 0.9565, 1.9121, 1.9572, 2.5085, 1.9508, 1.0518], + device='cuda:3'), covar=tensor([0.1376, 0.0782, 0.2024, 0.1282, 0.1125, 0.1066, 0.1609, 0.1977], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0134, 0.0120, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 01:27:22,908 INFO [finetune.py:976] (3/7) Epoch 27, batch 3950, loss[loss=0.1489, simple_loss=0.2192, pruned_loss=0.03933, over 4846.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2372, pruned_loss=0.0453, over 954491.31 frames. ], batch size: 49, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:27:26,326 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.447e+02 1.748e+02 2.157e+02 5.230e+02, threshold=3.496e+02, percent-clipped=2.0 +2023-04-28 01:27:36,714 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152881.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:28:27,785 INFO [finetune.py:976] (3/7) Epoch 27, batch 4000, loss[loss=0.1704, simple_loss=0.2363, pruned_loss=0.05226, over 4900.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.236, pruned_loss=0.04522, over 952867.02 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:28:39,703 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:28:39,752 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:28:50,586 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152938.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:29:00,137 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152945.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:29:32,517 INFO [finetune.py:976] (3/7) Epoch 27, batch 4050, loss[loss=0.2077, simple_loss=0.2921, pruned_loss=0.0617, over 4840.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2388, pruned_loss=0.0456, over 953960.79 frames. ], batch size: 47, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:29:35,453 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.640e+02 1.918e+02 2.261e+02 4.754e+02, threshold=3.835e+02, percent-clipped=2.0 +2023-04-28 01:29:43,789 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152977.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:30:36,409 INFO [finetune.py:976] (3/7) Epoch 27, batch 4100, loss[loss=0.1394, simple_loss=0.2151, pruned_loss=0.03184, over 4793.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2395, pruned_loss=0.04568, over 950888.38 frames. ], batch size: 25, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:30:58,681 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3534, 2.1639, 2.5037, 2.8549, 2.8142, 2.3312, 2.0750, 2.5245], + device='cuda:3'), covar=tensor([0.0799, 0.0967, 0.0620, 0.0512, 0.0546, 0.0754, 0.0653, 0.0534], + device='cuda:3'), in_proj_covar=tensor([0.0183, 0.0199, 0.0182, 0.0169, 0.0176, 0.0176, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:31:40,930 INFO [finetune.py:976] (3/7) Epoch 27, batch 4150, loss[loss=0.208, simple_loss=0.2752, pruned_loss=0.07044, over 4737.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2415, pruned_loss=0.04654, over 952170.80 frames. ], batch size: 54, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:31:43,381 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.659e+02 1.933e+02 2.318e+02 5.641e+02, threshold=3.866e+02, percent-clipped=2.0 +2023-04-28 01:31:54,613 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153082.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:32:22,862 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153101.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:32:43,337 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153117.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:32:45,074 INFO [finetune.py:976] (3/7) Epoch 27, batch 4200, loss[loss=0.2024, simple_loss=0.2661, pruned_loss=0.06939, over 4901.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2419, pruned_loss=0.04645, over 954168.47 frames. ], batch size: 46, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:33:26,254 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153149.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:33:49,691 INFO [finetune.py:976] (3/7) Epoch 27, batch 4250, loss[loss=0.153, simple_loss=0.2264, pruned_loss=0.03979, over 4902.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2399, pruned_loss=0.04616, over 953381.47 frames. ], batch size: 46, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:33:57,256 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.449e+02 1.716e+02 2.053e+02 3.736e+02, threshold=3.432e+02, percent-clipped=0.0 +2023-04-28 01:33:59,838 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153178.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:34:54,745 INFO [finetune.py:976] (3/7) Epoch 27, batch 4300, loss[loss=0.1333, simple_loss=0.2107, pruned_loss=0.02797, over 4925.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2376, pruned_loss=0.0456, over 955823.86 frames. ], batch size: 33, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:35:03,528 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1532, 1.8564, 2.3557, 2.5539, 2.1732, 2.0914, 2.2362, 2.1200], + device='cuda:3'), covar=tensor([0.4588, 0.7511, 0.6922, 0.5230, 0.5950, 0.8823, 0.8127, 1.0164], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0422, 0.0517, 0.0507, 0.0469, 0.0506, 0.0507, 0.0523], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:35:17,368 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153238.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:35:23,028 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4497, 1.9144, 2.3480, 2.7501, 2.3377, 1.8909, 1.6450, 2.0821], + device='cuda:3'), covar=tensor([0.3064, 0.2829, 0.1595, 0.1932, 0.2283, 0.2508, 0.3755, 0.1817], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0246, 0.0229, 0.0315, 0.0222, 0.0235, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 01:35:23,569 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153245.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:35:38,705 INFO [finetune.py:976] (3/7) Epoch 27, batch 4350, loss[loss=0.1456, simple_loss=0.2129, pruned_loss=0.03913, over 4822.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2347, pruned_loss=0.04502, over 956905.94 frames. ], batch size: 30, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:35:38,839 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6749, 1.5580, 2.0189, 2.0731, 1.5046, 1.3923, 1.6885, 1.2039], + device='cuda:3'), covar=tensor([0.0496, 0.0603, 0.0320, 0.0447, 0.0685, 0.1070, 0.0528, 0.0523], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 01:35:41,098 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.585e+02 1.848e+02 2.377e+02 5.336e+02, threshold=3.696e+02, percent-clipped=3.0 +2023-04-28 01:35:44,151 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153278.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:35:48,973 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153286.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:35:53,613 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8468, 2.8004, 2.1627, 3.2601, 2.8885, 2.8362, 1.0906, 2.8064], + device='cuda:3'), covar=tensor([0.2324, 0.1935, 0.3611, 0.3085, 0.4442, 0.2220, 0.6120, 0.3134], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0218, 0.0251, 0.0303, 0.0299, 0.0247, 0.0274, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 01:35:54,193 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153293.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:36:12,553 INFO [finetune.py:976] (3/7) Epoch 27, batch 4400, loss[loss=0.1401, simple_loss=0.2296, pruned_loss=0.02535, over 4818.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2373, pruned_loss=0.04615, over 957017.97 frames. ], batch size: 40, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:36:15,824 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-28 01:36:33,329 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153339.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:37:16,504 INFO [finetune.py:976] (3/7) Epoch 27, batch 4450, loss[loss=0.1938, simple_loss=0.2653, pruned_loss=0.06112, over 4821.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2401, pruned_loss=0.04692, over 956695.21 frames. ], batch size: 40, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:37:18,377 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5289, 3.2016, 0.9322, 1.8076, 1.6722, 2.2174, 1.8125, 1.0712], + device='cuda:3'), covar=tensor([0.1340, 0.0964, 0.1961, 0.1283, 0.1176, 0.1112, 0.1611, 0.2034], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0240, 0.0136, 0.0121, 0.0132, 0.0153, 0.0118, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 01:37:18,878 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.598e+02 1.820e+02 2.307e+02 3.312e+02, threshold=3.640e+02, percent-clipped=0.0 +2023-04-28 01:37:30,038 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153382.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:38:10,739 INFO [finetune.py:976] (3/7) Epoch 27, batch 4500, loss[loss=0.2101, simple_loss=0.2739, pruned_loss=0.07317, over 4905.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2425, pruned_loss=0.04786, over 956150.85 frames. ], batch size: 37, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:38:11,439 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153421.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:38:16,908 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153430.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:38:18,778 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2752, 1.5139, 1.3183, 1.7210, 1.6582, 1.8337, 1.3683, 3.3425], + device='cuda:3'), covar=tensor([0.0617, 0.0824, 0.0840, 0.1259, 0.0674, 0.0509, 0.0822, 0.0164], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 01:38:34,207 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0902, 1.3857, 1.2030, 1.5960, 1.4904, 1.5319, 1.3457, 2.3966], + device='cuda:3'), covar=tensor([0.0601, 0.0793, 0.0827, 0.1248, 0.0646, 0.0460, 0.0745, 0.0219], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 01:38:43,736 INFO [finetune.py:976] (3/7) Epoch 27, batch 4550, loss[loss=0.144, simple_loss=0.2297, pruned_loss=0.02916, over 4808.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2429, pruned_loss=0.04738, over 955334.56 frames. ], batch size: 41, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:38:45,569 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153473.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:38:46,107 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.179e+01 1.577e+02 1.859e+02 2.196e+02 3.775e+02, threshold=3.717e+02, percent-clipped=3.0 +2023-04-28 01:38:51,089 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153482.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:39:05,222 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6348, 1.5040, 1.9177, 2.0320, 1.4698, 1.3667, 1.6325, 0.9651], + device='cuda:3'), covar=tensor([0.0462, 0.0664, 0.0349, 0.0491, 0.0715, 0.1078, 0.0528, 0.0706], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0069, 0.0075, 0.0095, 0.0072, 0.0063], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 01:39:08,061 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1686, 1.4648, 1.3210, 1.6275, 1.5355, 1.6371, 1.3517, 2.9950], + device='cuda:3'), covar=tensor([0.0639, 0.0808, 0.0793, 0.1233, 0.0621, 0.0531, 0.0748, 0.0175], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 01:39:15,748 INFO [finetune.py:976] (3/7) Epoch 27, batch 4600, loss[loss=0.1415, simple_loss=0.2171, pruned_loss=0.03299, over 4790.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2424, pruned_loss=0.04693, over 956252.18 frames. ], batch size: 29, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:39:48,522 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8824, 1.4458, 1.4181, 1.5961, 2.0256, 1.5972, 1.3971, 1.3680], + device='cuda:3'), covar=tensor([0.1442, 0.1390, 0.1991, 0.1205, 0.0814, 0.1539, 0.2095, 0.2402], + device='cuda:3'), in_proj_covar=tensor([0.0319, 0.0311, 0.0356, 0.0289, 0.0331, 0.0309, 0.0304, 0.0380], + device='cuda:3'), out_proj_covar=tensor([6.5071e-05, 6.3830e-05, 7.4695e-05, 5.7888e-05, 6.7718e-05, 6.4473e-05, + 6.3000e-05, 8.0391e-05], device='cuda:3') +2023-04-28 01:39:48,984 INFO [finetune.py:976] (3/7) Epoch 27, batch 4650, loss[loss=0.1798, simple_loss=0.2453, pruned_loss=0.05721, over 4863.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.24, pruned_loss=0.04672, over 958759.12 frames. ], batch size: 34, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:39:51,380 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.905e+01 1.489e+02 1.829e+02 2.275e+02 4.563e+02, threshold=3.657e+02, percent-clipped=2.0 +2023-04-28 01:40:27,423 INFO [finetune.py:976] (3/7) Epoch 27, batch 4700, loss[loss=0.1762, simple_loss=0.2357, pruned_loss=0.05837, over 4865.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2371, pruned_loss=0.04614, over 956957.58 frames. ], batch size: 31, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:40:30,723 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-28 01:40:41,151 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153634.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:40:43,101 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-28 01:41:06,296 INFO [finetune.py:976] (3/7) Epoch 27, batch 4750, loss[loss=0.198, simple_loss=0.2605, pruned_loss=0.06777, over 4826.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2357, pruned_loss=0.04609, over 958134.84 frames. ], batch size: 30, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:41:07,622 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7662, 1.2072, 1.8671, 2.2346, 1.8195, 1.7203, 1.7759, 1.7563], + device='cuda:3'), covar=tensor([0.4855, 0.7235, 0.6628, 0.5756, 0.6229, 0.8122, 0.8446, 0.9044], + device='cuda:3'), in_proj_covar=tensor([0.0444, 0.0425, 0.0520, 0.0509, 0.0472, 0.0510, 0.0511, 0.0526], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:41:08,702 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.498e+02 1.801e+02 2.144e+02 5.776e+02, threshold=3.603e+02, percent-clipped=1.0 +2023-04-28 01:41:11,252 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2730, 1.1865, 4.1129, 3.9145, 3.5806, 3.9252, 3.9023, 3.5475], + device='cuda:3'), covar=tensor([0.7104, 0.5917, 0.1107, 0.1702, 0.1173, 0.1976, 0.1429, 0.1676], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0311, 0.0410, 0.0410, 0.0350, 0.0419, 0.0319, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:41:11,293 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8917, 1.8946, 1.7572, 1.5123, 1.8586, 1.5324, 2.5037, 1.4296], + device='cuda:3'), covar=tensor([0.3990, 0.1918, 0.4621, 0.3402, 0.1911, 0.2625, 0.1402, 0.5162], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0356, 0.0425, 0.0353, 0.0382, 0.0375, 0.0371, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:41:39,703 INFO [finetune.py:976] (3/7) Epoch 27, batch 4800, loss[loss=0.1884, simple_loss=0.2739, pruned_loss=0.05142, over 4807.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2386, pruned_loss=0.04668, over 954358.23 frames. ], batch size: 45, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:42:13,213 INFO [finetune.py:976] (3/7) Epoch 27, batch 4850, loss[loss=0.1916, simple_loss=0.2707, pruned_loss=0.05628, over 4899.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2423, pruned_loss=0.04726, over 955047.83 frames. ], batch size: 43, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:42:15,631 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153773.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:42:16,121 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.558e+02 1.785e+02 2.146e+02 3.572e+02, threshold=3.570e+02, percent-clipped=0.0 +2023-04-28 01:42:17,991 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153777.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:42:44,722 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1853, 2.1452, 1.8620, 1.8175, 2.3137, 1.9162, 2.9671, 1.7178], + device='cuda:3'), covar=tensor([0.4046, 0.2064, 0.4743, 0.3368, 0.1918, 0.2586, 0.1574, 0.4626], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0355, 0.0424, 0.0351, 0.0380, 0.0374, 0.0369, 0.0422], + device='cuda:3'), out_proj_covar=tensor([9.9878e-05, 1.0563e-04, 1.2844e-04, 1.0511e-04, 1.1236e-04, 1.1121e-04, + 1.0798e-04, 1.2682e-04], device='cuda:3') +2023-04-28 01:43:04,182 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0264, 2.5280, 2.2678, 2.4275, 1.7386, 2.2349, 2.0915, 1.6504], + device='cuda:3'), covar=tensor([0.2199, 0.1188, 0.0729, 0.1309, 0.3370, 0.1178, 0.2228, 0.2649], + device='cuda:3'), in_proj_covar=tensor([0.0283, 0.0300, 0.0215, 0.0276, 0.0315, 0.0253, 0.0247, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1275e-04, 1.1801e-04, 8.4550e-05, 1.0840e-04, 1.2692e-04, 9.9387e-05, + 9.9598e-05, 1.0362e-04], device='cuda:3') +2023-04-28 01:43:05,706 INFO [finetune.py:976] (3/7) Epoch 27, batch 4900, loss[loss=0.1977, simple_loss=0.2676, pruned_loss=0.0639, over 4829.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2439, pruned_loss=0.04788, over 955994.26 frames. ], batch size: 30, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:43:05,831 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6339, 2.7031, 2.2317, 2.3135, 2.6191, 2.4481, 3.7746, 2.1052], + device='cuda:3'), covar=tensor([0.3795, 0.2554, 0.4815, 0.3657, 0.1935, 0.2623, 0.1364, 0.4274], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0355, 0.0424, 0.0351, 0.0379, 0.0374, 0.0369, 0.0422], + device='cuda:3'), out_proj_covar=tensor([9.9797e-05, 1.0564e-04, 1.2840e-04, 1.0506e-04, 1.1227e-04, 1.1111e-04, + 1.0791e-04, 1.2680e-04], device='cuda:3') +2023-04-28 01:43:06,908 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153821.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:44:09,303 INFO [finetune.py:976] (3/7) Epoch 27, batch 4950, loss[loss=0.2462, simple_loss=0.3105, pruned_loss=0.09099, over 4819.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2453, pruned_loss=0.04812, over 954287.53 frames. ], batch size: 38, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:44:17,526 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5260, 1.7830, 0.9399, 1.2495, 2.0184, 1.3563, 1.3116, 1.3646], + device='cuda:3'), covar=tensor([0.0548, 0.0300, 0.0335, 0.0564, 0.0276, 0.0590, 0.0579, 0.0596], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0052], + device='cuda:3') +2023-04-28 01:44:18,027 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.601e+02 1.869e+02 2.254e+02 5.628e+02, threshold=3.738e+02, percent-clipped=5.0 +2023-04-28 01:44:51,832 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9679, 1.9289, 1.2553, 1.6727, 1.9915, 1.8094, 1.7416, 1.8647], + device='cuda:3'), covar=tensor([0.0433, 0.0324, 0.0290, 0.0484, 0.0252, 0.0430, 0.0399, 0.0494], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0051, 0.0052], + device='cuda:3') +2023-04-28 01:45:13,176 INFO [finetune.py:976] (3/7) Epoch 27, batch 5000, loss[loss=0.1453, simple_loss=0.2176, pruned_loss=0.03646, over 4770.00 frames. ], tot_loss[loss=0.169, simple_loss=0.243, pruned_loss=0.04747, over 954656.51 frames. ], batch size: 26, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:45:21,157 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153923.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:45:34,548 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153934.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:46:15,668 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8159, 1.5921, 1.9102, 2.1838, 2.2253, 1.6894, 1.5867, 1.9365], + device='cuda:3'), covar=tensor([0.0825, 0.1248, 0.0723, 0.0527, 0.0544, 0.0774, 0.0655, 0.0533], + device='cuda:3'), in_proj_covar=tensor([0.0186, 0.0203, 0.0185, 0.0173, 0.0179, 0.0180, 0.0153, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 01:46:17,422 INFO [finetune.py:976] (3/7) Epoch 27, batch 5050, loss[loss=0.1323, simple_loss=0.2093, pruned_loss=0.02771, over 4876.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2389, pruned_loss=0.04595, over 955804.88 frames. ], batch size: 34, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:46:25,377 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.912e+01 1.615e+02 1.861e+02 2.291e+02 3.934e+02, threshold=3.722e+02, percent-clipped=1.0 +2023-04-28 01:46:32,334 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153982.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:46:34,105 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153984.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:46:57,619 INFO [finetune.py:976] (3/7) Epoch 27, batch 5100, loss[loss=0.1781, simple_loss=0.2415, pruned_loss=0.05735, over 4871.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2369, pruned_loss=0.04542, over 957277.05 frames. ], batch size: 49, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:47:18,518 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154049.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 01:47:28,932 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6887, 2.7940, 2.2206, 2.4254, 2.7342, 2.3216, 3.6813, 2.0723], + device='cuda:3'), covar=tensor([0.3600, 0.2058, 0.4411, 0.3107, 0.1803, 0.2626, 0.1224, 0.4118], + device='cuda:3'), in_proj_covar=tensor([0.0338, 0.0353, 0.0422, 0.0350, 0.0379, 0.0374, 0.0368, 0.0420], + device='cuda:3'), out_proj_covar=tensor([9.9649e-05, 1.0522e-04, 1.2782e-04, 1.0480e-04, 1.1214e-04, 1.1119e-04, + 1.0749e-04, 1.2628e-04], device='cuda:3') +2023-04-28 01:47:31,251 INFO [finetune.py:976] (3/7) Epoch 27, batch 5150, loss[loss=0.1755, simple_loss=0.2533, pruned_loss=0.04883, over 4836.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2374, pruned_loss=0.04547, over 956584.63 frames. ], batch size: 30, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:47:33,690 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.527e+01 1.424e+02 1.690e+02 2.200e+02 4.419e+02, threshold=3.381e+02, percent-clipped=1.0 +2023-04-28 01:47:35,047 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154076.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:47:35,620 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154077.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:47:41,334 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154084.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:47:41,350 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1452, 0.7998, 0.9255, 0.8152, 1.2088, 1.0050, 0.8366, 0.9905], + device='cuda:3'), covar=tensor([0.1731, 0.1361, 0.1920, 0.1461, 0.0962, 0.1342, 0.1595, 0.2267], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0305, 0.0349, 0.0284, 0.0325, 0.0303, 0.0299, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.3812e-05, 6.2691e-05, 7.3277e-05, 5.6744e-05, 6.6439e-05, 6.3204e-05, + 6.1810e-05, 7.9120e-05], device='cuda:3') +2023-04-28 01:48:15,445 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154110.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 01:48:26,894 INFO [finetune.py:976] (3/7) Epoch 27, batch 5200, loss[loss=0.1663, simple_loss=0.2421, pruned_loss=0.04521, over 4872.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2403, pruned_loss=0.04619, over 956132.76 frames. ], batch size: 34, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:48:35,436 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154125.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:48:36,957 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-28 01:48:49,923 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154137.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 01:49:00,232 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154145.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:49:20,755 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9289, 2.9071, 2.1689, 3.3283, 2.9285, 2.9611, 1.2550, 2.8974], + device='cuda:3'), covar=tensor([0.1996, 0.1578, 0.3067, 0.2637, 0.3843, 0.2050, 0.5399, 0.2603], + device='cuda:3'), in_proj_covar=tensor([0.0244, 0.0218, 0.0250, 0.0303, 0.0298, 0.0247, 0.0272, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 01:49:21,701 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-28 01:49:31,678 INFO [finetune.py:976] (3/7) Epoch 27, batch 5250, loss[loss=0.189, simple_loss=0.2583, pruned_loss=0.05985, over 4895.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2424, pruned_loss=0.04679, over 956431.49 frames. ], batch size: 36, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:49:34,130 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.655e+02 1.963e+02 2.223e+02 5.480e+02, threshold=3.927e+02, percent-clipped=5.0 +2023-04-28 01:50:28,039 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1608, 2.0592, 1.8267, 1.7724, 2.2090, 1.8782, 2.6556, 1.6398], + device='cuda:3'), covar=tensor([0.3611, 0.1931, 0.4925, 0.3106, 0.1698, 0.2358, 0.1459, 0.4488], + device='cuda:3'), in_proj_covar=tensor([0.0337, 0.0353, 0.0423, 0.0349, 0.0378, 0.0374, 0.0368, 0.0421], + device='cuda:3'), out_proj_covar=tensor([9.9585e-05, 1.0508e-04, 1.2804e-04, 1.0459e-04, 1.1200e-04, 1.1120e-04, + 1.0758e-04, 1.2663e-04], device='cuda:3') +2023-04-28 01:50:35,789 INFO [finetune.py:976] (3/7) Epoch 27, batch 5300, loss[loss=0.1771, simple_loss=0.2521, pruned_loss=0.05107, over 4192.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2434, pruned_loss=0.04712, over 956238.88 frames. ], batch size: 65, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:50:36,526 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154221.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:51:15,586 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 01:51:16,116 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.5287, 1.3726, 1.3879, 1.0250, 1.2379, 1.1609, 1.6701, 1.1860], + device='cuda:3'), covar=tensor([0.3501, 0.1914, 0.5092, 0.2706, 0.1779, 0.2270, 0.1806, 0.5469], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0354, 0.0426, 0.0351, 0.0380, 0.0376, 0.0369, 0.0423], + device='cuda:3'), out_proj_covar=tensor([9.9951e-05, 1.0546e-04, 1.2872e-04, 1.0497e-04, 1.1249e-04, 1.1173e-04, + 1.0801e-04, 1.2722e-04], device='cuda:3') +2023-04-28 01:51:41,568 INFO [finetune.py:976] (3/7) Epoch 27, batch 5350, loss[loss=0.1327, simple_loss=0.2035, pruned_loss=0.03093, over 4817.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2436, pruned_loss=0.04724, over 955813.62 frames. ], batch size: 30, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:51:49,112 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.757e+01 1.510e+02 1.847e+02 2.221e+02 4.452e+02, threshold=3.694e+02, percent-clipped=2.0 +2023-04-28 01:51:52,262 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154279.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:51:54,117 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154282.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 01:51:56,615 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154286.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:00,116 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154291.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:19,625 INFO [finetune.py:976] (3/7) Epoch 27, batch 5400, loss[loss=0.1331, simple_loss=0.2053, pruned_loss=0.03038, over 4912.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2412, pruned_loss=0.04665, over 956732.26 frames. ], batch size: 37, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:52:30,055 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154337.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 01:52:37,666 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154347.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:41,234 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154352.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:52,200 INFO [finetune.py:976] (3/7) Epoch 27, batch 5450, loss[loss=0.1347, simple_loss=0.2048, pruned_loss=0.03233, over 4757.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2373, pruned_loss=0.04541, over 957195.88 frames. ], batch size: 27, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:52:54,630 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.557e+02 1.812e+02 2.063e+02 3.462e+02, threshold=3.624e+02, percent-clipped=0.0 +2023-04-28 01:52:57,817 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7970, 2.0647, 0.7856, 1.1321, 1.3408, 1.1174, 2.1422, 1.2689], + device='cuda:3'), covar=tensor([0.0525, 0.0479, 0.0571, 0.1020, 0.0400, 0.0806, 0.0294, 0.0606], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 01:53:09,466 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154398.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 01:53:16,763 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154405.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 01:53:31,461 INFO [finetune.py:976] (3/7) Epoch 27, batch 5500, loss[loss=0.1415, simple_loss=0.2158, pruned_loss=0.03363, over 4907.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2345, pruned_loss=0.04476, over 954883.99 frames. ], batch size: 36, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:53:44,348 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154432.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 01:53:54,047 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154440.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:54:12,935 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9344, 2.9171, 2.1388, 3.3323, 2.9234, 2.9676, 1.2461, 2.8443], + device='cuda:3'), covar=tensor([0.2237, 0.1609, 0.3594, 0.2958, 0.4044, 0.2151, 0.5715, 0.2952], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0217, 0.0250, 0.0302, 0.0296, 0.0246, 0.0272, 0.0271], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 01:54:26,680 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3963, 1.7600, 1.6871, 2.0288, 2.0291, 2.1773, 1.7367, 3.7158], + device='cuda:3'), covar=tensor([0.0520, 0.0664, 0.0652, 0.1030, 0.0510, 0.0574, 0.0650, 0.0147], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 01:54:35,188 INFO [finetune.py:976] (3/7) Epoch 27, batch 5550, loss[loss=0.1773, simple_loss=0.2693, pruned_loss=0.04266, over 4924.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2377, pruned_loss=0.04608, over 953860.43 frames. ], batch size: 33, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:54:37,625 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.508e+02 1.736e+02 2.106e+02 4.174e+02, threshold=3.472e+02, percent-clipped=1.0 +2023-04-28 01:54:37,743 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154474.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:55:19,706 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-28 01:55:31,960 INFO [finetune.py:976] (3/7) Epoch 27, batch 5600, loss[loss=0.1491, simple_loss=0.2275, pruned_loss=0.03537, over 4819.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2402, pruned_loss=0.04683, over 950176.37 frames. ], batch size: 33, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:55:51,030 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154535.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:56:01,149 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-28 01:56:32,739 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5148, 1.7790, 0.7759, 1.2272, 1.6726, 1.3548, 1.2668, 1.4096], + device='cuda:3'), covar=tensor([0.0489, 0.0354, 0.0342, 0.0547, 0.0267, 0.0496, 0.0502, 0.0566], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 01:56:35,594 INFO [finetune.py:976] (3/7) Epoch 27, batch 5650, loss[loss=0.1956, simple_loss=0.2647, pruned_loss=0.06326, over 4899.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2418, pruned_loss=0.0466, over 951180.21 frames. ], batch size: 36, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:56:42,718 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.767e+01 1.509e+02 1.920e+02 2.295e+02 5.018e+02, threshold=3.840e+02, percent-clipped=4.0 +2023-04-28 01:56:44,488 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154577.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 01:56:45,665 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154579.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:57:36,096 INFO [finetune.py:976] (3/7) Epoch 27, batch 5700, loss[loss=0.1164, simple_loss=0.1781, pruned_loss=0.02737, over 4044.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2367, pruned_loss=0.04588, over 929838.81 frames. ], batch size: 17, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:57:45,549 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154627.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:57:46,196 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154628.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:57:59,105 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154642.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:58:12,011 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154647.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:58:12,561 INFO [finetune.py:976] (3/7) Epoch 28, batch 0, loss[loss=0.1398, simple_loss=0.2246, pruned_loss=0.02747, over 4878.00 frames. ], tot_loss[loss=0.1398, simple_loss=0.2246, pruned_loss=0.02747, over 4878.00 frames. ], batch size: 43, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:58:12,561 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 01:58:17,734 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6458, 1.2676, 1.4549, 1.3767, 1.7864, 1.5143, 1.2784, 1.4436], + device='cuda:3'), covar=tensor([0.1843, 0.1566, 0.1850, 0.1423, 0.0904, 0.1534, 0.1861, 0.2607], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0306, 0.0350, 0.0286, 0.0325, 0.0304, 0.0300, 0.0375], + device='cuda:3'), out_proj_covar=tensor([6.3781e-05, 6.2941e-05, 7.3412e-05, 5.7246e-05, 6.6436e-05, 6.3401e-05, + 6.2023e-05, 7.9490e-05], device='cuda:3') +2023-04-28 01:58:29,430 INFO [finetune.py:1010] (3/7) Epoch 28, validation: loss=0.1549, simple_loss=0.224, pruned_loss=0.04297, over 2265189.00 frames. +2023-04-28 01:58:29,430 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 01:58:33,233 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3278, 2.9227, 1.0037, 1.5472, 2.1644, 1.2338, 3.8342, 1.9951], + device='cuda:3'), covar=tensor([0.0667, 0.0787, 0.0869, 0.1209, 0.0518, 0.0999, 0.0219, 0.0605], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0050, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 01:58:55,769 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.710e+01 1.461e+02 1.801e+02 2.323e+02 7.600e+02, threshold=3.601e+02, percent-clipped=3.0 +2023-04-28 01:59:04,490 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154689.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:06,870 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154693.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 01:59:09,842 INFO [finetune.py:976] (3/7) Epoch 28, batch 50, loss[loss=0.1936, simple_loss=0.2607, pruned_loss=0.06328, over 4847.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2462, pruned_loss=0.04983, over 216497.44 frames. ], batch size: 44, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:59:17,141 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154705.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 01:59:26,298 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154720.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:33,633 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154732.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:38,565 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154740.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:43,317 INFO [finetune.py:976] (3/7) Epoch 28, batch 100, loss[loss=0.1486, simple_loss=0.2192, pruned_loss=0.039, over 4902.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2393, pruned_loss=0.04717, over 380484.75 frames. ], batch size: 36, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:59:48,389 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154753.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:00:02,208 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.686e+01 1.538e+02 1.845e+02 2.159e+02 3.671e+02, threshold=3.690e+02, percent-clipped=1.0 +2023-04-28 02:00:05,349 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154780.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:00:06,015 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154781.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:00:10,228 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154788.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:00:16,273 INFO [finetune.py:976] (3/7) Epoch 28, batch 150, loss[loss=0.1542, simple_loss=0.2273, pruned_loss=0.04062, over 4937.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2345, pruned_loss=0.04614, over 508769.63 frames. ], batch size: 38, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 02:00:17,051 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-28 02:00:38,343 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154830.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:00:49,255 INFO [finetune.py:976] (3/7) Epoch 28, batch 200, loss[loss=0.154, simple_loss=0.2274, pruned_loss=0.04027, over 4905.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2332, pruned_loss=0.04596, over 608685.95 frames. ], batch size: 36, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:01:08,233 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.505e+02 1.793e+02 2.296e+02 3.844e+02, threshold=3.586e+02, percent-clipped=2.0 +2023-04-28 02:01:09,539 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154877.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:01:10,737 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0966, 1.4079, 1.3504, 1.6907, 1.5716, 1.5441, 1.3901, 2.4757], + device='cuda:3'), covar=tensor([0.0641, 0.0843, 0.0810, 0.1280, 0.0642, 0.0489, 0.0752, 0.0221], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 02:01:22,151 INFO [finetune.py:976] (3/7) Epoch 28, batch 250, loss[loss=0.1429, simple_loss=0.2179, pruned_loss=0.03394, over 4780.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2369, pruned_loss=0.0474, over 685101.04 frames. ], batch size: 26, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:01:41,470 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154925.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:01:51,919 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154942.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:01:54,972 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154947.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:00,648 INFO [finetune.py:976] (3/7) Epoch 28, batch 300, loss[loss=0.1813, simple_loss=0.2609, pruned_loss=0.05085, over 4854.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2409, pruned_loss=0.04784, over 743538.17 frames. ], batch size: 44, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:02:36,612 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.547e+02 1.784e+02 2.160e+02 3.859e+02, threshold=3.568e+02, percent-clipped=1.0 +2023-04-28 02:02:37,364 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154976.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:47,500 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154984.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:56,120 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154990.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:57,999 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154993.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:02:59,161 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154995.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:03:06,551 INFO [finetune.py:976] (3/7) Epoch 28, batch 350, loss[loss=0.2387, simple_loss=0.3064, pruned_loss=0.08546, over 4202.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2447, pruned_loss=0.04923, over 787330.78 frames. ], batch size: 65, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:03:29,261 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1062, 1.5674, 1.9904, 2.1836, 1.9383, 1.5629, 1.0893, 1.6834], + device='cuda:3'), covar=tensor([0.2850, 0.2838, 0.1557, 0.1823, 0.2420, 0.2521, 0.4009, 0.1769], + device='cuda:3'), in_proj_covar=tensor([0.0295, 0.0248, 0.0230, 0.0316, 0.0224, 0.0237, 0.0230, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 02:03:33,792 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5367, 1.4030, 4.2578, 4.0258, 3.7188, 4.0426, 4.0114, 3.7509], + device='cuda:3'), covar=tensor([0.6723, 0.5369, 0.1071, 0.1710, 0.1053, 0.1350, 0.1422, 0.1487], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0310, 0.0408, 0.0410, 0.0350, 0.0418, 0.0319, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:03:45,316 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155037.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:03:47,614 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155041.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:03:51,696 INFO [finetune.py:976] (3/7) Epoch 28, batch 400, loss[loss=0.1563, simple_loss=0.2358, pruned_loss=0.03839, over 4795.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2447, pruned_loss=0.04869, over 824439.10 frames. ], batch size: 45, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:04:06,197 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155069.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:11,282 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.524e+02 1.961e+02 2.438e+02 3.962e+02, threshold=3.922e+02, percent-clipped=3.0 +2023-04-28 02:04:11,989 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155076.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:13,369 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-04-28 02:04:25,389 INFO [finetune.py:976] (3/7) Epoch 28, batch 450, loss[loss=0.1372, simple_loss=0.2133, pruned_loss=0.03051, over 4861.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2428, pruned_loss=0.04768, over 853253.84 frames. ], batch size: 34, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:04:48,044 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:48,089 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:57,577 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-28 02:04:58,937 INFO [finetune.py:976] (3/7) Epoch 28, batch 500, loss[loss=0.1746, simple_loss=0.2436, pruned_loss=0.05275, over 4818.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2396, pruned_loss=0.04655, over 876735.89 frames. ], batch size: 33, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:05:17,832 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.583e+02 1.761e+02 2.203e+02 6.801e+02, threshold=3.523e+02, percent-clipped=1.0 +2023-04-28 02:05:20,243 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155178.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:05:32,274 INFO [finetune.py:976] (3/7) Epoch 28, batch 550, loss[loss=0.1568, simple_loss=0.2235, pruned_loss=0.04503, over 4910.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2355, pruned_loss=0.04504, over 895381.71 frames. ], batch size: 35, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:05:42,337 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3394, 2.1733, 2.0156, 1.8907, 2.3799, 1.9065, 2.9248, 1.8173], + device='cuda:3'), covar=tensor([0.3584, 0.1832, 0.4388, 0.3094, 0.1700, 0.2432, 0.1306, 0.4177], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0358, 0.0425, 0.0353, 0.0382, 0.0378, 0.0373, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:05:53,404 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2779, 1.8222, 2.1522, 2.2509, 2.1845, 1.7878, 1.1981, 1.8491], + device='cuda:3'), covar=tensor([0.3175, 0.2875, 0.1532, 0.2082, 0.2227, 0.2482, 0.3952, 0.1881], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0247, 0.0230, 0.0316, 0.0223, 0.0236, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 02:05:56,884 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3821, 2.9741, 2.4449, 2.8197, 1.9304, 2.4402, 2.6692, 2.2743], + device='cuda:3'), covar=tensor([0.2149, 0.1227, 0.0982, 0.1230, 0.3773, 0.1185, 0.1939, 0.2509], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0304, 0.0217, 0.0276, 0.0314, 0.0255, 0.0249, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1316e-04, 1.1979e-04, 8.5319e-05, 1.0878e-04, 1.2647e-04, 1.0021e-04, + 1.0045e-04, 1.0378e-04], device='cuda:3') +2023-04-28 02:06:05,933 INFO [finetune.py:976] (3/7) Epoch 28, batch 600, loss[loss=0.1699, simple_loss=0.2372, pruned_loss=0.05123, over 4910.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2377, pruned_loss=0.04659, over 907617.54 frames. ], batch size: 43, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:06:23,751 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.531e+01 1.578e+02 2.066e+02 2.357e+02 4.358e+02, threshold=4.132e+02, percent-clipped=3.0 +2023-04-28 02:06:30,725 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155284.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:06:39,099 INFO [finetune.py:976] (3/7) Epoch 28, batch 650, loss[loss=0.1299, simple_loss=0.1945, pruned_loss=0.03269, over 4791.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2394, pruned_loss=0.04643, over 916895.83 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:07:01,142 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8955, 1.1359, 1.6917, 1.7815, 1.7217, 1.8309, 1.7350, 1.7311], + device='cuda:3'), covar=tensor([0.3006, 0.4292, 0.3514, 0.3458, 0.4515, 0.5909, 0.3319, 0.3421], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0376, 0.0331, 0.0342, 0.0351, 0.0394, 0.0362, 0.0334], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 02:07:02,743 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155332.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:07:02,753 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155332.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:07:12,885 INFO [finetune.py:976] (3/7) Epoch 28, batch 700, loss[loss=0.116, simple_loss=0.1869, pruned_loss=0.02258, over 4741.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2407, pruned_loss=0.04681, over 926473.03 frames. ], batch size: 23, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:07:34,635 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.574e+02 1.852e+02 2.166e+02 4.284e+02, threshold=3.703e+02, percent-clipped=1.0 +2023-04-28 02:07:40,183 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155376.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:07:51,753 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9470, 1.3691, 4.8703, 4.6026, 4.1784, 4.7091, 4.3763, 4.2588], + device='cuda:3'), covar=tensor([0.7001, 0.6089, 0.1000, 0.1649, 0.1199, 0.1239, 0.1494, 0.1665], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0309, 0.0407, 0.0408, 0.0349, 0.0417, 0.0317, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:08:09,445 INFO [finetune.py:976] (3/7) Epoch 28, batch 750, loss[loss=0.1489, simple_loss=0.2261, pruned_loss=0.03588, over 4919.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2421, pruned_loss=0.04757, over 934492.47 frames. ], batch size: 33, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:08:41,160 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155424.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:08:41,839 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155425.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:08:56,441 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6239, 1.7213, 0.7196, 1.3222, 1.7947, 1.5036, 1.3929, 1.4460], + device='cuda:3'), covar=tensor([0.0487, 0.0345, 0.0346, 0.0540, 0.0266, 0.0489, 0.0475, 0.0565], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0052], + device='cuda:3') +2023-04-28 02:09:12,229 INFO [finetune.py:976] (3/7) Epoch 28, batch 800, loss[loss=0.1531, simple_loss=0.2369, pruned_loss=0.0346, over 4793.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2425, pruned_loss=0.04711, over 940442.89 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:09:23,004 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155457.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:09:36,895 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-28 02:09:42,707 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.932e+01 1.490e+02 1.749e+02 2.216e+02 3.602e+02, threshold=3.499e+02, percent-clipped=0.0 +2023-04-28 02:09:59,777 INFO [finetune.py:976] (3/7) Epoch 28, batch 850, loss[loss=0.1707, simple_loss=0.2397, pruned_loss=0.05086, over 4818.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.241, pruned_loss=0.04663, over 946502.73 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:10:12,007 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155518.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:10:21,074 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7424, 1.0663, 1.7114, 2.1572, 1.8209, 1.7274, 1.7569, 1.7760], + device='cuda:3'), covar=tensor([0.4542, 0.6638, 0.5933, 0.5764, 0.5765, 0.7607, 0.7482, 0.8218], + device='cuda:3'), in_proj_covar=tensor([0.0442, 0.0423, 0.0518, 0.0507, 0.0471, 0.0508, 0.0509, 0.0524], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:10:27,083 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3024, 1.9161, 2.1310, 2.6895, 2.1807, 1.7813, 1.7157, 2.0831], + device='cuda:3'), covar=tensor([0.2451, 0.2543, 0.1435, 0.1564, 0.2190, 0.2232, 0.3673, 0.1759], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0247, 0.0230, 0.0316, 0.0223, 0.0236, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 02:10:28,300 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7280, 2.2523, 2.5845, 3.3371, 2.5177, 2.0061, 2.1495, 2.5828], + device='cuda:3'), covar=tensor([0.3231, 0.3178, 0.1584, 0.2244, 0.2700, 0.2672, 0.3548, 0.1855], + device='cuda:3'), in_proj_covar=tensor([0.0294, 0.0247, 0.0230, 0.0316, 0.0223, 0.0236, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 02:10:32,817 INFO [finetune.py:976] (3/7) Epoch 28, batch 900, loss[loss=0.2014, simple_loss=0.259, pruned_loss=0.07193, over 4826.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2388, pruned_loss=0.04641, over 947355.61 frames. ], batch size: 51, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:10:47,340 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7591, 1.0780, 1.7429, 2.1554, 1.8348, 1.7090, 1.7426, 1.7332], + device='cuda:3'), covar=tensor([0.4197, 0.6397, 0.6128, 0.5458, 0.5634, 0.7825, 0.7336, 0.8979], + device='cuda:3'), in_proj_covar=tensor([0.0441, 0.0422, 0.0517, 0.0506, 0.0470, 0.0506, 0.0508, 0.0522], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:10:49,546 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.500e+01 1.455e+02 1.735e+02 2.204e+02 5.182e+02, threshold=3.469e+02, percent-clipped=3.0 +2023-04-28 02:10:53,745 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-28 02:11:05,998 INFO [finetune.py:976] (3/7) Epoch 28, batch 950, loss[loss=0.2146, simple_loss=0.2795, pruned_loss=0.07482, over 4912.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2379, pruned_loss=0.04641, over 948452.39 frames. ], batch size: 46, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:11:10,954 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155605.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:11:23,101 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9553, 1.7166, 1.6836, 1.4658, 1.8547, 1.5229, 2.2795, 1.4448], + device='cuda:3'), covar=tensor([0.3145, 0.1827, 0.4587, 0.2224, 0.1420, 0.2057, 0.1448, 0.4756], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0359, 0.0428, 0.0353, 0.0384, 0.0380, 0.0374, 0.0428], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:11:27,350 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155632.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:11:39,450 INFO [finetune.py:976] (3/7) Epoch 28, batch 1000, loss[loss=0.1158, simple_loss=0.1901, pruned_loss=0.02077, over 4739.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2417, pruned_loss=0.04779, over 949232.51 frames. ], batch size: 23, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:11:45,544 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155657.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:11:51,059 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155666.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:11:56,469 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.942e+01 1.547e+02 1.789e+02 2.038e+02 3.472e+02, threshold=3.578e+02, percent-clipped=1.0 +2023-04-28 02:11:59,554 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155680.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:12:12,447 INFO [finetune.py:976] (3/7) Epoch 28, batch 1050, loss[loss=0.2033, simple_loss=0.275, pruned_loss=0.06583, over 4829.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2444, pruned_loss=0.04802, over 950240.24 frames. ], batch size: 47, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:12:25,716 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155718.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:12:28,224 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 02:12:29,959 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155725.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:12:33,916 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-28 02:12:43,019 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0512, 2.7019, 2.0769, 2.2115, 1.4334, 1.4622, 2.1769, 1.4285], + device='cuda:3'), covar=tensor([0.1689, 0.1355, 0.1287, 0.1616, 0.2178, 0.1976, 0.0903, 0.2010], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0210, 0.0171, 0.0206, 0.0201, 0.0187, 0.0157, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 02:12:45,201 INFO [finetune.py:976] (3/7) Epoch 28, batch 1100, loss[loss=0.1651, simple_loss=0.2475, pruned_loss=0.04132, over 4897.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2445, pruned_loss=0.048, over 951241.28 frames. ], batch size: 37, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:12:48,608 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-28 02:13:01,284 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3967, 2.2021, 1.8246, 1.9776, 2.3116, 1.7870, 2.6613, 1.6856], + device='cuda:3'), covar=tensor([0.3617, 0.2000, 0.4451, 0.2910, 0.1600, 0.2623, 0.1590, 0.4116], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0355, 0.0423, 0.0350, 0.0380, 0.0376, 0.0370, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:13:13,292 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155773.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:13:19,855 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.609e+02 1.883e+02 2.305e+02 3.828e+02, threshold=3.767e+02, percent-clipped=2.0 +2023-04-28 02:13:30,940 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155784.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:13:33,203 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-04-28 02:13:46,058 INFO [finetune.py:976] (3/7) Epoch 28, batch 1150, loss[loss=0.163, simple_loss=0.2438, pruned_loss=0.04115, over 4757.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2446, pruned_loss=0.04778, over 954221.32 frames. ], batch size: 54, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:13:49,637 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 02:13:58,502 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155813.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:14:13,221 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2733, 2.1874, 1.8483, 1.8888, 2.2119, 1.8271, 2.6069, 1.5861], + device='cuda:3'), covar=tensor([0.3430, 0.1574, 0.4227, 0.2493, 0.1526, 0.2252, 0.1235, 0.4737], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0356, 0.0426, 0.0352, 0.0381, 0.0377, 0.0371, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:14:17,555 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-28 02:14:18,091 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155845.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:14:19,790 INFO [finetune.py:976] (3/7) Epoch 28, batch 1200, loss[loss=0.1856, simple_loss=0.2536, pruned_loss=0.05877, over 4900.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2418, pruned_loss=0.04697, over 954102.01 frames. ], batch size: 37, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:14:54,753 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.563e+02 1.837e+02 2.158e+02 4.161e+02, threshold=3.673e+02, percent-clipped=1.0 +2023-04-28 02:15:05,569 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7667, 1.2262, 1.2899, 1.4943, 1.8059, 1.4896, 1.3491, 1.2563], + device='cuda:3'), covar=tensor([0.1865, 0.2106, 0.2028, 0.1539, 0.1158, 0.2132, 0.2220, 0.2492], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0307, 0.0350, 0.0285, 0.0325, 0.0306, 0.0301, 0.0376], + device='cuda:3'), out_proj_covar=tensor([6.3992e-05, 6.3096e-05, 7.3472e-05, 5.7137e-05, 6.6300e-05, 6.3609e-05, + 6.2290e-05, 7.9569e-05], device='cuda:3') +2023-04-28 02:15:25,827 INFO [finetune.py:976] (3/7) Epoch 28, batch 1250, loss[loss=0.1287, simple_loss=0.2022, pruned_loss=0.02764, over 4834.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.239, pruned_loss=0.04628, over 952862.79 frames. ], batch size: 30, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:15:48,274 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9712, 1.3935, 4.9859, 4.7202, 4.2269, 4.7801, 4.4065, 4.4016], + device='cuda:3'), covar=tensor([0.7013, 0.6066, 0.1028, 0.1652, 0.1144, 0.1452, 0.1799, 0.1691], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0311, 0.0409, 0.0410, 0.0352, 0.0419, 0.0320, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:16:33,390 INFO [finetune.py:976] (3/7) Epoch 28, batch 1300, loss[loss=0.1874, simple_loss=0.2601, pruned_loss=0.05735, over 4933.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2375, pruned_loss=0.04616, over 953825.36 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:16:53,065 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155961.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:17:07,930 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.397e+01 1.571e+02 1.820e+02 2.327e+02 4.182e+02, threshold=3.640e+02, percent-clipped=2.0 +2023-04-28 02:17:35,611 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9231, 1.2061, 1.5919, 1.6862, 1.6494, 1.7144, 1.6346, 1.6185], + device='cuda:3'), covar=tensor([0.3594, 0.4529, 0.4005, 0.4095, 0.4917, 0.6490, 0.4297, 0.3902], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0374, 0.0329, 0.0340, 0.0348, 0.0390, 0.0359, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 02:17:37,316 INFO [finetune.py:976] (3/7) Epoch 28, batch 1350, loss[loss=0.1809, simple_loss=0.257, pruned_loss=0.05236, over 4822.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2381, pruned_loss=0.04631, over 955478.25 frames. ], batch size: 40, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:17:38,003 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3505, 3.3322, 2.4581, 3.8167, 3.3591, 3.3148, 1.4054, 3.3258], + device='cuda:3'), covar=tensor([0.2090, 0.1477, 0.3449, 0.2418, 0.4497, 0.2011, 0.6110, 0.2750], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0220, 0.0252, 0.0305, 0.0300, 0.0250, 0.0275, 0.0273], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 02:17:56,780 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156013.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:18:38,553 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4566, 3.0031, 1.0126, 1.6502, 1.7764, 2.2429, 1.7113, 0.9715], + device='cuda:3'), covar=tensor([0.1308, 0.1030, 0.1790, 0.1272, 0.1016, 0.0924, 0.1507, 0.1994], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0239, 0.0135, 0.0121, 0.0131, 0.0153, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 02:18:39,734 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6985, 1.6997, 4.6702, 4.4101, 4.0385, 4.4307, 4.3280, 4.1688], + device='cuda:3'), covar=tensor([0.6723, 0.4957, 0.1027, 0.1525, 0.1112, 0.1707, 0.0993, 0.1385], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0310, 0.0409, 0.0409, 0.0350, 0.0418, 0.0319, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:18:40,868 INFO [finetune.py:976] (3/7) Epoch 28, batch 1400, loss[loss=0.168, simple_loss=0.2457, pruned_loss=0.0451, over 4868.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.242, pruned_loss=0.04754, over 954960.30 frames. ], batch size: 34, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:19:12,180 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156070.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:19:20,862 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.674e+02 1.966e+02 2.354e+02 4.293e+02, threshold=3.933e+02, percent-clipped=2.0 +2023-04-28 02:19:36,469 INFO [finetune.py:976] (3/7) Epoch 28, batch 1450, loss[loss=0.1641, simple_loss=0.2341, pruned_loss=0.04708, over 4732.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.242, pruned_loss=0.04728, over 954725.25 frames. ], batch size: 27, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:19:40,218 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9616, 2.6513, 1.9783, 1.8807, 1.4245, 1.4540, 2.0617, 1.3482], + device='cuda:3'), covar=tensor([0.1653, 0.1188, 0.1306, 0.1700, 0.2263, 0.1908, 0.0933, 0.2133], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0209, 0.0170, 0.0205, 0.0201, 0.0186, 0.0156, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 02:19:46,179 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156113.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:19:59,741 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156131.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 02:20:05,154 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156140.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:20:09,927 INFO [finetune.py:976] (3/7) Epoch 28, batch 1500, loss[loss=0.2028, simple_loss=0.275, pruned_loss=0.06529, over 4738.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2443, pruned_loss=0.04812, over 954894.05 frames. ], batch size: 54, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:20:11,276 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-28 02:20:23,250 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156161.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:20:43,692 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.591e+02 1.859e+02 2.303e+02 4.450e+02, threshold=3.717e+02, percent-clipped=1.0 +2023-04-28 02:21:08,784 INFO [finetune.py:976] (3/7) Epoch 28, batch 1550, loss[loss=0.1867, simple_loss=0.2602, pruned_loss=0.0566, over 4886.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2426, pruned_loss=0.04764, over 952304.09 frames. ], batch size: 43, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:21:16,244 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-28 02:21:31,143 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156216.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:04,805 INFO [finetune.py:976] (3/7) Epoch 28, batch 1600, loss[loss=0.1223, simple_loss=0.186, pruned_loss=0.02934, over 4740.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2409, pruned_loss=0.04746, over 953444.06 frames. ], batch size: 23, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:22:12,844 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156261.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:24,229 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.624e+02 1.857e+02 2.215e+02 3.679e+02, threshold=3.713e+02, percent-clipped=0.0 +2023-04-28 02:22:25,609 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156277.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:34,041 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156290.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:38,824 INFO [finetune.py:976] (3/7) Epoch 28, batch 1650, loss[loss=0.164, simple_loss=0.2303, pruned_loss=0.04886, over 4909.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2371, pruned_loss=0.04583, over 955961.01 frames. ], batch size: 36, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:22:44,438 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9626, 2.6320, 2.0020, 2.0241, 1.4271, 1.5049, 2.0840, 1.4061], + device='cuda:3'), covar=tensor([0.1551, 0.1197, 0.1233, 0.1498, 0.2129, 0.1800, 0.0867, 0.1899], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0208, 0.0169, 0.0203, 0.0200, 0.0185, 0.0155, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 02:22:45,575 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156309.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:48,029 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156313.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:23:12,277 INFO [finetune.py:976] (3/7) Epoch 28, batch 1700, loss[loss=0.1728, simple_loss=0.2493, pruned_loss=0.04819, over 4735.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2355, pruned_loss=0.04506, over 956963.51 frames. ], batch size: 54, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:23:14,245 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156351.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:23:20,257 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156361.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:23:28,661 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.471e+01 1.503e+02 1.857e+02 2.212e+02 3.931e+02, threshold=3.714e+02, percent-clipped=1.0 +2023-04-28 02:23:45,144 INFO [finetune.py:976] (3/7) Epoch 28, batch 1750, loss[loss=0.19, simple_loss=0.2521, pruned_loss=0.0639, over 4829.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2373, pruned_loss=0.04577, over 956968.70 frames. ], batch size: 33, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:23:53,084 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-04-28 02:24:06,973 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-28 02:24:08,541 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156426.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:24:29,103 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156440.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:24:29,404 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-28 02:24:39,117 INFO [finetune.py:976] (3/7) Epoch 28, batch 1800, loss[loss=0.2377, simple_loss=0.3158, pruned_loss=0.07984, over 4860.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2415, pruned_loss=0.04702, over 959355.44 frames. ], batch size: 44, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:25:11,111 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.353e+01 1.610e+02 1.909e+02 2.260e+02 4.171e+02, threshold=3.817e+02, percent-clipped=3.0 +2023-04-28 02:25:31,043 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156488.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:25:42,454 INFO [finetune.py:976] (3/7) Epoch 28, batch 1850, loss[loss=0.1759, simple_loss=0.2432, pruned_loss=0.0543, over 4883.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2433, pruned_loss=0.04786, over 959365.55 frames. ], batch size: 35, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:26:48,596 INFO [finetune.py:976] (3/7) Epoch 28, batch 1900, loss[loss=0.1448, simple_loss=0.2303, pruned_loss=0.02967, over 4745.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2457, pruned_loss=0.04866, over 958115.82 frames. ], batch size: 54, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:27:01,544 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-28 02:27:20,593 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156572.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:27:22,349 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.575e+02 1.976e+02 2.372e+02 4.648e+02, threshold=3.953e+02, percent-clipped=2.0 +2023-04-28 02:27:52,842 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-04-28 02:27:54,786 INFO [finetune.py:976] (3/7) Epoch 28, batch 1950, loss[loss=0.199, simple_loss=0.2599, pruned_loss=0.06904, over 4827.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2437, pruned_loss=0.04775, over 957025.36 frames. ], batch size: 33, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:28:05,420 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7454, 1.4902, 2.0032, 2.0571, 1.5430, 1.4276, 1.6237, 0.9937], + device='cuda:3'), covar=tensor([0.0487, 0.0692, 0.0371, 0.0509, 0.0680, 0.1155, 0.0576, 0.0620], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0070, 0.0075, 0.0095, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 02:28:13,334 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4043, 1.0857, 1.2007, 1.1870, 1.5264, 1.2452, 1.1297, 1.1437], + device='cuda:3'), covar=tensor([0.1827, 0.1284, 0.1688, 0.1355, 0.0913, 0.1575, 0.1759, 0.2257], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0305, 0.0350, 0.0284, 0.0323, 0.0304, 0.0299, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.4010e-05, 6.2521e-05, 7.3368e-05, 5.6751e-05, 6.5794e-05, 6.3221e-05, + 6.1757e-05, 7.9085e-05], device='cuda:3') +2023-04-28 02:29:00,504 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156646.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:29:01,645 INFO [finetune.py:976] (3/7) Epoch 28, batch 2000, loss[loss=0.1828, simple_loss=0.2451, pruned_loss=0.0603, over 4929.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2414, pruned_loss=0.0473, over 957183.17 frames. ], batch size: 33, lr: 2.88e-03, grad_scale: 64.0 +2023-04-28 02:29:34,750 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.509e+02 1.833e+02 2.149e+02 4.980e+02, threshold=3.665e+02, percent-clipped=1.0 +2023-04-28 02:29:51,373 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-04-28 02:30:06,048 INFO [finetune.py:976] (3/7) Epoch 28, batch 2050, loss[loss=0.1568, simple_loss=0.2375, pruned_loss=0.03812, over 4753.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2383, pruned_loss=0.04666, over 956709.37 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 64.0 +2023-04-28 02:30:43,851 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156726.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 02:31:05,852 INFO [finetune.py:976] (3/7) Epoch 28, batch 2100, loss[loss=0.1364, simple_loss=0.2186, pruned_loss=0.02711, over 4794.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2374, pruned_loss=0.04608, over 955165.12 frames. ], batch size: 29, lr: 2.88e-03, grad_scale: 64.0 +2023-04-28 02:31:07,099 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156749.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:31:38,248 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-28 02:31:39,408 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156774.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:31:39,962 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.904e+01 1.566e+02 1.862e+02 2.081e+02 4.387e+02, threshold=3.725e+02, percent-clipped=1.0 +2023-04-28 02:32:10,054 INFO [finetune.py:976] (3/7) Epoch 28, batch 2150, loss[loss=0.179, simple_loss=0.2643, pruned_loss=0.04689, over 4764.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.24, pruned_loss=0.04627, over 956483.54 frames. ], batch size: 28, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:32:27,881 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156810.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:32:42,818 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-28 02:32:52,486 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-04-28 02:32:57,254 INFO [finetune.py:976] (3/7) Epoch 28, batch 2200, loss[loss=0.1717, simple_loss=0.2464, pruned_loss=0.04851, over 4859.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2412, pruned_loss=0.04683, over 955022.66 frames. ], batch size: 34, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:33:15,059 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156872.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:33:17,875 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.564e+02 1.773e+02 2.100e+02 3.301e+02, threshold=3.547e+02, percent-clipped=0.0 +2023-04-28 02:33:23,528 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156885.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:33:31,444 INFO [finetune.py:976] (3/7) Epoch 28, batch 2250, loss[loss=0.1439, simple_loss=0.2179, pruned_loss=0.03492, over 4925.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2424, pruned_loss=0.04722, over 955400.92 frames. ], batch size: 33, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:33:47,682 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-28 02:33:47,975 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156920.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:33:49,607 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-28 02:34:04,148 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:04,170 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:05,295 INFO [finetune.py:976] (3/7) Epoch 28, batch 2300, loss[loss=0.1672, simple_loss=0.2418, pruned_loss=0.0463, over 4803.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2421, pruned_loss=0.04627, over 955492.89 frames. ], batch size: 25, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:34:25,202 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.469e+02 1.728e+02 2.151e+02 3.571e+02, threshold=3.456e+02, percent-clipped=1.0 +2023-04-28 02:34:36,692 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156994.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:39,084 INFO [finetune.py:976] (3/7) Epoch 28, batch 2350, loss[loss=0.1467, simple_loss=0.215, pruned_loss=0.03918, over 4778.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2402, pruned_loss=0.04589, over 957272.78 frames. ], batch size: 28, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:34:47,991 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157010.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:35:23,596 INFO [finetune.py:976] (3/7) Epoch 28, batch 2400, loss[loss=0.1301, simple_loss=0.2034, pruned_loss=0.02837, over 4729.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2375, pruned_loss=0.04532, over 956604.37 frames. ], batch size: 59, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:35:49,030 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5231, 1.7881, 2.0158, 2.0888, 2.0219, 1.9702, 2.0026, 2.0564], + device='cuda:3'), covar=tensor([0.3643, 0.4977, 0.3797, 0.4324, 0.4731, 0.6435, 0.4573, 0.4061], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0374, 0.0328, 0.0340, 0.0351, 0.0392, 0.0360, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 02:35:50,214 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157071.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:35:53,120 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.553e+02 1.893e+02 2.191e+02 4.408e+02, threshold=3.786e+02, percent-clipped=4.0 +2023-04-28 02:36:07,063 INFO [finetune.py:976] (3/7) Epoch 28, batch 2450, loss[loss=0.1968, simple_loss=0.265, pruned_loss=0.06429, over 4894.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2337, pruned_loss=0.04379, over 958004.80 frames. ], batch size: 32, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:36:11,410 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157105.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:36:12,684 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6207, 1.6422, 0.8224, 1.3624, 1.7537, 1.5096, 1.4280, 1.4790], + device='cuda:3'), covar=tensor([0.0472, 0.0361, 0.0326, 0.0520, 0.0280, 0.0485, 0.0454, 0.0536], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 02:36:58,723 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157142.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:37:07,638 INFO [finetune.py:976] (3/7) Epoch 28, batch 2500, loss[loss=0.1807, simple_loss=0.2571, pruned_loss=0.05213, over 4897.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2357, pruned_loss=0.04486, over 955425.72 frames. ], batch size: 32, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:37:44,331 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.608e+02 1.980e+02 2.407e+02 6.066e+02, threshold=3.960e+02, percent-clipped=5.0 +2023-04-28 02:38:00,173 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3487, 1.2190, 3.8661, 3.6822, 3.4435, 3.7911, 3.7683, 3.4476], + device='cuda:3'), covar=tensor([0.7402, 0.5772, 0.1164, 0.1536, 0.1077, 0.2195, 0.1232, 0.1617], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0309, 0.0409, 0.0408, 0.0349, 0.0417, 0.0319, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:38:14,042 INFO [finetune.py:976] (3/7) Epoch 28, batch 2550, loss[loss=0.2048, simple_loss=0.2979, pruned_loss=0.05589, over 4833.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2397, pruned_loss=0.04644, over 954104.98 frames. ], batch size: 47, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:38:22,905 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157203.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:39:11,687 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157241.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:39:16,446 INFO [finetune.py:976] (3/7) Epoch 28, batch 2600, loss[loss=0.1608, simple_loss=0.2371, pruned_loss=0.04223, over 4766.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2414, pruned_loss=0.04705, over 954491.92 frames. ], batch size: 28, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:39:34,408 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1621, 1.4862, 1.3354, 1.6453, 1.5170, 1.6977, 1.3487, 2.9542], + device='cuda:3'), covar=tensor([0.0624, 0.0809, 0.0790, 0.1179, 0.0615, 0.0491, 0.0727, 0.0178], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 02:39:50,659 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.630e+02 1.851e+02 2.303e+02 4.332e+02, threshold=3.701e+02, percent-clipped=1.0 +2023-04-28 02:40:05,057 INFO [finetune.py:976] (3/7) Epoch 28, batch 2650, loss[loss=0.1259, simple_loss=0.2001, pruned_loss=0.02588, over 4772.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2418, pruned_loss=0.04662, over 956936.03 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:40:21,902 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8315, 1.4340, 1.9882, 2.3553, 1.9881, 1.8259, 1.9137, 1.8872], + device='cuda:3'), covar=tensor([0.4491, 0.6884, 0.6218, 0.5290, 0.5706, 0.7809, 0.8098, 0.8710], + device='cuda:3'), in_proj_covar=tensor([0.0446, 0.0425, 0.0521, 0.0508, 0.0474, 0.0511, 0.0511, 0.0527], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:40:53,194 INFO [finetune.py:976] (3/7) Epoch 28, batch 2700, loss[loss=0.1255, simple_loss=0.205, pruned_loss=0.02304, over 4892.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2403, pruned_loss=0.04602, over 956540.87 frames. ], batch size: 35, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:40:58,713 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157356.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:04,718 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157366.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:10,643 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.425e+02 1.670e+02 2.111e+02 3.242e+02, threshold=3.340e+02, percent-clipped=0.0 +2023-04-28 02:41:15,936 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157382.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:26,099 INFO [finetune.py:976] (3/7) Epoch 28, batch 2750, loss[loss=0.207, simple_loss=0.2603, pruned_loss=0.07688, over 4226.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2379, pruned_loss=0.04547, over 956186.67 frames. ], batch size: 65, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:41:30,005 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157404.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:41:31,101 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157405.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:31,732 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8707, 2.4561, 1.8215, 1.8299, 1.3582, 1.3952, 1.9706, 1.2815], + device='cuda:3'), covar=tensor([0.1561, 0.1083, 0.1261, 0.1449, 0.2135, 0.1812, 0.0900, 0.2008], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0209, 0.0170, 0.0205, 0.0201, 0.0187, 0.0157, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 02:41:33,081 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-04-28 02:41:38,348 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157417.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:42,047 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3216, 1.7953, 2.2138, 2.7562, 2.1533, 1.7160, 1.6005, 2.0217], + device='cuda:3'), covar=tensor([0.2619, 0.2790, 0.1483, 0.1755, 0.2392, 0.2459, 0.3725, 0.1814], + device='cuda:3'), in_proj_covar=tensor([0.0291, 0.0245, 0.0229, 0.0314, 0.0222, 0.0235, 0.0228, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 02:41:46,766 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6494, 3.5400, 2.6713, 4.2028, 3.6047, 3.6497, 1.7296, 3.6300], + device='cuda:3'), covar=tensor([0.1626, 0.1389, 0.3477, 0.1722, 0.3332, 0.1957, 0.5188, 0.2261], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0221, 0.0252, 0.0304, 0.0300, 0.0248, 0.0275, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 02:42:02,288 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157443.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:42:05,180 INFO [finetune.py:976] (3/7) Epoch 28, batch 2800, loss[loss=0.18, simple_loss=0.2404, pruned_loss=0.05986, over 4282.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2348, pruned_loss=0.0448, over 956213.64 frames. ], batch size: 65, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:42:13,772 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157453.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:42:13,825 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7609, 1.8028, 0.8422, 1.4772, 1.8972, 1.6386, 1.4976, 1.5983], + device='cuda:3'), covar=tensor([0.0467, 0.0356, 0.0325, 0.0526, 0.0244, 0.0496, 0.0500, 0.0523], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 02:42:17,620 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-04-28 02:42:26,578 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157465.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:42:38,115 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.553e+02 1.775e+02 2.169e+02 3.521e+02, threshold=3.549e+02, percent-clipped=1.0 +2023-04-28 02:42:58,312 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6912, 1.4118, 1.8090, 1.9390, 1.5841, 1.4424, 1.5831, 1.0347], + device='cuda:3'), covar=tensor([0.0450, 0.0713, 0.0393, 0.0506, 0.0626, 0.0959, 0.0505, 0.0527], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0068, 0.0065, 0.0069, 0.0075, 0.0095, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 02:43:07,287 INFO [finetune.py:976] (3/7) Epoch 28, batch 2850, loss[loss=0.1536, simple_loss=0.2314, pruned_loss=0.03795, over 4893.00 frames. ], tot_loss[loss=0.1603, simple_loss=0.233, pruned_loss=0.0438, over 956997.12 frames. ], batch size: 35, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:43:07,363 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157498.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:43:17,204 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157506.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:44:00,026 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157541.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:44:09,939 INFO [finetune.py:976] (3/7) Epoch 28, batch 2900, loss[loss=0.1761, simple_loss=0.2536, pruned_loss=0.04927, over 4918.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2371, pruned_loss=0.04517, over 955003.38 frames. ], batch size: 36, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:44:33,704 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157567.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:44:34,286 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157568.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:44:44,343 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.234e+02 1.564e+02 1.889e+02 2.208e+02 3.535e+02, threshold=3.777e+02, percent-clipped=0.0 +2023-04-28 02:44:52,206 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8790, 2.8510, 2.1512, 3.2986, 2.8198, 2.8563, 1.3291, 2.8176], + device='cuda:3'), covar=tensor([0.2117, 0.1748, 0.3660, 0.2842, 0.4212, 0.2159, 0.5357, 0.3171], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0220, 0.0251, 0.0303, 0.0299, 0.0247, 0.0274, 0.0272], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 02:45:03,636 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157589.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:45:14,533 INFO [finetune.py:976] (3/7) Epoch 28, batch 2950, loss[loss=0.1486, simple_loss=0.2342, pruned_loss=0.03151, over 4811.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.24, pruned_loss=0.04617, over 955097.87 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:45:18,067 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-28 02:45:50,082 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157629.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:46:10,745 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.7489, 1.6895, 1.7149, 1.3526, 1.7588, 1.5309, 2.2866, 1.5372], + device='cuda:3'), covar=tensor([0.3690, 0.2010, 0.4782, 0.3060, 0.1586, 0.2541, 0.1401, 0.4481], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0357, 0.0427, 0.0352, 0.0385, 0.0380, 0.0374, 0.0426], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:46:18,333 INFO [finetune.py:976] (3/7) Epoch 28, batch 3000, loss[loss=0.1734, simple_loss=0.2497, pruned_loss=0.04858, over 4819.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2412, pruned_loss=0.04676, over 953468.14 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:46:18,333 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 02:46:34,886 INFO [finetune.py:1010] (3/7) Epoch 28, validation: loss=0.153, simple_loss=0.2217, pruned_loss=0.04213, over 2265189.00 frames. +2023-04-28 02:46:34,887 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 02:46:46,700 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1973, 3.2401, 2.7482, 3.0118, 3.3107, 3.0296, 4.1202, 2.6522], + device='cuda:3'), covar=tensor([0.2910, 0.1456, 0.2901, 0.2423, 0.1336, 0.1852, 0.0919, 0.2894], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0356, 0.0426, 0.0351, 0.0383, 0.0379, 0.0373, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:46:49,099 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157666.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:46:55,060 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.557e+02 1.829e+02 2.170e+02 4.324e+02, threshold=3.658e+02, percent-clipped=1.0 +2023-04-28 02:47:00,639 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-28 02:47:00,853 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-04-28 02:47:09,009 INFO [finetune.py:976] (3/7) Epoch 28, batch 3050, loss[loss=0.1403, simple_loss=0.2236, pruned_loss=0.02856, over 4926.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2431, pruned_loss=0.0473, over 953245.36 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:47:27,242 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157712.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:47:28,440 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157714.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:47:58,457 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157738.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:48:09,325 INFO [finetune.py:976] (3/7) Epoch 28, batch 3100, loss[loss=0.14, simple_loss=0.22, pruned_loss=0.03001, over 4762.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2412, pruned_loss=0.04642, over 955432.27 frames. ], batch size: 27, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:48:22,593 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157760.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 02:48:32,272 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.835e+01 1.560e+02 1.859e+02 2.154e+02 3.848e+02, threshold=3.719e+02, percent-clipped=2.0 +2023-04-28 02:48:46,288 INFO [finetune.py:976] (3/7) Epoch 28, batch 3150, loss[loss=0.1902, simple_loss=0.2518, pruned_loss=0.06427, over 4795.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.238, pruned_loss=0.04534, over 955985.05 frames. ], batch size: 51, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:48:46,380 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157798.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:49:18,159 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157846.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:49:19,357 INFO [finetune.py:976] (3/7) Epoch 28, batch 3200, loss[loss=0.1158, simple_loss=0.1959, pruned_loss=0.01779, over 4788.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2358, pruned_loss=0.04516, over 956146.96 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:49:22,526 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157853.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:49:27,983 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157862.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 02:49:38,913 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.560e+02 1.811e+02 2.192e+02 5.644e+02, threshold=3.622e+02, percent-clipped=2.0 +2023-04-28 02:50:07,745 INFO [finetune.py:976] (3/7) Epoch 28, batch 3250, loss[loss=0.1318, simple_loss=0.2078, pruned_loss=0.02786, over 4720.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2369, pruned_loss=0.04577, over 956167.60 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:50:29,710 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157914.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:50:43,192 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157924.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:51:14,743 INFO [finetune.py:976] (3/7) Epoch 28, batch 3300, loss[loss=0.1591, simple_loss=0.2412, pruned_loss=0.0385, over 4818.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2381, pruned_loss=0.04559, over 953692.76 frames. ], batch size: 30, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:51:51,135 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.721e+02 1.967e+02 2.277e+02 5.584e+02, threshold=3.933e+02, percent-clipped=3.0 +2023-04-28 02:52:10,302 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3240, 3.3439, 2.6803, 3.9059, 3.2785, 3.3599, 1.8907, 3.3397], + device='cuda:3'), covar=tensor([0.1823, 0.1353, 0.4132, 0.1835, 0.3031, 0.1863, 0.4666, 0.2376], + device='cuda:3'), in_proj_covar=tensor([0.0243, 0.0217, 0.0250, 0.0301, 0.0296, 0.0246, 0.0272, 0.0269], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 02:52:20,397 INFO [finetune.py:976] (3/7) Epoch 28, batch 3350, loss[loss=0.1792, simple_loss=0.2622, pruned_loss=0.04813, over 4814.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2406, pruned_loss=0.04662, over 952888.10 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:52:41,471 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158012.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:53:03,025 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4445, 1.3967, 4.0733, 3.7849, 3.5719, 3.8833, 3.8518, 3.5725], + device='cuda:3'), covar=tensor([0.7245, 0.5459, 0.1114, 0.1886, 0.1183, 0.1289, 0.1465, 0.1705], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0314, 0.0410, 0.0411, 0.0354, 0.0421, 0.0322, 0.0368], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:53:12,138 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158038.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:53:23,377 INFO [finetune.py:976] (3/7) Epoch 28, batch 3400, loss[loss=0.1682, simple_loss=0.2357, pruned_loss=0.05035, over 4855.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2409, pruned_loss=0.04657, over 954485.78 frames. ], batch size: 31, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:53:41,205 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158060.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:53:41,249 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158060.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:53:56,202 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.114e+02 1.582e+02 1.900e+02 2.246e+02 3.787e+02, threshold=3.800e+02, percent-clipped=0.0 +2023-04-28 02:54:10,639 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158086.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:54:23,031 INFO [finetune.py:976] (3/7) Epoch 28, batch 3450, loss[loss=0.1524, simple_loss=0.2277, pruned_loss=0.03859, over 4179.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2405, pruned_loss=0.04616, over 954623.41 frames. ], batch size: 66, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:54:34,533 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158108.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:54:54,852 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158122.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:55:29,823 INFO [finetune.py:976] (3/7) Epoch 28, batch 3500, loss[loss=0.155, simple_loss=0.2271, pruned_loss=0.04141, over 4935.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.238, pruned_loss=0.04551, over 955815.48 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:55:40,242 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-28 02:55:48,573 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158162.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:55:59,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8869, 1.3045, 3.2750, 3.0568, 2.9512, 3.2257, 3.1800, 2.8812], + device='cuda:3'), covar=tensor([0.7272, 0.5033, 0.1502, 0.2257, 0.1422, 0.1959, 0.1718, 0.1944], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0311, 0.0407, 0.0408, 0.0351, 0.0417, 0.0319, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:56:07,658 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.474e+02 1.724e+02 2.129e+02 5.829e+02, threshold=3.447e+02, percent-clipped=1.0 +2023-04-28 02:56:12,597 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158183.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:56:16,570 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7804, 1.6478, 0.8795, 1.4549, 1.9003, 1.6826, 1.5121, 1.5559], + device='cuda:3'), covar=tensor([0.0463, 0.0355, 0.0311, 0.0521, 0.0248, 0.0468, 0.0456, 0.0531], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0051, 0.0052], + device='cuda:3') +2023-04-28 02:56:16,616 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8577, 1.4178, 1.9555, 2.3684, 1.9494, 1.8147, 1.9068, 1.7959], + device='cuda:3'), covar=tensor([0.4356, 0.7106, 0.5964, 0.5350, 0.5934, 0.7683, 0.8085, 0.9943], + device='cuda:3'), in_proj_covar=tensor([0.0445, 0.0425, 0.0521, 0.0508, 0.0474, 0.0511, 0.0512, 0.0526], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:56:23,134 INFO [finetune.py:976] (3/7) Epoch 28, batch 3550, loss[loss=0.1455, simple_loss=0.2094, pruned_loss=0.0408, over 4756.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2365, pruned_loss=0.04544, over 955365.79 frames. ], batch size: 54, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:56:29,701 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-04-28 02:56:30,181 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158209.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:56:30,777 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158210.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:56:39,862 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158224.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:56:46,957 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9796, 1.8234, 1.6833, 1.4453, 1.8288, 1.5789, 2.1663, 1.4622], + device='cuda:3'), covar=tensor([0.2914, 0.1550, 0.3640, 0.2197, 0.1432, 0.1826, 0.1526, 0.4093], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0358, 0.0427, 0.0355, 0.0385, 0.0380, 0.0374, 0.0426], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:56:56,634 INFO [finetune.py:976] (3/7) Epoch 28, batch 3600, loss[loss=0.1978, simple_loss=0.2666, pruned_loss=0.06449, over 4833.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2349, pruned_loss=0.04531, over 956435.90 frames. ], batch size: 40, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:57:11,762 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158272.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:57:14,138 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.946e+01 1.583e+02 1.935e+02 2.237e+02 3.659e+02, threshold=3.870e+02, percent-clipped=1.0 +2023-04-28 02:57:29,419 INFO [finetune.py:976] (3/7) Epoch 28, batch 3650, loss[loss=0.18, simple_loss=0.2606, pruned_loss=0.04967, over 4831.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2371, pruned_loss=0.046, over 957110.96 frames. ], batch size: 47, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:58:02,906 INFO [finetune.py:976] (3/7) Epoch 28, batch 3700, loss[loss=0.1563, simple_loss=0.233, pruned_loss=0.03981, over 4811.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2388, pruned_loss=0.04573, over 955686.89 frames. ], batch size: 45, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:58:19,866 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.537e+02 1.915e+02 2.211e+02 4.327e+02, threshold=3.830e+02, percent-clipped=1.0 +2023-04-28 02:58:25,397 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158384.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:58:35,209 INFO [finetune.py:976] (3/7) Epoch 28, batch 3750, loss[loss=0.1689, simple_loss=0.2356, pruned_loss=0.05115, over 4779.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2414, pruned_loss=0.04707, over 954585.96 frames. ], batch size: 26, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:58:44,946 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9444, 1.8529, 1.6865, 1.5049, 1.9891, 1.6360, 2.5047, 1.5840], + device='cuda:3'), covar=tensor([0.3939, 0.2283, 0.4705, 0.3133, 0.1969, 0.2724, 0.1566, 0.4622], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0356, 0.0424, 0.0352, 0.0383, 0.0378, 0.0373, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:58:52,185 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0839, 1.7507, 1.9980, 2.3056, 2.2932, 1.8886, 1.8798, 2.1072], + device='cuda:3'), covar=tensor([0.0703, 0.1089, 0.0650, 0.0544, 0.0601, 0.0849, 0.0573, 0.0502], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0185, 0.0171, 0.0178, 0.0177, 0.0150, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 02:58:52,396 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-28 02:59:05,527 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158445.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:59:07,665 INFO [finetune.py:976] (3/7) Epoch 28, batch 3800, loss[loss=0.2171, simple_loss=0.2761, pruned_loss=0.07907, over 4819.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2429, pruned_loss=0.04748, over 955804.86 frames. ], batch size: 30, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:59:09,410 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3930, 3.1070, 1.1638, 1.6340, 2.6797, 1.6244, 4.4482, 2.1814], + device='cuda:3'), covar=tensor([0.0681, 0.0665, 0.0833, 0.1247, 0.0470, 0.0994, 0.0214, 0.0595], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 02:59:31,092 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-28 02:59:42,092 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.322e+01 1.536e+02 1.803e+02 2.140e+02 3.917e+02, threshold=3.606e+02, percent-clipped=1.0 +2023-04-28 02:59:43,423 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158478.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:00:04,618 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 03:00:07,899 INFO [finetune.py:976] (3/7) Epoch 28, batch 3850, loss[loss=0.1268, simple_loss=0.2021, pruned_loss=0.02573, over 4874.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2409, pruned_loss=0.04652, over 954979.79 frames. ], batch size: 31, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:00:26,829 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158509.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:00:46,797 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8123, 1.6823, 2.2522, 2.2773, 1.6941, 1.5280, 1.8449, 1.0704], + device='cuda:3'), covar=tensor([0.0611, 0.0789, 0.0425, 0.0884, 0.0739, 0.1063, 0.0631, 0.0650], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 03:01:10,984 INFO [finetune.py:976] (3/7) Epoch 28, batch 3900, loss[loss=0.1701, simple_loss=0.2352, pruned_loss=0.05246, over 4891.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2385, pruned_loss=0.04605, over 953141.97 frames. ], batch size: 32, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:01:28,078 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158557.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:01:51,819 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.216e+01 1.537e+02 1.820e+02 2.187e+02 3.711e+02, threshold=3.639e+02, percent-clipped=1.0 +2023-04-28 03:02:23,132 INFO [finetune.py:976] (3/7) Epoch 28, batch 3950, loss[loss=0.1812, simple_loss=0.2404, pruned_loss=0.06101, over 4866.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2359, pruned_loss=0.04563, over 953091.27 frames. ], batch size: 34, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:03:31,310 INFO [finetune.py:976] (3/7) Epoch 28, batch 4000, loss[loss=0.1736, simple_loss=0.2346, pruned_loss=0.05632, over 4882.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2365, pruned_loss=0.04652, over 953647.57 frames. ], batch size: 32, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:04:13,593 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.149e+01 1.505e+02 1.815e+02 2.168e+02 4.885e+02, threshold=3.630e+02, percent-clipped=2.0 +2023-04-28 03:04:37,743 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 03:04:38,152 INFO [finetune.py:976] (3/7) Epoch 28, batch 4050, loss[loss=0.1574, simple_loss=0.229, pruned_loss=0.04288, over 4789.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2385, pruned_loss=0.04731, over 952813.40 frames. ], batch size: 25, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:05:33,920 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158740.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:05:43,859 INFO [finetune.py:976] (3/7) Epoch 28, batch 4100, loss[loss=0.1201, simple_loss=0.1907, pruned_loss=0.02477, over 4719.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2408, pruned_loss=0.04728, over 952610.55 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:06:05,223 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.3523, 3.4426, 2.4756, 3.9019, 3.4210, 3.3407, 1.6091, 3.3099], + device='cuda:3'), covar=tensor([0.2019, 0.1287, 0.3486, 0.2213, 0.3463, 0.2071, 0.5552, 0.2694], + device='cuda:3'), in_proj_covar=tensor([0.0247, 0.0221, 0.0253, 0.0305, 0.0302, 0.0251, 0.0278, 0.0276], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 03:06:25,664 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.629e+02 1.870e+02 2.384e+02 6.257e+02, threshold=3.741e+02, percent-clipped=1.0 +2023-04-28 03:06:26,415 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158777.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:06:26,740 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-04-28 03:06:27,003 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158778.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:06:35,231 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158783.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:06:37,068 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4595, 1.8112, 1.6241, 2.3160, 2.4314, 1.9973, 1.9302, 1.7367], + device='cuda:3'), covar=tensor([0.1973, 0.1785, 0.1955, 0.1673, 0.1250, 0.1898, 0.2111, 0.2453], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0306, 0.0349, 0.0285, 0.0324, 0.0304, 0.0299, 0.0374], + device='cuda:3'), out_proj_covar=tensor([6.4284e-05, 6.2707e-05, 7.3116e-05, 5.6954e-05, 6.6088e-05, 6.3236e-05, + 6.1769e-05, 7.9164e-05], device='cuda:3') +2023-04-28 03:06:49,753 INFO [finetune.py:976] (3/7) Epoch 28, batch 4150, loss[loss=0.1843, simple_loss=0.2619, pruned_loss=0.0533, over 4734.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2426, pruned_loss=0.04786, over 953642.93 frames. ], batch size: 27, lr: 2.87e-03, grad_scale: 64.0 +2023-04-28 03:07:30,856 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158826.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:07:32,731 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5168, 1.4863, 4.2924, 4.0017, 3.7431, 4.1110, 4.0736, 3.8395], + device='cuda:3'), covar=tensor([0.7280, 0.5455, 0.1013, 0.1639, 0.1187, 0.1599, 0.1457, 0.1444], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0309, 0.0404, 0.0406, 0.0350, 0.0416, 0.0318, 0.0363], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:07:41,988 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158835.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:07:43,876 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158838.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:07:52,631 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158844.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:07:54,973 INFO [finetune.py:976] (3/7) Epoch 28, batch 4200, loss[loss=0.1524, simple_loss=0.2242, pruned_loss=0.04034, over 4166.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2425, pruned_loss=0.04698, over 954950.38 frames. ], batch size: 65, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:08:11,160 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.4725, 3.6015, 2.6698, 4.1046, 3.5327, 3.4930, 1.6749, 3.4493], + device='cuda:3'), covar=tensor([0.1900, 0.1220, 0.3206, 0.1901, 0.2667, 0.2032, 0.5743, 0.2545], + device='cuda:3'), in_proj_covar=tensor([0.0246, 0.0220, 0.0252, 0.0304, 0.0301, 0.0251, 0.0277, 0.0274], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 03:08:37,325 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.525e+02 1.688e+02 1.998e+02 3.338e+02, threshold=3.377e+02, percent-clipped=0.0 +2023-04-28 03:08:54,591 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5378, 1.3572, 0.4908, 1.2642, 1.4352, 1.4166, 1.3287, 1.3534], + device='cuda:3'), covar=tensor([0.0480, 0.0366, 0.0400, 0.0542, 0.0296, 0.0477, 0.0449, 0.0541], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 03:08:56,472 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3253, 1.5465, 1.8611, 1.9386, 1.8945, 1.9586, 1.8783, 1.8518], + device='cuda:3'), covar=tensor([0.3542, 0.4968, 0.4268, 0.4197, 0.4969, 0.6430, 0.4733, 0.4436], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0376, 0.0329, 0.0342, 0.0351, 0.0392, 0.0361, 0.0334], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 03:09:04,983 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158896.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:09:06,083 INFO [finetune.py:976] (3/7) Epoch 28, batch 4250, loss[loss=0.1337, simple_loss=0.2138, pruned_loss=0.02679, over 4755.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2402, pruned_loss=0.04584, over 956224.27 frames. ], batch size: 26, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:09:15,890 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9072, 1.4874, 2.0167, 2.3469, 1.9966, 1.8765, 1.9486, 1.8675], + device='cuda:3'), covar=tensor([0.4512, 0.6768, 0.6049, 0.5497, 0.5723, 0.7878, 0.8218, 0.9279], + device='cuda:3'), in_proj_covar=tensor([0.0446, 0.0425, 0.0522, 0.0510, 0.0475, 0.0511, 0.0513, 0.0528], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:10:00,120 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.65 vs. limit=5.0 +2023-04-28 03:10:12,866 INFO [finetune.py:976] (3/7) Epoch 28, batch 4300, loss[loss=0.1424, simple_loss=0.2146, pruned_loss=0.03512, over 4912.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2372, pruned_loss=0.04516, over 956109.56 frames. ], batch size: 43, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:10:36,575 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-28 03:10:48,758 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.467e+01 1.425e+02 1.689e+02 1.969e+02 4.397e+02, threshold=3.377e+02, percent-clipped=1.0 +2023-04-28 03:11:18,966 INFO [finetune.py:976] (3/7) Epoch 28, batch 4350, loss[loss=0.1659, simple_loss=0.2362, pruned_loss=0.04776, over 4227.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2347, pruned_loss=0.04446, over 956532.35 frames. ], batch size: 65, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:12:15,673 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159040.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:12:26,355 INFO [finetune.py:976] (3/7) Epoch 28, batch 4400, loss[loss=0.1559, simple_loss=0.2362, pruned_loss=0.03782, over 4695.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2354, pruned_loss=0.04487, over 952934.43 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:12:34,183 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2036, 1.4259, 1.2503, 1.5887, 1.5306, 1.5999, 1.3248, 2.9396], + device='cuda:3'), covar=tensor([0.0651, 0.0831, 0.0820, 0.1222, 0.0641, 0.0583, 0.0772, 0.0169], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 03:12:54,459 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159069.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:12:55,078 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8637, 1.3292, 1.5419, 2.2030, 2.3342, 1.8893, 1.6635, 1.9179], + device='cuda:3'), covar=tensor([0.0942, 0.1755, 0.1167, 0.0601, 0.0613, 0.0928, 0.0848, 0.0675], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0184, 0.0170, 0.0178, 0.0177, 0.0150, 0.0175], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:13:05,081 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.784e+01 1.599e+02 1.852e+02 2.200e+02 4.625e+02, threshold=3.703e+02, percent-clipped=1.0 +2023-04-28 03:13:08,314 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-28 03:13:19,038 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159088.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:13:29,732 INFO [finetune.py:976] (3/7) Epoch 28, batch 4450, loss[loss=0.1603, simple_loss=0.2124, pruned_loss=0.05412, over 4036.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2381, pruned_loss=0.04577, over 951044.08 frames. ], batch size: 17, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:13:49,843 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4650, 0.9884, 0.4479, 1.1821, 1.1011, 1.3632, 1.2747, 1.2329], + device='cuda:3'), covar=tensor([0.0499, 0.0421, 0.0393, 0.0586, 0.0318, 0.0511, 0.0497, 0.0578], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 03:14:09,975 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159130.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:14:11,692 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1348, 2.6637, 2.4066, 2.3444, 1.5552, 1.6071, 2.5401, 1.7059], + device='cuda:3'), covar=tensor([0.1622, 0.1485, 0.1246, 0.1507, 0.2184, 0.1864, 0.0814, 0.1883], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0210, 0.0170, 0.0205, 0.0201, 0.0186, 0.0157, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 03:14:12,212 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159133.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:14:19,203 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-28 03:14:21,841 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159139.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:14:38,437 INFO [finetune.py:976] (3/7) Epoch 28, batch 4500, loss[loss=0.1734, simple_loss=0.2418, pruned_loss=0.0525, over 4869.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2407, pruned_loss=0.04683, over 952852.41 frames. ], batch size: 34, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:14:40,017 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-28 03:14:49,771 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159159.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:15:11,105 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.553e+02 1.824e+02 2.169e+02 4.146e+02, threshold=3.648e+02, percent-clipped=1.0 +2023-04-28 03:15:23,329 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-28 03:15:31,031 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159191.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 03:15:41,715 INFO [finetune.py:976] (3/7) Epoch 28, batch 4550, loss[loss=0.1664, simple_loss=0.2432, pruned_loss=0.04478, over 4733.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2407, pruned_loss=0.04608, over 954416.93 frames. ], batch size: 27, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:16:04,932 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.7467, 3.7805, 2.7237, 4.3989, 3.8617, 3.7903, 1.5343, 3.7658], + device='cuda:3'), covar=tensor([0.1593, 0.1327, 0.3065, 0.1544, 0.2601, 0.1723, 0.6027, 0.2373], + device='cuda:3'), in_proj_covar=tensor([0.0248, 0.0222, 0.0254, 0.0307, 0.0303, 0.0253, 0.0279, 0.0277], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 03:16:05,614 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159220.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:16:45,688 INFO [finetune.py:976] (3/7) Epoch 28, batch 4600, loss[loss=0.1505, simple_loss=0.2223, pruned_loss=0.0393, over 4915.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2406, pruned_loss=0.04609, over 953045.37 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:16:56,975 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4353, 2.3329, 2.6504, 2.8567, 2.8103, 2.3834, 2.1351, 2.5877], + device='cuda:3'), covar=tensor([0.0896, 0.1072, 0.0668, 0.0618, 0.0640, 0.0876, 0.0730, 0.0598], + device='cuda:3'), in_proj_covar=tensor([0.0182, 0.0202, 0.0182, 0.0169, 0.0177, 0.0176, 0.0149, 0.0174], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:17:15,990 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-28 03:17:18,838 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.827e+01 1.517e+02 1.769e+02 2.322e+02 3.935e+02, threshold=3.539e+02, percent-clipped=1.0 +2023-04-28 03:17:38,396 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159290.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:17:49,023 INFO [finetune.py:976] (3/7) Epoch 28, batch 4650, loss[loss=0.1328, simple_loss=0.2087, pruned_loss=0.02843, over 4757.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2385, pruned_loss=0.04554, over 952021.52 frames. ], batch size: 28, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:18:08,865 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6863, 2.1437, 2.1869, 2.2539, 2.1094, 2.1583, 2.3088, 2.2414], + device='cuda:3'), covar=tensor([0.3558, 0.4866, 0.4248, 0.4113, 0.5069, 0.6270, 0.4390, 0.4117], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0378, 0.0331, 0.0344, 0.0353, 0.0393, 0.0363, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 03:18:54,509 INFO [finetune.py:976] (3/7) Epoch 28, batch 4700, loss[loss=0.1366, simple_loss=0.2094, pruned_loss=0.03193, over 4905.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2358, pruned_loss=0.04453, over 954043.19 frames. ], batch size: 37, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:19:03,240 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159351.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:19:36,242 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.039e+02 1.471e+02 1.707e+02 1.992e+02 4.348e+02, threshold=3.414e+02, percent-clipped=1.0 +2023-04-28 03:20:00,615 INFO [finetune.py:976] (3/7) Epoch 28, batch 4750, loss[loss=0.1872, simple_loss=0.2723, pruned_loss=0.05106, over 4845.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2351, pruned_loss=0.04469, over 955181.09 frames. ], batch size: 47, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:20:41,283 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159425.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:20:51,586 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159433.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:20:55,260 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159439.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:21:06,689 INFO [finetune.py:976] (3/7) Epoch 28, batch 4800, loss[loss=0.1654, simple_loss=0.2414, pruned_loss=0.04472, over 4899.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2381, pruned_loss=0.04574, over 953507.26 frames. ], batch size: 36, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:21:25,685 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0043, 1.8707, 2.1072, 2.5440, 2.4641, 2.1365, 1.8488, 2.2474], + device='cuda:3'), covar=tensor([0.0916, 0.1077, 0.0676, 0.0500, 0.0570, 0.0758, 0.0676, 0.0547], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0203, 0.0183, 0.0170, 0.0179, 0.0177, 0.0150, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:21:40,929 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5199, 1.4296, 1.8909, 1.8779, 1.3716, 1.2843, 1.4756, 0.8868], + device='cuda:3'), covar=tensor([0.0502, 0.0540, 0.0318, 0.0477, 0.0698, 0.1062, 0.0523, 0.0577], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 03:21:47,685 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-28 03:21:48,625 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.679e+02 2.036e+02 2.333e+02 4.105e+02, threshold=4.072e+02, percent-clipped=2.0 +2023-04-28 03:21:51,198 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159481.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:21:59,480 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159487.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:22:07,518 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159491.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:22:12,296 INFO [finetune.py:976] (3/7) Epoch 28, batch 4850, loss[loss=0.1519, simple_loss=0.2415, pruned_loss=0.03118, over 4770.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2407, pruned_loss=0.04646, over 952708.98 frames. ], batch size: 26, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:22:40,593 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159515.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:22:43,670 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4362, 1.7471, 1.6207, 2.2899, 2.4187, 1.9730, 1.8944, 1.6748], + device='cuda:3'), covar=tensor([0.1532, 0.1825, 0.1860, 0.1354, 0.1188, 0.1944, 0.1961, 0.2215], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0305, 0.0347, 0.0285, 0.0322, 0.0303, 0.0298, 0.0371], + device='cuda:3'), out_proj_covar=tensor([6.3616e-05, 6.2656e-05, 7.2590e-05, 5.7001e-05, 6.5722e-05, 6.3002e-05, + 6.1452e-05, 7.8292e-05], device='cuda:3') +2023-04-28 03:23:05,857 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159539.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:23:16,522 INFO [finetune.py:976] (3/7) Epoch 28, batch 4900, loss[loss=0.1558, simple_loss=0.2346, pruned_loss=0.03849, over 4816.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2428, pruned_loss=0.04713, over 952357.93 frames. ], batch size: 51, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:23:56,944 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.586e+02 1.930e+02 2.270e+02 5.838e+02, threshold=3.860e+02, percent-clipped=1.0 +2023-04-28 03:24:20,243 INFO [finetune.py:976] (3/7) Epoch 28, batch 4950, loss[loss=0.1635, simple_loss=0.2394, pruned_loss=0.04383, over 4897.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2434, pruned_loss=0.04728, over 952479.41 frames. ], batch size: 37, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:25:08,569 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159631.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:25:21,913 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-04-28 03:25:23,013 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159646.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:25:24,156 INFO [finetune.py:976] (3/7) Epoch 28, batch 5000, loss[loss=0.1813, simple_loss=0.2494, pruned_loss=0.05655, over 4899.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2414, pruned_loss=0.04671, over 953489.65 frames. ], batch size: 43, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:25:24,267 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4086, 1.4711, 1.3436, 1.6610, 1.6284, 1.7805, 1.4361, 3.0492], + device='cuda:3'), covar=tensor([0.0557, 0.0768, 0.0767, 0.1239, 0.0619, 0.0506, 0.0720, 0.0159], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 03:25:55,258 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6570, 2.0743, 1.8532, 2.0592, 1.6342, 1.8310, 1.7204, 1.4432], + device='cuda:3'), covar=tensor([0.1807, 0.1083, 0.0797, 0.1110, 0.3154, 0.0959, 0.1695, 0.2193], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0301, 0.0217, 0.0275, 0.0313, 0.0252, 0.0247, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1247e-04, 1.1850e-04, 8.5338e-05, 1.0800e-04, 1.2631e-04, 9.9109e-05, + 9.9501e-05, 1.0345e-04], device='cuda:3') +2023-04-28 03:26:05,240 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.569e+02 1.814e+02 2.278e+02 4.708e+02, threshold=3.628e+02, percent-clipped=1.0 +2023-04-28 03:26:24,454 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159692.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:26:32,518 INFO [finetune.py:976] (3/7) Epoch 28, batch 5050, loss[loss=0.1517, simple_loss=0.2176, pruned_loss=0.04284, over 4804.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2392, pruned_loss=0.04679, over 954058.21 frames. ], batch size: 51, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:27:06,779 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159725.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:27:35,600 INFO [finetune.py:976] (3/7) Epoch 28, batch 5100, loss[loss=0.1708, simple_loss=0.244, pruned_loss=0.04876, over 4904.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.236, pruned_loss=0.04576, over 953326.94 frames. ], batch size: 46, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:27:38,413 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-28 03:28:09,664 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159773.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:28:13,003 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.052e+01 1.474e+02 1.729e+02 2.168e+02 4.974e+02, threshold=3.458e+02, percent-clipped=1.0 +2023-04-28 03:28:41,783 INFO [finetune.py:976] (3/7) Epoch 28, batch 5150, loss[loss=0.1722, simple_loss=0.25, pruned_loss=0.04724, over 4903.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2362, pruned_loss=0.04577, over 953381.17 frames. ], batch size: 43, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:29:03,528 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159815.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:29:45,762 INFO [finetune.py:976] (3/7) Epoch 28, batch 5200, loss[loss=0.169, simple_loss=0.2453, pruned_loss=0.04633, over 4009.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2399, pruned_loss=0.04667, over 953821.60 frames. ], batch size: 65, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:29:55,502 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159855.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:30:05,394 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159863.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:30:15,182 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-04-28 03:30:26,727 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.682e+02 1.955e+02 2.433e+02 4.346e+02, threshold=3.909e+02, percent-clipped=3.0 +2023-04-28 03:30:51,955 INFO [finetune.py:976] (3/7) Epoch 28, batch 5250, loss[loss=0.1467, simple_loss=0.2236, pruned_loss=0.03496, over 4800.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2409, pruned_loss=0.04713, over 950766.51 frames. ], batch size: 51, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:31:09,938 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159910.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:31:18,998 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159916.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:31:55,447 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159946.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:32:02,065 INFO [finetune.py:976] (3/7) Epoch 28, batch 5300, loss[loss=0.1626, simple_loss=0.233, pruned_loss=0.04613, over 4760.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2428, pruned_loss=0.0477, over 949012.07 frames. ], batch size: 26, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:32:26,787 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159971.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:32:35,449 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.404e+01 1.544e+02 1.829e+02 2.287e+02 5.054e+02, threshold=3.659e+02, percent-clipped=1.0 +2023-04-28 03:32:48,731 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159987.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:32:58,790 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159994.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:33:06,691 INFO [finetune.py:976] (3/7) Epoch 28, batch 5350, loss[loss=0.1549, simple_loss=0.242, pruned_loss=0.03389, over 4903.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2437, pruned_loss=0.04791, over 950617.93 frames. ], batch size: 37, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:33:16,416 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1278, 1.3484, 1.6205, 1.7521, 1.6383, 1.7122, 1.6625, 1.6491], + device='cuda:3'), covar=tensor([0.4038, 0.4796, 0.4182, 0.4041, 0.5231, 0.6960, 0.4265, 0.4259], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0377, 0.0331, 0.0344, 0.0353, 0.0394, 0.0364, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 03:34:13,400 INFO [finetune.py:976] (3/7) Epoch 28, batch 5400, loss[loss=0.182, simple_loss=0.2433, pruned_loss=0.06031, over 4892.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.241, pruned_loss=0.04724, over 951122.61 frames. ], batch size: 35, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:34:24,957 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8521, 1.4182, 1.9258, 2.3137, 1.9498, 1.8041, 1.8870, 1.8028], + device='cuda:3'), covar=tensor([0.4290, 0.7088, 0.6221, 0.5340, 0.5757, 0.7805, 0.8014, 1.0392], + device='cuda:3'), in_proj_covar=tensor([0.0445, 0.0425, 0.0520, 0.0509, 0.0474, 0.0510, 0.0512, 0.0527], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:34:46,562 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.246e+02 1.579e+02 1.912e+02 2.220e+02 4.270e+02, threshold=3.825e+02, percent-clipped=1.0 +2023-04-28 03:35:18,038 INFO [finetune.py:976] (3/7) Epoch 28, batch 5450, loss[loss=0.1467, simple_loss=0.2225, pruned_loss=0.03548, over 4772.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2377, pruned_loss=0.04626, over 952585.48 frames. ], batch size: 27, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:36:22,773 INFO [finetune.py:976] (3/7) Epoch 28, batch 5500, loss[loss=0.1773, simple_loss=0.2515, pruned_loss=0.05157, over 4936.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2347, pruned_loss=0.04483, over 955331.49 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:37:00,973 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 7.993e+01 1.507e+02 1.787e+02 2.170e+02 4.329e+02, threshold=3.573e+02, percent-clipped=2.0 +2023-04-28 03:37:01,132 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0841, 2.5744, 2.0495, 1.9045, 1.4376, 1.5384, 2.1368, 1.4109], + device='cuda:3'), covar=tensor([0.1591, 0.1200, 0.1379, 0.1571, 0.2297, 0.1827, 0.0920, 0.1995], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0208, 0.0169, 0.0204, 0.0200, 0.0185, 0.0156, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 03:37:04,205 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0430, 1.5647, 1.8451, 1.7306, 1.8442, 1.5345, 0.8682, 1.5053], + device='cuda:3'), covar=tensor([0.3115, 0.3192, 0.1671, 0.2083, 0.2344, 0.2678, 0.4116, 0.1969], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0246, 0.0227, 0.0313, 0.0222, 0.0234, 0.0228, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 03:37:27,649 INFO [finetune.py:976] (3/7) Epoch 28, batch 5550, loss[loss=0.1851, simple_loss=0.2683, pruned_loss=0.05097, over 4824.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2369, pruned_loss=0.04544, over 952594.63 frames. ], batch size: 40, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:37:41,111 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160211.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:38:03,243 INFO [finetune.py:976] (3/7) Epoch 28, batch 5600, loss[loss=0.2166, simple_loss=0.2786, pruned_loss=0.07732, over 4822.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2387, pruned_loss=0.0454, over 953024.54 frames. ], batch size: 39, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:38:13,711 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160266.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:38:17,925 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3780, 1.9645, 2.2869, 3.0078, 2.2885, 1.8076, 1.9017, 2.1701], + device='cuda:3'), covar=tensor([0.3029, 0.2826, 0.1425, 0.2029, 0.2618, 0.2363, 0.3343, 0.1986], + device='cuda:3'), in_proj_covar=tensor([0.0290, 0.0245, 0.0226, 0.0312, 0.0222, 0.0234, 0.0227, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 03:38:20,147 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.566e+02 2.005e+02 2.303e+02 3.818e+02, threshold=4.010e+02, percent-clipped=2.0 +2023-04-28 03:38:24,572 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-28 03:38:26,535 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160287.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:38:32,858 INFO [finetune.py:976] (3/7) Epoch 28, batch 5650, loss[loss=0.204, simple_loss=0.2753, pruned_loss=0.06634, over 4783.00 frames. ], tot_loss[loss=0.166, simple_loss=0.241, pruned_loss=0.04553, over 953872.55 frames. ], batch size: 54, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:38:38,222 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160307.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 03:38:54,998 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160335.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:39:02,717 INFO [finetune.py:976] (3/7) Epoch 28, batch 5700, loss[loss=0.153, simple_loss=0.2061, pruned_loss=0.04998, over 4383.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2369, pruned_loss=0.04537, over 934480.02 frames. ], batch size: 19, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:39:14,601 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160368.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 03:39:31,533 INFO [finetune.py:976] (3/7) Epoch 29, batch 0, loss[loss=0.173, simple_loss=0.2484, pruned_loss=0.04882, over 4863.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2484, pruned_loss=0.04882, over 4863.00 frames. ], batch size: 34, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:39:31,534 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 03:39:42,780 INFO [finetune.py:1010] (3/7) Epoch 29, validation: loss=0.1546, simple_loss=0.2236, pruned_loss=0.04278, over 2265189.00 frames. +2023-04-28 03:39:42,780 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 03:39:44,401 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.451e+01 1.469e+02 1.726e+02 2.023e+02 3.272e+02, threshold=3.452e+02, percent-clipped=0.0 +2023-04-28 03:39:45,705 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4647, 3.4077, 0.8565, 1.9635, 1.8266, 2.4478, 2.0282, 0.9934], + device='cuda:3'), covar=tensor([0.1320, 0.0911, 0.1878, 0.1054, 0.1003, 0.0907, 0.1294, 0.1992], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0120, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 03:40:16,116 INFO [finetune.py:976] (3/7) Epoch 29, batch 50, loss[loss=0.1834, simple_loss=0.2679, pruned_loss=0.04947, over 4730.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2415, pruned_loss=0.04567, over 215393.11 frames. ], batch size: 59, lr: 2.85e-03, grad_scale: 32.0 +2023-04-28 03:40:21,801 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 03:40:33,629 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-04-28 03:40:45,250 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-28 03:41:16,377 INFO [finetune.py:976] (3/7) Epoch 29, batch 100, loss[loss=0.1371, simple_loss=0.2128, pruned_loss=0.03073, over 4746.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2349, pruned_loss=0.04428, over 379057.99 frames. ], batch size: 27, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:41:24,294 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.575e+02 1.942e+02 2.371e+02 3.324e+02, threshold=3.883e+02, percent-clipped=0.0 +2023-04-28 03:42:07,347 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160511.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:42:22,985 INFO [finetune.py:976] (3/7) Epoch 29, batch 150, loss[loss=0.1659, simple_loss=0.2356, pruned_loss=0.04803, over 4759.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2334, pruned_loss=0.04453, over 506262.84 frames. ], batch size: 59, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:42:43,464 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9513, 1.4451, 1.7745, 1.7311, 1.7551, 1.4398, 0.7784, 1.4366], + device='cuda:3'), covar=tensor([0.3174, 0.3246, 0.1712, 0.1996, 0.2374, 0.2570, 0.4180, 0.1994], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0228, 0.0315, 0.0223, 0.0236, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 03:43:05,314 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160559.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:43:14,855 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160566.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:43:27,103 INFO [finetune.py:976] (3/7) Epoch 29, batch 200, loss[loss=0.1885, simple_loss=0.2518, pruned_loss=0.06257, over 4900.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2332, pruned_loss=0.0446, over 607025.51 frames. ], batch size: 43, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:43:34,237 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.562e+02 1.851e+02 2.230e+02 3.985e+02, threshold=3.702e+02, percent-clipped=1.0 +2023-04-28 03:43:37,298 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9355, 4.3079, 1.0107, 2.3635, 2.4463, 2.8853, 2.4747, 1.1266], + device='cuda:3'), covar=tensor([0.1394, 0.0841, 0.1967, 0.1128, 0.0995, 0.1032, 0.1384, 0.2140], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0135, 0.0119, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 03:44:17,861 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160614.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:44:30,031 INFO [finetune.py:976] (3/7) Epoch 29, batch 250, loss[loss=0.1501, simple_loss=0.2209, pruned_loss=0.0396, over 4762.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2357, pruned_loss=0.04569, over 684349.69 frames. ], batch size: 28, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:45:20,094 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160663.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 03:45:32,745 INFO [finetune.py:976] (3/7) Epoch 29, batch 300, loss[loss=0.1726, simple_loss=0.2449, pruned_loss=0.05019, over 4806.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2392, pruned_loss=0.04609, over 744142.67 frames. ], batch size: 41, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:45:38,926 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.518e+02 1.888e+02 2.303e+02 4.692e+02, threshold=3.776e+02, percent-clipped=1.0 +2023-04-28 03:46:01,451 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-28 03:46:15,442 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 03:46:37,390 INFO [finetune.py:976] (3/7) Epoch 29, batch 350, loss[loss=0.1841, simple_loss=0.2533, pruned_loss=0.05744, over 4922.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.242, pruned_loss=0.04734, over 788537.79 frames. ], batch size: 42, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:47:08,174 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2061, 2.0637, 2.2720, 2.6298, 2.6166, 2.1474, 2.0028, 2.4313], + device='cuda:3'), covar=tensor([0.0785, 0.1076, 0.0684, 0.0600, 0.0649, 0.0838, 0.0673, 0.0512], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0205, 0.0184, 0.0171, 0.0178, 0.0177, 0.0151, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:47:35,778 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 03:47:45,909 INFO [finetune.py:976] (3/7) Epoch 29, batch 400, loss[loss=0.1651, simple_loss=0.2451, pruned_loss=0.04256, over 4908.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.242, pruned_loss=0.04673, over 824367.26 frames. ], batch size: 36, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:47:47,780 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.194e+02 1.618e+02 1.920e+02 2.400e+02 5.071e+02, threshold=3.839e+02, percent-clipped=2.0 +2023-04-28 03:48:21,587 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1436, 1.7096, 1.5399, 2.0688, 2.1644, 1.9034, 1.8608, 1.6127], + device='cuda:3'), covar=tensor([0.1297, 0.1636, 0.1373, 0.1412, 0.1268, 0.1794, 0.1671, 0.1733], + device='cuda:3'), in_proj_covar=tensor([0.0312, 0.0305, 0.0346, 0.0284, 0.0322, 0.0302, 0.0297, 0.0372], + device='cuda:3'), out_proj_covar=tensor([6.3550e-05, 6.2497e-05, 7.2385e-05, 5.6854e-05, 6.5639e-05, 6.2742e-05, + 6.1197e-05, 7.8507e-05], device='cuda:3') +2023-04-28 03:48:36,944 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5752, 1.5191, 0.6994, 1.2997, 1.5064, 1.4722, 1.3735, 1.3662], + device='cuda:3'), covar=tensor([0.0438, 0.0325, 0.0368, 0.0485, 0.0281, 0.0442, 0.0425, 0.0502], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 03:48:48,103 INFO [finetune.py:976] (3/7) Epoch 29, batch 450, loss[loss=0.1497, simple_loss=0.2226, pruned_loss=0.03841, over 4927.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2411, pruned_loss=0.04627, over 852365.09 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:49:19,869 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.8439, 3.7032, 2.7404, 4.3987, 3.8980, 3.7977, 1.5263, 3.7966], + device='cuda:3'), covar=tensor([0.1481, 0.1159, 0.2829, 0.1499, 0.2301, 0.1619, 0.6110, 0.2337], + device='cuda:3'), in_proj_covar=tensor([0.0250, 0.0223, 0.0254, 0.0308, 0.0303, 0.0254, 0.0280, 0.0278], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 03:49:51,964 INFO [finetune.py:976] (3/7) Epoch 29, batch 500, loss[loss=0.135, simple_loss=0.2104, pruned_loss=0.02974, over 4753.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2371, pruned_loss=0.04514, over 873158.13 frames. ], batch size: 28, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:49:53,843 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.560e+02 1.871e+02 2.250e+02 5.218e+02, threshold=3.742e+02, percent-clipped=1.0 +2023-04-28 03:50:02,890 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-28 03:50:25,299 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160901.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:50:56,033 INFO [finetune.py:976] (3/7) Epoch 29, batch 550, loss[loss=0.1465, simple_loss=0.2183, pruned_loss=0.0374, over 4894.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2347, pruned_loss=0.04456, over 891529.82 frames. ], batch size: 35, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:51:48,586 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160962.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:51:49,195 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160963.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 03:51:50,437 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160965.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:51:51,655 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4653, 1.2940, 1.6183, 1.6351, 1.3291, 1.2970, 1.3011, 0.7131], + device='cuda:3'), covar=tensor([0.0458, 0.0583, 0.0359, 0.0555, 0.0705, 0.1051, 0.0540, 0.0586], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0066, 0.0065, 0.0068, 0.0074, 0.0093, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 03:52:01,358 INFO [finetune.py:976] (3/7) Epoch 29, batch 600, loss[loss=0.1801, simple_loss=0.2616, pruned_loss=0.0493, over 4855.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2344, pruned_loss=0.04437, over 906303.27 frames. ], batch size: 44, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:52:03,146 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.581e+02 1.979e+02 2.385e+02 4.353e+02, threshold=3.958e+02, percent-clipped=2.0 +2023-04-28 03:52:35,726 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161011.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:52:36,362 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8300, 2.3703, 2.0881, 2.2379, 1.7719, 1.9711, 1.7927, 1.5180], + device='cuda:3'), covar=tensor([0.1884, 0.1029, 0.0732, 0.1233, 0.3391, 0.1101, 0.1978, 0.2597], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0299, 0.0216, 0.0275, 0.0313, 0.0253, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1216e-04, 1.1769e-04, 8.4773e-05, 1.0791e-04, 1.2607e-04, 9.9356e-05, + 9.9928e-05, 1.0375e-04], device='cuda:3') +2023-04-28 03:52:38,795 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4241, 2.2963, 2.4297, 2.9426, 2.8202, 2.3640, 2.1948, 2.6941], + device='cuda:3'), covar=tensor([0.0847, 0.0982, 0.0695, 0.0548, 0.0609, 0.0880, 0.0675, 0.0498], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0203, 0.0183, 0.0171, 0.0177, 0.0177, 0.0150, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:52:44,113 INFO [finetune.py:976] (3/7) Epoch 29, batch 650, loss[loss=0.1612, simple_loss=0.2418, pruned_loss=0.04036, over 4830.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2359, pruned_loss=0.04453, over 916434.38 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:52:44,873 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161026.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:52:51,969 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-28 03:53:17,105 INFO [finetune.py:976] (3/7) Epoch 29, batch 700, loss[loss=0.1914, simple_loss=0.2707, pruned_loss=0.05602, over 4923.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2382, pruned_loss=0.04512, over 925700.26 frames. ], batch size: 42, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:53:18,921 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.470e+01 1.599e+02 1.884e+02 2.263e+02 4.345e+02, threshold=3.768e+02, percent-clipped=2.0 +2023-04-28 03:53:50,552 INFO [finetune.py:976] (3/7) Epoch 29, batch 750, loss[loss=0.1502, simple_loss=0.2438, pruned_loss=0.02833, over 4736.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2395, pruned_loss=0.04523, over 931341.55 frames. ], batch size: 27, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:53:52,525 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4390, 1.3759, 1.6695, 1.7230, 1.2929, 1.2324, 1.3647, 0.8725], + device='cuda:3'), covar=tensor([0.0492, 0.0579, 0.0356, 0.0411, 0.0626, 0.0981, 0.0551, 0.0543], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 03:54:24,404 INFO [finetune.py:976] (3/7) Epoch 29, batch 800, loss[loss=0.1712, simple_loss=0.2437, pruned_loss=0.04936, over 4890.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2393, pruned_loss=0.04477, over 936218.26 frames. ], batch size: 43, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:54:26,213 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.992e+01 1.674e+02 1.972e+02 2.388e+02 4.465e+02, threshold=3.944e+02, percent-clipped=4.0 +2023-04-28 03:54:52,595 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161216.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:54:53,811 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2105, 1.5763, 1.5111, 1.7510, 1.7142, 1.8639, 1.5094, 3.2138], + device='cuda:3'), covar=tensor([0.0601, 0.0707, 0.0685, 0.1095, 0.0554, 0.0459, 0.0644, 0.0165], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 03:54:56,221 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3582, 2.2102, 2.2399, 1.9344, 2.4026, 2.1482, 2.9454, 1.8925], + device='cuda:3'), covar=tensor([0.3006, 0.1783, 0.3423, 0.2621, 0.1357, 0.1930, 0.1159, 0.4003], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0357, 0.0426, 0.0355, 0.0389, 0.0378, 0.0375, 0.0428], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 03:54:57,905 INFO [finetune.py:976] (3/7) Epoch 29, batch 850, loss[loss=0.1515, simple_loss=0.2279, pruned_loss=0.03752, over 4779.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2374, pruned_loss=0.04463, over 939903.37 frames. ], batch size: 29, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:55:01,670 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161231.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:55:29,547 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161257.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:55:49,584 INFO [finetune.py:976] (3/7) Epoch 29, batch 900, loss[loss=0.1627, simple_loss=0.2293, pruned_loss=0.048, over 4935.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2349, pruned_loss=0.04383, over 941570.85 frames. ], batch size: 38, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:55:50,927 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161277.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:55:51,421 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.567e+01 1.487e+02 1.850e+02 2.194e+02 4.508e+02, threshold=3.700e+02, percent-clipped=1.0 +2023-04-28 03:56:00,187 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161292.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:56:21,348 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161321.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:56:23,751 INFO [finetune.py:976] (3/7) Epoch 29, batch 950, loss[loss=0.1508, simple_loss=0.2242, pruned_loss=0.03875, over 4776.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2338, pruned_loss=0.04394, over 945248.15 frames. ], batch size: 29, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:56:35,384 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1070, 2.4565, 1.0910, 1.3581, 2.0399, 1.1705, 3.3555, 1.7493], + device='cuda:3'), covar=tensor([0.0730, 0.0583, 0.0792, 0.1343, 0.0498, 0.1126, 0.0256, 0.0641], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0065, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 03:56:38,258 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-28 03:57:07,673 INFO [finetune.py:976] (3/7) Epoch 29, batch 1000, loss[loss=0.194, simple_loss=0.2719, pruned_loss=0.05806, over 4811.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2355, pruned_loss=0.04437, over 946424.58 frames. ], batch size: 40, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:57:09,493 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.605e+02 1.983e+02 2.337e+02 3.564e+02, threshold=3.965e+02, percent-clipped=0.0 +2023-04-28 03:57:38,279 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3904, 1.8274, 1.6708, 2.2906, 2.3641, 2.0373, 1.9326, 1.6995], + device='cuda:3'), covar=tensor([0.1791, 0.1863, 0.1760, 0.1429, 0.1351, 0.1645, 0.2031, 0.2288], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0309, 0.0351, 0.0288, 0.0328, 0.0307, 0.0301, 0.0377], + device='cuda:3'), out_proj_covar=tensor([6.4496e-05, 6.3351e-05, 7.3511e-05, 5.7688e-05, 6.6889e-05, 6.3702e-05, + 6.2145e-05, 7.9699e-05], device='cuda:3') +2023-04-28 03:57:39,573 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161401.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:58:12,304 INFO [finetune.py:976] (3/7) Epoch 29, batch 1050, loss[loss=0.1396, simple_loss=0.2198, pruned_loss=0.02967, over 4782.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2385, pruned_loss=0.04526, over 950058.95 frames. ], batch size: 26, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:58:24,034 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8817, 1.7427, 2.3713, 2.4283, 1.6442, 1.5522, 1.8567, 1.0826], + device='cuda:3'), covar=tensor([0.0626, 0.0893, 0.0368, 0.0744, 0.0751, 0.1114, 0.0666, 0.0692], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0073, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 03:58:55,487 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161460.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:58:56,749 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161462.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:59:17,654 INFO [finetune.py:976] (3/7) Epoch 29, batch 1100, loss[loss=0.155, simple_loss=0.2376, pruned_loss=0.03621, over 4903.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2417, pruned_loss=0.04677, over 951687.60 frames. ], batch size: 37, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:59:20,410 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.507e+02 1.843e+02 2.398e+02 4.910e+02, threshold=3.687e+02, percent-clipped=4.0 +2023-04-28 04:00:20,265 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161521.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:00:23,055 INFO [finetune.py:976] (3/7) Epoch 29, batch 1150, loss[loss=0.1703, simple_loss=0.247, pruned_loss=0.04678, over 4914.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2422, pruned_loss=0.04706, over 952458.24 frames. ], batch size: 42, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:00:23,850 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-28 04:00:47,542 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 04:01:04,587 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161557.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:01:25,598 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161572.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:01:27,861 INFO [finetune.py:976] (3/7) Epoch 29, batch 1200, loss[loss=0.1956, simple_loss=0.267, pruned_loss=0.06211, over 4882.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2403, pruned_loss=0.04643, over 952493.61 frames. ], batch size: 35, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:01:29,679 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.552e+02 1.827e+02 2.258e+02 5.032e+02, threshold=3.654e+02, percent-clipped=3.0 +2023-04-28 04:01:47,189 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161587.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:02:02,695 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-28 04:02:08,809 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-28 04:02:09,297 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161605.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:02:23,986 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161621.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:02:33,462 INFO [finetune.py:976] (3/7) Epoch 29, batch 1250, loss[loss=0.1351, simple_loss=0.2064, pruned_loss=0.03187, over 4907.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2385, pruned_loss=0.04647, over 951995.45 frames. ], batch size: 46, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:02:42,897 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 04:03:05,226 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6440, 0.7170, 1.6178, 2.0087, 1.7403, 1.5910, 1.6257, 1.6014], + device='cuda:3'), covar=tensor([0.4460, 0.6232, 0.5606, 0.5984, 0.5508, 0.7415, 0.7310, 0.7678], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0427, 0.0522, 0.0511, 0.0475, 0.0514, 0.0515, 0.0530], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:03:28,500 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161669.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:03:28,565 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0807, 1.9268, 1.7837, 1.6298, 2.1030, 1.8075, 2.4957, 1.6182], + device='cuda:3'), covar=tensor([0.3534, 0.1942, 0.4672, 0.2888, 0.1618, 0.2196, 0.1433, 0.4526], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0356, 0.0424, 0.0353, 0.0386, 0.0377, 0.0373, 0.0424], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:03:28,576 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7235, 1.3657, 1.3336, 1.4978, 1.8535, 1.5664, 1.2624, 1.3123], + device='cuda:3'), covar=tensor([0.1640, 0.1356, 0.1801, 0.1160, 0.0812, 0.1452, 0.1906, 0.2357], + device='cuda:3'), in_proj_covar=tensor([0.0317, 0.0309, 0.0351, 0.0288, 0.0328, 0.0308, 0.0301, 0.0378], + device='cuda:3'), out_proj_covar=tensor([6.4600e-05, 6.3231e-05, 7.3462e-05, 5.7609e-05, 6.6919e-05, 6.3900e-05, + 6.2001e-05, 7.9786e-05], device='cuda:3') +2023-04-28 04:03:35,474 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4313, 1.3977, 1.7505, 1.7481, 1.3026, 1.2436, 1.4927, 0.8924], + device='cuda:3'), covar=tensor([0.0553, 0.0492, 0.0349, 0.0593, 0.0839, 0.1020, 0.0495, 0.0555], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 04:03:38,165 INFO [finetune.py:976] (3/7) Epoch 29, batch 1300, loss[loss=0.1597, simple_loss=0.2353, pruned_loss=0.04202, over 4873.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2358, pruned_loss=0.04518, over 950815.62 frames. ], batch size: 34, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:03:46,267 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.807e+01 1.476e+02 1.722e+02 2.175e+02 4.011e+02, threshold=3.444e+02, percent-clipped=1.0 +2023-04-28 04:04:30,912 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-28 04:04:42,648 INFO [finetune.py:976] (3/7) Epoch 29, batch 1350, loss[loss=0.1829, simple_loss=0.2549, pruned_loss=0.05542, over 4753.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.237, pruned_loss=0.04573, over 953381.66 frames. ], batch size: 27, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:05:27,903 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161757.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:05:50,111 INFO [finetune.py:976] (3/7) Epoch 29, batch 1400, loss[loss=0.1555, simple_loss=0.2245, pruned_loss=0.04324, over 4782.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2397, pruned_loss=0.04677, over 952214.93 frames. ], batch size: 26, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:05:58,003 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.598e+02 1.895e+02 2.318e+02 6.343e+02, threshold=3.789e+02, percent-clipped=7.0 +2023-04-28 04:06:51,809 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161815.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:06:52,365 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161816.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:07:03,300 INFO [finetune.py:976] (3/7) Epoch 29, batch 1450, loss[loss=0.1125, simple_loss=0.175, pruned_loss=0.025, over 4456.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2405, pruned_loss=0.04669, over 950941.22 frames. ], batch size: 19, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:07:18,019 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-04-28 04:08:00,389 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161872.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:08:02,915 INFO [finetune.py:976] (3/7) Epoch 29, batch 1500, loss[loss=0.1445, simple_loss=0.221, pruned_loss=0.03396, over 4781.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2413, pruned_loss=0.04705, over 951517.32 frames. ], batch size: 26, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:08:03,669 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161876.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:08:04,722 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.641e+02 1.910e+02 2.350e+02 4.691e+02, threshold=3.820e+02, percent-clipped=1.0 +2023-04-28 04:08:16,009 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161887.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:09:01,027 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161920.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:09:09,621 INFO [finetune.py:976] (3/7) Epoch 29, batch 1550, loss[loss=0.1162, simple_loss=0.191, pruned_loss=0.02068, over 4694.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2418, pruned_loss=0.04683, over 951256.44 frames. ], batch size: 23, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:09:21,319 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161935.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:10:14,712 INFO [finetune.py:976] (3/7) Epoch 29, batch 1600, loss[loss=0.156, simple_loss=0.2325, pruned_loss=0.03971, over 4921.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.24, pruned_loss=0.04662, over 952078.84 frames. ], batch size: 38, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:10:16,470 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 7.098e+01 1.531e+02 1.823e+02 2.197e+02 4.043e+02, threshold=3.646e+02, percent-clipped=1.0 +2023-04-28 04:11:10,810 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-28 04:11:20,434 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0060, 1.5007, 1.8342, 1.7091, 1.7888, 1.4707, 0.7804, 1.4726], + device='cuda:3'), covar=tensor([0.3126, 0.3174, 0.1719, 0.2296, 0.2565, 0.2712, 0.4014, 0.2058], + device='cuda:3'), in_proj_covar=tensor([0.0289, 0.0244, 0.0226, 0.0311, 0.0221, 0.0233, 0.0226, 0.0184], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 04:11:22,119 INFO [finetune.py:976] (3/7) Epoch 29, batch 1650, loss[loss=0.1573, simple_loss=0.2336, pruned_loss=0.04055, over 4907.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.238, pruned_loss=0.04609, over 952672.70 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:12:07,061 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162057.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:12:29,332 INFO [finetune.py:976] (3/7) Epoch 29, batch 1700, loss[loss=0.183, simple_loss=0.262, pruned_loss=0.05202, over 4891.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2352, pruned_loss=0.04494, over 954227.56 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:12:36,095 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162077.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:12:36,610 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.915e+01 1.582e+02 1.975e+02 2.280e+02 6.731e+02, threshold=3.951e+02, percent-clipped=4.0 +2023-04-28 04:13:11,494 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162105.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:13:24,530 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162116.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:13:29,894 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4094, 1.0345, 0.5748, 1.1527, 1.1465, 1.2983, 1.2298, 1.2149], + device='cuda:3'), covar=tensor([0.0495, 0.0391, 0.0390, 0.0547, 0.0314, 0.0489, 0.0466, 0.0566], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0054, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 04:13:34,111 INFO [finetune.py:976] (3/7) Epoch 29, batch 1750, loss[loss=0.1781, simple_loss=0.2628, pruned_loss=0.04671, over 4822.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2381, pruned_loss=0.04603, over 954399.92 frames. ], batch size: 40, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:13:40,745 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1173, 2.5530, 1.0555, 1.3886, 1.8368, 1.2533, 3.3607, 1.8228], + device='cuda:3'), covar=tensor([0.0651, 0.0533, 0.0793, 0.1244, 0.0505, 0.0994, 0.0253, 0.0585], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 04:13:51,942 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162138.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:14:27,313 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162164.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:14:37,342 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162171.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:14:45,211 INFO [finetune.py:976] (3/7) Epoch 29, batch 1800, loss[loss=0.1852, simple_loss=0.2703, pruned_loss=0.0501, over 4246.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2407, pruned_loss=0.04633, over 954768.56 frames. ], batch size: 65, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:14:47,021 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.596e+02 1.871e+02 2.266e+02 6.327e+02, threshold=3.743e+02, percent-clipped=2.0 +2023-04-28 04:14:56,441 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0854, 1.3149, 1.2287, 1.5401, 1.4266, 1.4338, 1.2567, 2.4546], + device='cuda:3'), covar=tensor([0.0614, 0.0856, 0.0848, 0.1316, 0.0681, 0.0526, 0.0806, 0.0215], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 04:15:20,863 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9159, 2.2569, 1.0247, 1.2589, 1.7160, 1.1192, 2.9895, 1.5532], + device='cuda:3'), covar=tensor([0.0699, 0.0613, 0.0748, 0.1183, 0.0481, 0.1026, 0.0258, 0.0582], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 04:15:40,570 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1316, 2.7835, 2.4512, 2.6673, 1.8542, 2.4678, 2.5041, 1.8955], + device='cuda:3'), covar=tensor([0.2110, 0.1076, 0.0759, 0.1138, 0.3486, 0.0868, 0.1813, 0.2588], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0300, 0.0216, 0.0276, 0.0314, 0.0252, 0.0248, 0.0263], + device='cuda:3'), out_proj_covar=tensor([1.1244e-04, 1.1792e-04, 8.4774e-05, 1.0846e-04, 1.2629e-04, 9.9113e-05, + 9.9742e-05, 1.0366e-04], device='cuda:3') +2023-04-28 04:15:49,785 INFO [finetune.py:976] (3/7) Epoch 29, batch 1850, loss[loss=0.1574, simple_loss=0.2401, pruned_loss=0.03735, over 4830.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2421, pruned_loss=0.04726, over 956625.30 frames. ], batch size: 47, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:15:53,544 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162231.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:16:53,806 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-28 04:16:53,967 INFO [finetune.py:976] (3/7) Epoch 29, batch 1900, loss[loss=0.1399, simple_loss=0.2189, pruned_loss=0.03044, over 4807.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2427, pruned_loss=0.04708, over 956585.78 frames. ], batch size: 38, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:16:55,788 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.192e+01 1.518e+02 1.776e+02 2.128e+02 3.542e+02, threshold=3.553e+02, percent-clipped=0.0 +2023-04-28 04:17:14,638 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162292.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:17:35,885 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-28 04:17:58,314 INFO [finetune.py:976] (3/7) Epoch 29, batch 1950, loss[loss=0.1563, simple_loss=0.2291, pruned_loss=0.04176, over 4817.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2414, pruned_loss=0.04653, over 957544.86 frames. ], batch size: 30, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:18:51,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6919, 2.5092, 2.6954, 3.2848, 3.0728, 2.6053, 2.5249, 2.9995], + device='cuda:3'), covar=tensor([0.0761, 0.0906, 0.0640, 0.0451, 0.0523, 0.0771, 0.0594, 0.0450], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0203, 0.0184, 0.0170, 0.0177, 0.0177, 0.0151, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:19:03,739 INFO [finetune.py:976] (3/7) Epoch 29, batch 2000, loss[loss=0.1743, simple_loss=0.2515, pruned_loss=0.04859, over 4895.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2384, pruned_loss=0.0456, over 957593.89 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:19:05,551 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.225e+01 1.505e+02 1.783e+02 2.109e+02 5.594e+02, threshold=3.566e+02, percent-clipped=2.0 +2023-04-28 04:19:30,715 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.75 vs. limit=5.0 +2023-04-28 04:19:43,204 INFO [finetune.py:976] (3/7) Epoch 29, batch 2050, loss[loss=0.1388, simple_loss=0.2131, pruned_loss=0.03231, over 4829.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2351, pruned_loss=0.04481, over 956555.94 frames. ], batch size: 39, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:19:48,144 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162433.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:20:13,304 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162471.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:20:16,687 INFO [finetune.py:976] (3/7) Epoch 29, batch 2100, loss[loss=0.1429, simple_loss=0.2106, pruned_loss=0.03762, over 4707.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2344, pruned_loss=0.04428, over 958192.23 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:20:19,020 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.161e+01 1.530e+02 1.785e+02 2.249e+02 3.474e+02, threshold=3.570e+02, percent-clipped=1.0 +2023-04-28 04:20:39,346 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162511.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:20:45,709 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162519.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:20:50,371 INFO [finetune.py:976] (3/7) Epoch 29, batch 2150, loss[loss=0.2015, simple_loss=0.2853, pruned_loss=0.05891, over 4815.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2376, pruned_loss=0.04509, over 956762.38 frames. ], batch size: 45, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:21:01,284 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 04:21:21,602 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162572.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 04:21:23,001 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.19 vs. limit=5.0 +2023-04-28 04:21:23,762 INFO [finetune.py:976] (3/7) Epoch 29, batch 2200, loss[loss=0.1906, simple_loss=0.2593, pruned_loss=0.06101, over 4891.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2403, pruned_loss=0.04621, over 953715.58 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:21:26,034 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.398e+01 1.632e+02 1.891e+02 2.223e+02 3.490e+02, threshold=3.782e+02, percent-clipped=0.0 +2023-04-28 04:21:37,950 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162587.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:22:23,020 INFO [finetune.py:976] (3/7) Epoch 29, batch 2250, loss[loss=0.2058, simple_loss=0.274, pruned_loss=0.06882, over 4889.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2404, pruned_loss=0.04619, over 953224.88 frames. ], batch size: 43, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:23:28,884 INFO [finetune.py:976] (3/7) Epoch 29, batch 2300, loss[loss=0.191, simple_loss=0.2607, pruned_loss=0.06061, over 4817.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2413, pruned_loss=0.04626, over 953416.88 frames. ], batch size: 33, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:23:35,928 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2813, 3.2529, 0.8288, 1.6747, 1.6406, 2.3033, 1.8971, 0.9801], + device='cuda:3'), covar=tensor([0.1820, 0.1453, 0.2399, 0.1761, 0.1404, 0.1321, 0.1690, 0.2269], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0237, 0.0136, 0.0120, 0.0131, 0.0153, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 04:23:36,413 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.802e+01 1.609e+02 1.813e+02 2.047e+02 3.617e+02, threshold=3.626e+02, percent-clipped=0.0 +2023-04-28 04:24:08,647 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.1004, 2.3577, 2.2914, 2.4884, 2.2827, 2.3830, 2.3836, 2.3549], + device='cuda:3'), covar=tensor([0.3169, 0.5664, 0.4784, 0.4126, 0.5563, 0.6466, 0.5632, 0.5031], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0378, 0.0335, 0.0346, 0.0354, 0.0399, 0.0366, 0.0338], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 04:24:12,276 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5812, 0.7099, 1.5162, 1.9696, 1.6941, 1.5193, 1.5292, 1.5050], + device='cuda:3'), covar=tensor([0.3884, 0.5758, 0.5163, 0.4938, 0.4674, 0.6069, 0.6419, 0.7662], + device='cuda:3'), in_proj_covar=tensor([0.0449, 0.0426, 0.0522, 0.0511, 0.0477, 0.0513, 0.0516, 0.0530], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:24:31,810 INFO [finetune.py:976] (3/7) Epoch 29, batch 2350, loss[loss=0.1492, simple_loss=0.2224, pruned_loss=0.038, over 4904.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2393, pruned_loss=0.0456, over 954076.76 frames. ], batch size: 36, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:24:41,993 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162733.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:25:30,909 INFO [finetune.py:976] (3/7) Epoch 29, batch 2400, loss[loss=0.1456, simple_loss=0.2123, pruned_loss=0.03948, over 4730.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2367, pruned_loss=0.04471, over 954107.59 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:25:33,145 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.191e+01 1.527e+02 1.810e+02 2.223e+02 4.938e+02, threshold=3.619e+02, percent-clipped=3.0 +2023-04-28 04:25:35,557 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1142, 1.8866, 2.1263, 2.4753, 2.5707, 1.9911, 1.8598, 2.1757], + device='cuda:3'), covar=tensor([0.0771, 0.1079, 0.0660, 0.0514, 0.0536, 0.0839, 0.0637, 0.0556], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0203, 0.0184, 0.0170, 0.0177, 0.0177, 0.0151, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:25:36,110 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162781.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:26:04,587 INFO [finetune.py:976] (3/7) Epoch 29, batch 2450, loss[loss=0.211, simple_loss=0.2861, pruned_loss=0.06794, over 4858.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2354, pruned_loss=0.04466, over 955037.83 frames. ], batch size: 47, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:26:05,954 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162827.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:26:30,508 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-04-28 04:26:33,771 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162867.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 04:26:38,581 INFO [finetune.py:976] (3/7) Epoch 29, batch 2500, loss[loss=0.1787, simple_loss=0.2521, pruned_loss=0.0526, over 4787.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2378, pruned_loss=0.04532, over 952289.31 frames. ], batch size: 51, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:26:40,362 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.423e+01 1.461e+02 1.777e+02 2.062e+02 3.489e+02, threshold=3.555e+02, percent-clipped=0.0 +2023-04-28 04:26:47,955 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162887.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:26:48,585 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162888.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:27:12,362 INFO [finetune.py:976] (3/7) Epoch 29, batch 2550, loss[loss=0.1723, simple_loss=0.249, pruned_loss=0.04782, over 4803.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2415, pruned_loss=0.04633, over 952994.03 frames. ], batch size: 51, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:27:18,508 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5415, 1.7476, 1.9025, 2.0010, 1.8327, 1.8906, 1.9615, 1.9329], + device='cuda:3'), covar=tensor([0.3573, 0.5167, 0.4384, 0.3997, 0.5175, 0.6427, 0.4970, 0.4521], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0375, 0.0332, 0.0343, 0.0352, 0.0396, 0.0363, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 04:27:19,032 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162935.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:27:46,026 INFO [finetune.py:976] (3/7) Epoch 29, batch 2600, loss[loss=0.166, simple_loss=0.2461, pruned_loss=0.04292, over 4821.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.243, pruned_loss=0.04705, over 952659.04 frames. ], batch size: 33, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:27:47,803 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.782e+01 1.602e+02 1.933e+02 2.328e+02 3.675e+02, threshold=3.867e+02, percent-clipped=1.0 +2023-04-28 04:28:12,095 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-28 04:28:19,511 INFO [finetune.py:976] (3/7) Epoch 29, batch 2650, loss[loss=0.1563, simple_loss=0.2317, pruned_loss=0.04048, over 4892.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.243, pruned_loss=0.04663, over 951575.18 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:28:57,567 INFO [finetune.py:976] (3/7) Epoch 29, batch 2700, loss[loss=0.167, simple_loss=0.2422, pruned_loss=0.04585, over 4912.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2404, pruned_loss=0.0456, over 951791.22 frames. ], batch size: 46, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:29:04,330 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.516e+02 1.799e+02 2.193e+02 4.304e+02, threshold=3.599e+02, percent-clipped=1.0 +2023-04-28 04:29:15,567 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6088, 1.7099, 1.3622, 1.0770, 1.2423, 1.2493, 1.2664, 1.1406], + device='cuda:3'), covar=tensor([0.1656, 0.1209, 0.1527, 0.1710, 0.2243, 0.2002, 0.1001, 0.2002], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0209, 0.0170, 0.0203, 0.0200, 0.0186, 0.0155, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 04:29:17,290 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163090.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:30:02,212 INFO [finetune.py:976] (3/7) Epoch 29, batch 2750, loss[loss=0.1739, simple_loss=0.2397, pruned_loss=0.05404, over 4822.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2389, pruned_loss=0.04572, over 953671.94 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:30:11,869 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1486, 2.7789, 2.1342, 2.1198, 1.5111, 1.5448, 2.3066, 1.5783], + device='cuda:3'), covar=tensor([0.1536, 0.1305, 0.1314, 0.1610, 0.2099, 0.1867, 0.0858, 0.1894], + device='cuda:3'), in_proj_covar=tensor([0.0197, 0.0209, 0.0169, 0.0203, 0.0200, 0.0186, 0.0155, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 04:30:38,998 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163151.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:30:50,660 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163167.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 04:31:01,089 INFO [finetune.py:976] (3/7) Epoch 29, batch 2800, loss[loss=0.1466, simple_loss=0.2288, pruned_loss=0.03218, over 4879.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2362, pruned_loss=0.04511, over 956050.72 frames. ], batch size: 34, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:31:08,097 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.442e+02 1.723e+02 2.053e+02 5.173e+02, threshold=3.446e+02, percent-clipped=1.0 +2023-04-28 04:31:11,253 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163183.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:31:52,469 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163215.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:31:59,017 INFO [finetune.py:976] (3/7) Epoch 29, batch 2850, loss[loss=0.1372, simple_loss=0.2148, pruned_loss=0.02977, over 4781.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2338, pruned_loss=0.04371, over 954713.25 frames. ], batch size: 29, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:32:08,286 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163240.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:32:31,405 INFO [finetune.py:976] (3/7) Epoch 29, batch 2900, loss[loss=0.1413, simple_loss=0.2205, pruned_loss=0.03105, over 4894.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2376, pruned_loss=0.0451, over 954539.16 frames. ], batch size: 37, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:32:33,720 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.435e+02 1.823e+02 2.184e+02 4.986e+02, threshold=3.647e+02, percent-clipped=1.0 +2023-04-28 04:32:47,862 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163301.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:33:04,179 INFO [finetune.py:976] (3/7) Epoch 29, batch 2950, loss[loss=0.1901, simple_loss=0.2654, pruned_loss=0.05737, over 4787.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.241, pruned_loss=0.04611, over 955657.17 frames. ], batch size: 29, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:33:27,664 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6921, 1.7500, 0.8766, 1.3539, 1.9012, 1.5745, 1.4455, 1.4853], + device='cuda:3'), covar=tensor([0.0451, 0.0326, 0.0323, 0.0513, 0.0268, 0.0458, 0.0441, 0.0503], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0054, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 04:33:37,006 INFO [finetune.py:976] (3/7) Epoch 29, batch 3000, loss[loss=0.1582, simple_loss=0.2249, pruned_loss=0.04571, over 4737.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2415, pruned_loss=0.04651, over 954463.05 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:33:37,006 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 04:33:44,054 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2069, 2.5859, 1.1454, 1.4292, 2.0111, 1.4066, 3.0780, 1.7680], + device='cuda:3'), covar=tensor([0.0558, 0.0547, 0.0643, 0.1139, 0.0388, 0.0825, 0.0269, 0.0529], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 04:33:47,842 INFO [finetune.py:1010] (3/7) Epoch 29, validation: loss=0.1535, simple_loss=0.222, pruned_loss=0.04252, over 2265189.00 frames. +2023-04-28 04:33:47,842 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 04:33:49,649 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.593e+02 1.913e+02 2.268e+02 4.606e+02, threshold=3.825e+02, percent-clipped=1.0 +2023-04-28 04:34:17,718 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-28 04:34:19,353 INFO [finetune.py:976] (3/7) Epoch 29, batch 3050, loss[loss=0.1442, simple_loss=0.2217, pruned_loss=0.03338, over 4841.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2421, pruned_loss=0.04646, over 956387.55 frames. ], batch size: 49, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:34:34,895 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163446.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:34:53,041 INFO [finetune.py:976] (3/7) Epoch 29, batch 3100, loss[loss=0.1964, simple_loss=0.2535, pruned_loss=0.06966, over 4823.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2399, pruned_loss=0.04582, over 956159.84 frames. ], batch size: 30, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:34:55,823 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.898e+01 1.464e+02 1.707e+02 2.164e+02 5.622e+02, threshold=3.413e+02, percent-clipped=1.0 +2023-04-28 04:34:59,438 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163483.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:35:12,454 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1451, 2.7543, 1.0634, 1.4460, 1.9125, 1.3385, 3.5914, 1.9429], + device='cuda:3'), covar=tensor([0.0658, 0.0566, 0.0760, 0.1144, 0.0542, 0.0935, 0.0300, 0.0576], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 04:35:27,056 INFO [finetune.py:976] (3/7) Epoch 29, batch 3150, loss[loss=0.1451, simple_loss=0.2156, pruned_loss=0.03734, over 4875.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2362, pruned_loss=0.04446, over 957218.32 frames. ], batch size: 31, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:35:28,961 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5654, 3.1256, 2.6024, 2.8926, 2.3479, 2.7063, 2.8316, 2.1554], + device='cuda:3'), covar=tensor([0.1912, 0.1243, 0.0765, 0.1264, 0.2809, 0.1114, 0.1663, 0.2602], + device='cuda:3'), in_proj_covar=tensor([0.0281, 0.0297, 0.0215, 0.0274, 0.0311, 0.0252, 0.0245, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1219e-04, 1.1683e-04, 8.4479e-05, 1.0772e-04, 1.2521e-04, 9.8830e-05, + 9.8769e-05, 1.0313e-04], device='cuda:3') +2023-04-28 04:35:31,734 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163531.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:35:52,640 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-04-28 04:36:10,391 INFO [finetune.py:976] (3/7) Epoch 29, batch 3200, loss[loss=0.197, simple_loss=0.2611, pruned_loss=0.06641, over 4850.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2329, pruned_loss=0.04293, over 957525.82 frames. ], batch size: 47, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:36:12,739 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.620e+01 1.506e+02 1.753e+02 2.068e+02 8.624e+02, threshold=3.506e+02, percent-clipped=5.0 +2023-04-28 04:36:42,504 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163596.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:36:44,381 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7261, 1.5150, 1.6797, 2.0298, 2.0783, 1.7030, 1.3881, 1.8420], + device='cuda:3'), covar=tensor([0.0743, 0.1236, 0.0818, 0.0531, 0.0623, 0.0755, 0.0751, 0.0560], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0202, 0.0183, 0.0171, 0.0177, 0.0176, 0.0151, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:37:17,128 INFO [finetune.py:976] (3/7) Epoch 29, batch 3250, loss[loss=0.1733, simple_loss=0.2433, pruned_loss=0.0517, over 4938.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2344, pruned_loss=0.04396, over 956866.64 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:38:20,407 INFO [finetune.py:976] (3/7) Epoch 29, batch 3300, loss[loss=0.1601, simple_loss=0.2562, pruned_loss=0.03205, over 4750.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2399, pruned_loss=0.04613, over 956027.38 frames. ], batch size: 54, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:38:22,208 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.605e+02 1.841e+02 2.298e+02 3.971e+02, threshold=3.681e+02, percent-clipped=3.0 +2023-04-28 04:39:13,622 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9180, 1.9684, 1.1206, 1.5912, 1.9063, 1.7926, 1.6329, 1.7434], + device='cuda:3'), covar=tensor([0.0430, 0.0334, 0.0326, 0.0524, 0.0251, 0.0445, 0.0475, 0.0503], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 04:39:14,483 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-04-28 04:39:22,573 INFO [finetune.py:976] (3/7) Epoch 29, batch 3350, loss[loss=0.115, simple_loss=0.1941, pruned_loss=0.01795, over 4702.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2404, pruned_loss=0.04581, over 954470.31 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:39:46,997 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163746.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:39:48,629 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 04:40:27,094 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7051, 1.3979, 1.3659, 1.3911, 1.8797, 1.5640, 1.2960, 1.3159], + device='cuda:3'), covar=tensor([0.1697, 0.1325, 0.1922, 0.1223, 0.0735, 0.1317, 0.1842, 0.2212], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0308, 0.0349, 0.0286, 0.0325, 0.0305, 0.0300, 0.0375], + device='cuda:3'), out_proj_covar=tensor([6.3896e-05, 6.3087e-05, 7.3067e-05, 5.7157e-05, 6.6349e-05, 6.3360e-05, + 6.1873e-05, 7.9099e-05], device='cuda:3') +2023-04-28 04:40:27,591 INFO [finetune.py:976] (3/7) Epoch 29, batch 3400, loss[loss=0.1641, simple_loss=0.2557, pruned_loss=0.0362, over 4829.00 frames. ], tot_loss[loss=0.167, simple_loss=0.241, pruned_loss=0.04643, over 953157.43 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:40:29,471 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.589e+02 1.829e+02 2.322e+02 5.159e+02, threshold=3.657e+02, percent-clipped=5.0 +2023-04-28 04:40:50,564 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163794.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:41:21,002 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-28 04:41:32,560 INFO [finetune.py:976] (3/7) Epoch 29, batch 3450, loss[loss=0.1387, simple_loss=0.212, pruned_loss=0.03266, over 4808.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2398, pruned_loss=0.04576, over 950988.75 frames. ], batch size: 45, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:41:34,581 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3260, 1.8816, 2.1903, 2.9515, 2.3124, 1.8749, 2.0191, 2.1847], + device='cuda:3'), covar=tensor([0.2794, 0.2875, 0.1509, 0.1858, 0.2311, 0.2313, 0.3253, 0.1977], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0246, 0.0229, 0.0314, 0.0223, 0.0237, 0.0229, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 04:42:36,322 INFO [finetune.py:976] (3/7) Epoch 29, batch 3500, loss[loss=0.1661, simple_loss=0.2421, pruned_loss=0.04508, over 4827.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2364, pruned_loss=0.04435, over 952995.59 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:42:37,685 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3763, 1.8044, 1.6701, 2.2518, 2.4567, 1.9982, 1.9882, 1.6868], + device='cuda:3'), covar=tensor([0.1586, 0.1780, 0.1713, 0.1350, 0.1034, 0.2099, 0.1869, 0.2120], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0307, 0.0347, 0.0285, 0.0324, 0.0304, 0.0298, 0.0373], + device='cuda:3'), out_proj_covar=tensor([6.3694e-05, 6.2839e-05, 7.2565e-05, 5.7053e-05, 6.6138e-05, 6.3091e-05, + 6.1513e-05, 7.8719e-05], device='cuda:3') +2023-04-28 04:42:38,147 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.447e+02 1.753e+02 2.110e+02 3.235e+02, threshold=3.507e+02, percent-clipped=0.0 +2023-04-28 04:42:59,215 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163896.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:43:28,397 INFO [finetune.py:976] (3/7) Epoch 29, batch 3550, loss[loss=0.1391, simple_loss=0.2099, pruned_loss=0.03413, over 3997.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2341, pruned_loss=0.0437, over 951246.16 frames. ], batch size: 17, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:43:39,967 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163944.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:43:55,245 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2068, 1.7334, 1.5904, 1.9799, 1.7958, 2.0616, 1.6573, 4.1197], + device='cuda:3'), covar=tensor([0.0560, 0.0720, 0.0703, 0.1079, 0.0556, 0.0542, 0.0659, 0.0102], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 04:44:01,287 INFO [finetune.py:976] (3/7) Epoch 29, batch 3600, loss[loss=0.1655, simple_loss=0.2379, pruned_loss=0.04659, over 4771.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2335, pruned_loss=0.04384, over 951678.43 frames. ], batch size: 27, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:44:03,546 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.301e+01 1.508e+02 1.790e+02 2.030e+02 3.807e+02, threshold=3.580e+02, percent-clipped=2.0 +2023-04-28 04:44:03,661 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9900, 2.4238, 1.0578, 1.2817, 1.6930, 1.2169, 3.0012, 1.5720], + device='cuda:3'), covar=tensor([0.0666, 0.0547, 0.0703, 0.1199, 0.0504, 0.0948, 0.0275, 0.0629], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 04:44:24,338 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6075, 3.0398, 1.3611, 1.9065, 1.8640, 2.4391, 1.9364, 1.3405], + device='cuda:3'), covar=tensor([0.1093, 0.0712, 0.1451, 0.1044, 0.0967, 0.0746, 0.1242, 0.1713], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0237, 0.0135, 0.0120, 0.0132, 0.0152, 0.0116, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 04:44:34,470 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2797, 1.5577, 1.7293, 1.8131, 1.6758, 1.7683, 1.7438, 1.7606], + device='cuda:3'), covar=tensor([0.4165, 0.5382, 0.4502, 0.4424, 0.5679, 0.7000, 0.5417, 0.4721], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0376, 0.0332, 0.0343, 0.0353, 0.0394, 0.0364, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 04:44:36,729 INFO [finetune.py:976] (3/7) Epoch 29, batch 3650, loss[loss=0.225, simple_loss=0.3022, pruned_loss=0.07387, over 4805.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2368, pruned_loss=0.04551, over 951429.84 frames. ], batch size: 51, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:45:03,414 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.09 vs. limit=2.0 +2023-04-28 04:45:10,039 INFO [finetune.py:976] (3/7) Epoch 29, batch 3700, loss[loss=0.1826, simple_loss=0.2561, pruned_loss=0.0546, over 4920.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.239, pruned_loss=0.04559, over 952826.66 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:45:11,860 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.572e+02 1.919e+02 2.476e+02 4.831e+02, threshold=3.838e+02, percent-clipped=5.0 +2023-04-28 04:45:12,092 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 04:45:19,699 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164090.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:45:38,973 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-28 04:45:43,377 INFO [finetune.py:976] (3/7) Epoch 29, batch 3750, loss[loss=0.1526, simple_loss=0.2194, pruned_loss=0.04295, over 4099.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2395, pruned_loss=0.04542, over 952055.76 frames. ], batch size: 65, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:46:11,064 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164151.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:46:18,653 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164155.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:46:43,101 INFO [finetune.py:976] (3/7) Epoch 29, batch 3800, loss[loss=0.1934, simple_loss=0.2712, pruned_loss=0.05775, over 4816.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2414, pruned_loss=0.04625, over 950376.79 frames. ], batch size: 47, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:46:47,404 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.309e+01 1.497e+02 1.761e+02 2.099e+02 4.096e+02, threshold=3.523e+02, percent-clipped=1.0 +2023-04-28 04:47:23,362 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164216.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:47:34,752 INFO [finetune.py:976] (3/7) Epoch 29, batch 3850, loss[loss=0.1593, simple_loss=0.2328, pruned_loss=0.04285, over 4899.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2408, pruned_loss=0.04667, over 951709.59 frames. ], batch size: 43, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:47:35,627 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-04-28 04:47:50,226 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2184, 1.4305, 1.3819, 1.6863, 1.5331, 1.6536, 1.3750, 3.0780], + device='cuda:3'), covar=tensor([0.0622, 0.0836, 0.0783, 0.1245, 0.0651, 0.0478, 0.0736, 0.0162], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 04:48:06,360 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0058, 2.7591, 1.0309, 1.5090, 2.0837, 1.2189, 3.5380, 1.8959], + device='cuda:3'), covar=tensor([0.0734, 0.0647, 0.0858, 0.1197, 0.0518, 0.1040, 0.0234, 0.0565], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 04:48:38,393 INFO [finetune.py:976] (3/7) Epoch 29, batch 3900, loss[loss=0.1683, simple_loss=0.2463, pruned_loss=0.04521, over 4915.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2387, pruned_loss=0.04587, over 952910.40 frames. ], batch size: 37, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:48:40,203 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.498e+02 1.753e+02 2.124e+02 6.066e+02, threshold=3.506e+02, percent-clipped=2.0 +2023-04-28 04:49:39,856 INFO [finetune.py:976] (3/7) Epoch 29, batch 3950, loss[loss=0.1666, simple_loss=0.2341, pruned_loss=0.04955, over 4759.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2358, pruned_loss=0.04475, over 952433.59 frames. ], batch size: 27, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:50:24,357 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164360.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:50:24,668 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.51 vs. limit=5.0 +2023-04-28 04:50:25,011 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9389, 2.4663, 2.1587, 2.0293, 1.4128, 1.4761, 2.1478, 1.4294], + device='cuda:3'), covar=tensor([0.1878, 0.1553, 0.1380, 0.1711, 0.2365, 0.2096, 0.0971, 0.2136], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0210, 0.0171, 0.0204, 0.0201, 0.0187, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 04:50:45,027 INFO [finetune.py:976] (3/7) Epoch 29, batch 4000, loss[loss=0.1947, simple_loss=0.2658, pruned_loss=0.06184, over 4909.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2354, pruned_loss=0.04475, over 953113.36 frames. ], batch size: 36, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:50:47,324 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.986e+01 1.466e+02 1.722e+02 2.043e+02 3.565e+02, threshold=3.444e+02, percent-clipped=1.0 +2023-04-28 04:51:41,777 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164421.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:51:48,794 INFO [finetune.py:976] (3/7) Epoch 29, batch 4050, loss[loss=0.1519, simple_loss=0.213, pruned_loss=0.04542, over 4716.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2385, pruned_loss=0.04563, over 953794.35 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:52:20,687 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164446.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:52:55,386 INFO [finetune.py:976] (3/7) Epoch 29, batch 4100, loss[loss=0.168, simple_loss=0.238, pruned_loss=0.04899, over 4808.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2403, pruned_loss=0.04575, over 953602.21 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:53:02,650 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.616e+02 1.841e+02 2.144e+02 4.180e+02, threshold=3.683e+02, percent-clipped=3.0 +2023-04-28 04:53:15,938 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6065, 1.8772, 2.0153, 2.1363, 1.9353, 1.9443, 2.1101, 2.0721], + device='cuda:3'), covar=tensor([0.3698, 0.5094, 0.4185, 0.3971, 0.5077, 0.6744, 0.4825, 0.4433], + device='cuda:3'), in_proj_covar=tensor([0.0346, 0.0376, 0.0334, 0.0344, 0.0354, 0.0396, 0.0365, 0.0336], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 04:53:27,836 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.4704, 1.3096, 1.4631, 1.0585, 1.3570, 1.2272, 1.7185, 1.4351], + device='cuda:3'), covar=tensor([0.3481, 0.2008, 0.4946, 0.2913, 0.1651, 0.2310, 0.1629, 0.4538], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0354, 0.0424, 0.0353, 0.0385, 0.0376, 0.0371, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:53:41,169 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164511.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:53:54,831 INFO [finetune.py:976] (3/7) Epoch 29, batch 4150, loss[loss=0.1366, simple_loss=0.2107, pruned_loss=0.03127, over 4832.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2415, pruned_loss=0.04623, over 954212.82 frames. ], batch size: 30, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:54:35,913 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164568.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:54:40,038 INFO [finetune.py:976] (3/7) Epoch 29, batch 4200, loss[loss=0.1794, simple_loss=0.2488, pruned_loss=0.05502, over 4197.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2425, pruned_loss=0.04624, over 954006.01 frames. ], batch size: 65, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:54:41,973 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164578.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:54:42,442 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.830e+01 1.587e+02 1.797e+02 2.330e+02 9.173e+02, threshold=3.593e+02, percent-clipped=2.0 +2023-04-28 04:55:14,302 INFO [finetune.py:976] (3/7) Epoch 29, batch 4250, loss[loss=0.166, simple_loss=0.2467, pruned_loss=0.04271, over 4790.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2388, pruned_loss=0.04457, over 952603.26 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:55:16,891 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164629.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:55:24,087 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164639.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 04:55:24,786 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.75 vs. limit=5.0 +2023-04-28 04:55:37,638 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164658.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:55:48,429 INFO [finetune.py:976] (3/7) Epoch 29, batch 4300, loss[loss=0.1769, simple_loss=0.2393, pruned_loss=0.0573, over 4938.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2366, pruned_loss=0.04426, over 953681.37 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:55:50,845 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.176e+01 1.533e+02 1.704e+02 2.198e+02 4.636e+02, threshold=3.409e+02, percent-clipped=3.0 +2023-04-28 04:55:54,012 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4765, 1.3967, 4.1726, 3.9107, 3.6860, 4.0465, 4.0285, 3.6497], + device='cuda:3'), covar=tensor([0.7651, 0.5894, 0.1124, 0.1793, 0.1186, 0.1808, 0.1320, 0.1458], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0312, 0.0411, 0.0413, 0.0353, 0.0417, 0.0321, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:56:16,392 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-28 04:56:16,862 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164716.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:56:18,770 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164719.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:56:22,291 INFO [finetune.py:976] (3/7) Epoch 29, batch 4350, loss[loss=0.126, simple_loss=0.2058, pruned_loss=0.0231, over 4795.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2348, pruned_loss=0.04412, over 954490.03 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:56:33,373 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7848, 2.9446, 2.4178, 2.7126, 3.0550, 2.7591, 3.8950, 2.4477], + device='cuda:3'), covar=tensor([0.3533, 0.2154, 0.4255, 0.2884, 0.1415, 0.2199, 0.1095, 0.3727], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0356, 0.0425, 0.0354, 0.0386, 0.0377, 0.0373, 0.0426], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:56:36,277 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:56:36,290 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:56:59,106 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4974, 3.5254, 0.8834, 1.7447, 1.9680, 2.5684, 1.9159, 0.9961], + device='cuda:3'), covar=tensor([0.1398, 0.0873, 0.2007, 0.1304, 0.1020, 0.0965, 0.1586, 0.2006], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0120, 0.0132, 0.0153, 0.0117, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 04:57:17,722 INFO [finetune.py:976] (3/7) Epoch 29, batch 4400, loss[loss=0.1568, simple_loss=0.2393, pruned_loss=0.03719, over 4835.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2356, pruned_loss=0.04502, over 954657.49 frames. ], batch size: 47, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:57:18,770 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-04-28 04:57:20,184 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.542e+02 1.919e+02 2.288e+02 3.219e+02, threshold=3.837e+02, percent-clipped=0.0 +2023-04-28 04:57:41,278 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164794.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:58:01,554 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164807.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:58:04,445 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164811.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:58:11,734 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2146, 1.5464, 1.3589, 1.9308, 1.6817, 1.9542, 1.5021, 4.1646], + device='cuda:3'), covar=tensor([0.0553, 0.0808, 0.0769, 0.1100, 0.0610, 0.0567, 0.0716, 0.0116], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 04:58:24,163 INFO [finetune.py:976] (3/7) Epoch 29, batch 4450, loss[loss=0.1673, simple_loss=0.247, pruned_loss=0.0438, over 4773.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2368, pruned_loss=0.04511, over 952760.10 frames. ], batch size: 26, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:58:59,955 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164859.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:59:01,214 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9393, 2.4505, 2.0209, 2.2550, 1.6435, 2.0583, 2.0392, 1.6270], + device='cuda:3'), covar=tensor([0.1754, 0.1016, 0.0788, 0.1099, 0.3355, 0.1123, 0.1633, 0.2379], + device='cuda:3'), in_proj_covar=tensor([0.0282, 0.0300, 0.0216, 0.0275, 0.0313, 0.0253, 0.0247, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1227e-04, 1.1781e-04, 8.4723e-05, 1.0817e-04, 1.2587e-04, 9.9325e-05, + 9.9223e-05, 1.0334e-04], device='cuda:3') +2023-04-28 04:59:02,517 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 04:59:04,188 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1462, 1.6757, 2.1964, 2.5282, 2.2005, 2.1222, 2.1884, 2.0700], + device='cuda:3'), covar=tensor([0.4187, 0.6247, 0.6097, 0.5026, 0.5769, 0.7490, 0.7242, 0.8217], + device='cuda:3'), in_proj_covar=tensor([0.0447, 0.0425, 0.0521, 0.0510, 0.0476, 0.0514, 0.0516, 0.0530], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 04:59:10,500 INFO [finetune.py:976] (3/7) Epoch 29, batch 4500, loss[loss=0.1951, simple_loss=0.2704, pruned_loss=0.05988, over 4879.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2385, pruned_loss=0.04579, over 950725.62 frames. ], batch size: 43, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:59:18,305 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.480e+02 1.804e+02 2.196e+02 5.471e+02, threshold=3.609e+02, percent-clipped=1.0 +2023-04-28 05:00:20,419 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164924.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:00:20,952 INFO [finetune.py:976] (3/7) Epoch 29, batch 4550, loss[loss=0.2367, simple_loss=0.3085, pruned_loss=0.08244, over 4889.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2403, pruned_loss=0.04597, over 951701.42 frames. ], batch size: 35, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 05:00:32,585 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164934.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 05:00:33,866 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9663, 1.4388, 1.8093, 1.6996, 1.8238, 1.4936, 0.8686, 1.4434], + device='cuda:3'), covar=tensor([0.3232, 0.3267, 0.1802, 0.2448, 0.2434, 0.2701, 0.4156, 0.2104], + device='cuda:3'), in_proj_covar=tensor([0.0298, 0.0250, 0.0232, 0.0320, 0.0226, 0.0240, 0.0233, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 05:01:26,332 INFO [finetune.py:976] (3/7) Epoch 29, batch 4600, loss[loss=0.1358, simple_loss=0.2047, pruned_loss=0.03346, over 4828.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2385, pruned_loss=0.04509, over 948502.02 frames. ], batch size: 30, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 05:01:29,217 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.461e+02 1.731e+02 1.968e+02 2.942e+02, threshold=3.463e+02, percent-clipped=1.0 +2023-04-28 05:01:36,145 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8350, 1.4011, 1.4269, 1.5969, 2.0113, 1.5967, 1.4351, 1.4309], + device='cuda:3'), covar=tensor([0.1599, 0.1525, 0.1983, 0.1309, 0.0774, 0.1746, 0.1896, 0.2401], + device='cuda:3'), in_proj_covar=tensor([0.0316, 0.0311, 0.0351, 0.0289, 0.0327, 0.0307, 0.0302, 0.0379], + device='cuda:3'), out_proj_covar=tensor([6.4270e-05, 6.3591e-05, 7.3405e-05, 5.7643e-05, 6.6704e-05, 6.3637e-05, + 6.2278e-05, 8.0043e-05], device='cuda:3') +2023-04-28 05:01:41,010 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3002, 1.7799, 2.2187, 2.6255, 2.2058, 1.7396, 1.4428, 1.9907], + device='cuda:3'), covar=tensor([0.3014, 0.2872, 0.1497, 0.2074, 0.2568, 0.2623, 0.4108, 0.1946], + device='cuda:3'), in_proj_covar=tensor([0.0295, 0.0248, 0.0230, 0.0318, 0.0224, 0.0238, 0.0231, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 05:02:19,693 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165014.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:02:20,900 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165016.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:02:32,274 INFO [finetune.py:976] (3/7) Epoch 29, batch 4650, loss[loss=0.1667, simple_loss=0.2314, pruned_loss=0.05101, over 4826.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2371, pruned_loss=0.04505, over 949693.08 frames. ], batch size: 39, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 05:03:05,234 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3843, 1.3061, 1.6318, 1.5806, 1.2543, 1.2059, 1.2609, 0.7272], + device='cuda:3'), covar=tensor([0.0481, 0.0530, 0.0332, 0.0512, 0.0631, 0.0982, 0.0460, 0.0519], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0076, 0.0096, 0.0073, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 05:03:18,783 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165064.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:03:37,007 INFO [finetune.py:976] (3/7) Epoch 29, batch 4700, loss[loss=0.1356, simple_loss=0.2065, pruned_loss=0.03236, over 4912.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.235, pruned_loss=0.04408, over 952721.16 frames. ], batch size: 36, lr: 2.83e-03, grad_scale: 16.0 +2023-04-28 05:03:40,047 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.538e+02 1.838e+02 2.257e+02 4.770e+02, threshold=3.676e+02, percent-clipped=2.0 +2023-04-28 05:04:09,169 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165102.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:04:41,314 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165123.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:04:42,458 INFO [finetune.py:976] (3/7) Epoch 29, batch 4750, loss[loss=0.1774, simple_loss=0.2519, pruned_loss=0.05144, over 4806.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2342, pruned_loss=0.04464, over 951767.12 frames. ], batch size: 41, lr: 2.83e-03, grad_scale: 16.0 +2023-04-28 05:05:25,306 INFO [finetune.py:976] (3/7) Epoch 29, batch 4800, loss[loss=0.1721, simple_loss=0.2489, pruned_loss=0.04768, over 4773.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2361, pruned_loss=0.045, over 952531.50 frames. ], batch size: 54, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:05:28,762 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.552e+02 1.836e+02 2.145e+02 4.672e+02, threshold=3.672e+02, percent-clipped=1.0 +2023-04-28 05:05:30,736 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.9351, 2.7771, 2.7888, 3.3318, 3.2231, 2.6970, 2.4091, 3.0988], + device='cuda:3'), covar=tensor([0.0732, 0.0854, 0.0647, 0.0548, 0.0494, 0.0867, 0.0661, 0.0448], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0202, 0.0185, 0.0171, 0.0178, 0.0178, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:05:31,316 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165184.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:05:42,375 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-28 05:05:57,721 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165224.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:05:58,219 INFO [finetune.py:976] (3/7) Epoch 29, batch 4850, loss[loss=0.1681, simple_loss=0.2488, pruned_loss=0.04368, over 4921.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2388, pruned_loss=0.0457, over 952172.73 frames. ], batch size: 42, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:06:03,035 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6947, 1.8310, 0.7975, 1.4336, 1.9977, 1.5822, 1.4044, 1.5728], + device='cuda:3'), covar=tensor([0.0476, 0.0352, 0.0327, 0.0536, 0.0237, 0.0489, 0.0489, 0.0562], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0054, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 05:06:05,321 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165234.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 05:06:29,993 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165272.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:06:31,766 INFO [finetune.py:976] (3/7) Epoch 29, batch 4900, loss[loss=0.1778, simple_loss=0.2575, pruned_loss=0.04902, over 4903.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2399, pruned_loss=0.04549, over 953366.76 frames. ], batch size: 36, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:06:35,317 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.561e+02 1.845e+02 2.104e+02 4.657e+02, threshold=3.691e+02, percent-clipped=1.0 +2023-04-28 05:06:37,062 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165282.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:06:38,955 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8241, 1.6667, 2.0246, 2.1781, 1.6113, 1.5013, 1.7676, 0.9932], + device='cuda:3'), covar=tensor([0.0569, 0.0567, 0.0493, 0.0774, 0.0782, 0.1071, 0.0548, 0.0629], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0076, 0.0095, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 05:06:56,342 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8957, 2.4652, 1.8144, 1.8533, 1.3524, 1.3837, 1.9423, 1.3042], + device='cuda:3'), covar=tensor([0.1809, 0.1272, 0.1518, 0.1631, 0.2387, 0.2050, 0.1011, 0.2120], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0209, 0.0171, 0.0204, 0.0201, 0.0186, 0.0156, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 05:06:57,502 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165314.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:07:01,797 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-04-28 05:07:04,967 INFO [finetune.py:976] (3/7) Epoch 29, batch 4950, loss[loss=0.1599, simple_loss=0.2337, pruned_loss=0.04305, over 4746.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2409, pruned_loss=0.04529, over 953796.98 frames. ], batch size: 27, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:07:29,753 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165362.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:07:32,225 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2005, 2.2146, 1.8728, 1.8941, 2.3066, 1.9295, 2.8488, 1.7198], + device='cuda:3'), covar=tensor([0.3969, 0.2126, 0.5212, 0.3392, 0.1789, 0.2739, 0.1280, 0.4780], + device='cuda:3'), in_proj_covar=tensor([0.0342, 0.0356, 0.0426, 0.0353, 0.0386, 0.0377, 0.0372, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:07:38,074 INFO [finetune.py:976] (3/7) Epoch 29, batch 5000, loss[loss=0.1289, simple_loss=0.2009, pruned_loss=0.02845, over 4267.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.239, pruned_loss=0.0448, over 953973.98 frames. ], batch size: 18, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:07:41,611 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.839e+01 1.556e+02 1.839e+02 2.298e+02 4.035e+02, threshold=3.677e+02, percent-clipped=3.0 +2023-04-28 05:07:57,197 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165402.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:08:01,455 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 05:08:12,132 INFO [finetune.py:976] (3/7) Epoch 29, batch 5050, loss[loss=0.2119, simple_loss=0.2769, pruned_loss=0.07345, over 4815.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.237, pruned_loss=0.04419, over 955068.61 frames. ], batch size: 40, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:08:29,767 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165450.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:08:45,561 INFO [finetune.py:976] (3/7) Epoch 29, batch 5100, loss[loss=0.1539, simple_loss=0.2236, pruned_loss=0.04207, over 4700.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2348, pruned_loss=0.04384, over 954009.84 frames. ], batch size: 23, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:08:48,562 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165479.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:08:49,075 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.177e+01 1.550e+02 1.878e+02 2.337e+02 3.681e+02, threshold=3.756e+02, percent-clipped=1.0 +2023-04-28 05:09:24,354 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0828, 2.0176, 1.8491, 1.7248, 2.2840, 1.7904, 2.7630, 1.7138], + device='cuda:3'), covar=tensor([0.3906, 0.2323, 0.4794, 0.3164, 0.1482, 0.2682, 0.1215, 0.4404], + device='cuda:3'), in_proj_covar=tensor([0.0339, 0.0353, 0.0423, 0.0351, 0.0383, 0.0374, 0.0369, 0.0422], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:09:47,677 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-28 05:09:48,075 INFO [finetune.py:976] (3/7) Epoch 29, batch 5150, loss[loss=0.206, simple_loss=0.2774, pruned_loss=0.06733, over 4804.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2355, pruned_loss=0.0443, over 953054.80 frames. ], batch size: 51, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:10:21,695 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165551.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:10:22,545 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.29 vs. limit=5.0 +2023-04-28 05:10:52,977 INFO [finetune.py:976] (3/7) Epoch 29, batch 5200, loss[loss=0.2357, simple_loss=0.3004, pruned_loss=0.08546, over 4817.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2383, pruned_loss=0.04519, over 953298.02 frames. ], batch size: 41, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:10:55,996 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.620e+02 1.941e+02 2.269e+02 3.767e+02, threshold=3.883e+02, percent-clipped=2.0 +2023-04-28 05:11:45,804 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165612.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:11:59,713 INFO [finetune.py:976] (3/7) Epoch 29, batch 5250, loss[loss=0.169, simple_loss=0.2563, pruned_loss=0.04084, over 4932.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2394, pruned_loss=0.04545, over 953281.65 frames. ], batch size: 42, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:13:07,360 INFO [finetune.py:976] (3/7) Epoch 29, batch 5300, loss[loss=0.1401, simple_loss=0.2159, pruned_loss=0.03209, over 4227.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2411, pruned_loss=0.04632, over 953493.59 frames. ], batch size: 18, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:13:16,017 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.838e+01 1.557e+02 1.826e+02 2.262e+02 6.421e+02, threshold=3.651e+02, percent-clipped=2.0 +2023-04-28 05:14:12,923 INFO [finetune.py:976] (3/7) Epoch 29, batch 5350, loss[loss=0.1845, simple_loss=0.2549, pruned_loss=0.05703, over 4814.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.241, pruned_loss=0.04612, over 954369.47 frames. ], batch size: 40, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:14:43,451 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165750.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:15:11,426 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1334, 2.5290, 0.7984, 1.4718, 1.5771, 1.9010, 1.6950, 0.8849], + device='cuda:3'), covar=tensor([0.1467, 0.1156, 0.1754, 0.1280, 0.1099, 0.0955, 0.1510, 0.1592], + device='cuda:3'), in_proj_covar=tensor([0.0117, 0.0236, 0.0135, 0.0120, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 05:15:21,306 INFO [finetune.py:976] (3/7) Epoch 29, batch 5400, loss[loss=0.1315, simple_loss=0.2016, pruned_loss=0.03072, over 4877.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2386, pruned_loss=0.04549, over 952489.32 frames. ], batch size: 31, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:15:24,348 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165779.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:15:24,845 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.500e+02 1.780e+02 2.167e+02 4.402e+02, threshold=3.560e+02, percent-clipped=1.0 +2023-04-28 05:15:35,372 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-28 05:15:42,340 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5598, 2.0347, 2.3653, 2.9814, 2.3357, 1.9282, 2.0563, 2.3543], + device='cuda:3'), covar=tensor([0.2759, 0.2909, 0.1560, 0.2049, 0.2630, 0.2444, 0.3408, 0.1918], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0315, 0.0223, 0.0236, 0.0230, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 05:15:55,091 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165803.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:07,215 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165811.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:21,415 INFO [finetune.py:976] (3/7) Epoch 29, batch 5450, loss[loss=0.1368, simple_loss=0.2156, pruned_loss=0.02898, over 4770.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2359, pruned_loss=0.04499, over 953946.92 frames. ], batch size: 28, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:16:22,709 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165827.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:48,228 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165864.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:55,273 INFO [finetune.py:976] (3/7) Epoch 29, batch 5500, loss[loss=0.175, simple_loss=0.2355, pruned_loss=0.05727, over 4904.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2321, pruned_loss=0.04336, over 955630.74 frames. ], batch size: 36, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:16:58,257 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.287e+01 1.513e+02 1.807e+02 2.218e+02 5.669e+02, threshold=3.614e+02, percent-clipped=2.0 +2023-04-28 05:17:16,268 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165907.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:17:29,075 INFO [finetune.py:976] (3/7) Epoch 29, batch 5550, loss[loss=0.1687, simple_loss=0.24, pruned_loss=0.04864, over 4810.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2341, pruned_loss=0.04403, over 956313.93 frames. ], batch size: 25, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:18:28,707 INFO [finetune.py:976] (3/7) Epoch 29, batch 5600, loss[loss=0.1818, simple_loss=0.2689, pruned_loss=0.04738, over 4810.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2373, pruned_loss=0.04444, over 955314.30 frames. ], batch size: 41, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:18:37,366 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.927e+01 1.529e+02 1.829e+02 2.121e+02 4.998e+02, threshold=3.659e+02, percent-clipped=4.0 +2023-04-28 05:19:32,708 INFO [finetune.py:976] (3/7) Epoch 29, batch 5650, loss[loss=0.219, simple_loss=0.2859, pruned_loss=0.07603, over 4758.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2413, pruned_loss=0.04542, over 954816.81 frames. ], batch size: 54, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:19:43,929 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9101, 1.8850, 1.8838, 1.6521, 2.1120, 1.6647, 2.6389, 1.6866], + device='cuda:3'), covar=tensor([0.3484, 0.1970, 0.4852, 0.2732, 0.1527, 0.2472, 0.1268, 0.4294], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0357, 0.0427, 0.0354, 0.0388, 0.0377, 0.0373, 0.0427], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:20:11,616 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166054.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 05:20:33,889 INFO [finetune.py:976] (3/7) Epoch 29, batch 5700, loss[loss=0.1467, simple_loss=0.2122, pruned_loss=0.04059, over 4173.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2366, pruned_loss=0.04386, over 941720.61 frames. ], batch size: 18, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:20:42,336 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 6.981e+01 1.433e+02 1.713e+02 2.148e+02 4.581e+02, threshold=3.425e+02, percent-clipped=1.0 +2023-04-28 05:21:19,169 INFO [finetune.py:976] (3/7) Epoch 30, batch 0, loss[loss=0.2225, simple_loss=0.2814, pruned_loss=0.08182, over 4914.00 frames. ], tot_loss[loss=0.2225, simple_loss=0.2814, pruned_loss=0.08182, over 4914.00 frames. ], batch size: 33, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:21:19,169 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 05:21:20,967 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9595, 1.2093, 1.8723, 2.3702, 2.0662, 1.8915, 1.8678, 1.8597], + device='cuda:3'), covar=tensor([0.4469, 0.7704, 0.7038, 0.5800, 0.6773, 0.7802, 0.8182, 0.8021], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0427, 0.0523, 0.0511, 0.0477, 0.0517, 0.0519, 0.0533], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:21:26,914 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3060, 1.6264, 1.9176, 2.0585, 2.0292, 2.0806, 1.9545, 1.9944], + device='cuda:3'), covar=tensor([0.3367, 0.5073, 0.3916, 0.3987, 0.4876, 0.5938, 0.4223, 0.4261], + device='cuda:3'), in_proj_covar=tensor([0.0347, 0.0377, 0.0333, 0.0346, 0.0354, 0.0396, 0.0365, 0.0336], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:21:34,681 INFO [finetune.py:1010] (3/7) Epoch 30, validation: loss=0.1551, simple_loss=0.2236, pruned_loss=0.04334, over 2265189.00 frames. +2023-04-28 05:21:34,682 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 05:21:37,611 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166106.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:21:48,580 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166115.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 05:22:23,307 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-28 05:22:31,247 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166147.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:22:35,225 INFO [finetune.py:976] (3/7) Epoch 30, batch 50, loss[loss=0.1866, simple_loss=0.2468, pruned_loss=0.06313, over 4769.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2453, pruned_loss=0.0487, over 218001.77 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:22:42,119 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-04-28 05:22:42,607 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-28 05:22:51,881 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166159.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:23:14,382 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 7.993e+01 1.354e+02 1.707e+02 2.159e+02 3.038e+02, threshold=3.414e+02, percent-clipped=0.0 +2023-04-28 05:23:47,367 INFO [finetune.py:976] (3/7) Epoch 30, batch 100, loss[loss=0.2038, simple_loss=0.2765, pruned_loss=0.0656, over 4902.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2365, pruned_loss=0.04514, over 381407.76 frames. ], batch size: 43, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:23:50,899 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166207.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:23:56,830 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166208.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:24:53,827 INFO [finetune.py:976] (3/7) Epoch 30, batch 150, loss[loss=0.1599, simple_loss=0.2285, pruned_loss=0.04562, over 4883.00 frames. ], tot_loss[loss=0.1603, simple_loss=0.2324, pruned_loss=0.04404, over 510135.47 frames. ], batch size: 35, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:24:55,631 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166255.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:25:06,806 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5189, 2.1425, 2.4438, 2.8486, 2.4032, 2.0358, 1.8178, 2.2411], + device='cuda:3'), covar=tensor([0.3020, 0.2815, 0.1613, 0.1975, 0.2581, 0.2614, 0.3464, 0.1750], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0316, 0.0223, 0.0237, 0.0230, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 05:25:14,637 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9235, 2.3332, 1.9623, 1.6345, 1.4273, 1.5208, 2.0039, 1.3864], + device='cuda:3'), covar=tensor([0.1689, 0.1313, 0.1463, 0.1755, 0.2264, 0.1970, 0.0975, 0.2067], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0210, 0.0170, 0.0204, 0.0201, 0.0187, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 05:25:15,374 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-28 05:25:28,825 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.799e+01 1.515e+02 1.784e+02 2.130e+02 3.221e+02, threshold=3.568e+02, percent-clipped=0.0 +2023-04-28 05:25:59,858 INFO [finetune.py:976] (3/7) Epoch 30, batch 200, loss[loss=0.119, simple_loss=0.1877, pruned_loss=0.02514, over 4766.00 frames. ], tot_loss[loss=0.1585, simple_loss=0.2306, pruned_loss=0.04319, over 609827.83 frames. ], batch size: 23, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:26:02,455 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166307.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:26:09,119 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1465, 1.6704, 2.0864, 2.3805, 2.0121, 1.6356, 1.2857, 1.7945], + device='cuda:3'), covar=tensor([0.3089, 0.2972, 0.1588, 0.2094, 0.2517, 0.2610, 0.4015, 0.1792], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0315, 0.0223, 0.0237, 0.0229, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 05:26:16,257 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5108, 2.0436, 2.3887, 2.9798, 2.3930, 1.8673, 1.9349, 2.3254], + device='cuda:3'), covar=tensor([0.2988, 0.2916, 0.1636, 0.2046, 0.2459, 0.2544, 0.3580, 0.1784], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0315, 0.0223, 0.0237, 0.0229, 0.0185], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 05:26:52,498 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8952, 2.3485, 1.9933, 1.7893, 1.4165, 1.4558, 1.9719, 1.3515], + device='cuda:3'), covar=tensor([0.1622, 0.1291, 0.1213, 0.1490, 0.2187, 0.1809, 0.0906, 0.1933], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0210, 0.0170, 0.0204, 0.0201, 0.0187, 0.0157, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 05:27:02,967 INFO [finetune.py:976] (3/7) Epoch 30, batch 250, loss[loss=0.1823, simple_loss=0.2577, pruned_loss=0.05346, over 4808.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2347, pruned_loss=0.04448, over 686684.18 frames. ], batch size: 51, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:27:19,832 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166368.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:27:38,574 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.540e+01 1.580e+02 1.903e+02 2.256e+02 3.495e+02, threshold=3.807e+02, percent-clipped=0.0 +2023-04-28 05:28:02,277 INFO [finetune.py:976] (3/7) Epoch 30, batch 300, loss[loss=0.1716, simple_loss=0.2584, pruned_loss=0.04238, over 4902.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2404, pruned_loss=0.04642, over 747750.09 frames. ], batch size: 35, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:28:07,537 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166406.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:28:09,963 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166410.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 05:29:01,975 INFO [finetune.py:976] (3/7) Epoch 30, batch 350, loss[loss=0.1438, simple_loss=0.2271, pruned_loss=0.03029, over 4789.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2429, pruned_loss=0.04722, over 794836.16 frames. ], batch size: 29, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:29:08,236 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166454.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:08,327 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8548, 2.1220, 2.1173, 2.3079, 2.0545, 2.0622, 2.1821, 2.0951], + device='cuda:3'), covar=tensor([0.3846, 0.6613, 0.5303, 0.4528, 0.5848, 0.7001, 0.6529, 0.5817], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0374, 0.0331, 0.0343, 0.0351, 0.0394, 0.0362, 0.0334], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:29:11,888 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166459.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:31,635 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.632e+01 1.563e+02 1.883e+02 2.303e+02 4.800e+02, threshold=3.766e+02, percent-clipped=2.0 +2023-04-28 05:29:32,406 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166481.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:36,890 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-28 05:29:38,379 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0331, 1.8040, 1.9770, 2.4018, 2.5402, 2.0005, 1.7330, 2.1937], + device='cuda:3'), covar=tensor([0.0856, 0.1146, 0.0762, 0.0591, 0.0551, 0.0824, 0.0737, 0.0547], + device='cuda:3'), in_proj_covar=tensor([0.0183, 0.0202, 0.0184, 0.0170, 0.0178, 0.0177, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:29:46,146 INFO [finetune.py:976] (3/7) Epoch 30, batch 400, loss[loss=0.1804, simple_loss=0.261, pruned_loss=0.04991, over 4835.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2421, pruned_loss=0.04611, over 831301.44 frames. ], batch size: 47, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:29:46,224 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166503.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:49,096 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166507.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:50,842 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166509.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:51,010 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-04-28 05:30:06,158 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166531.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:13,228 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166542.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:13,263 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.7582, 2.2777, 2.1466, 2.3070, 2.1225, 2.1776, 2.2065, 2.1407], + device='cuda:3'), covar=tensor([0.3848, 0.5263, 0.4730, 0.4114, 0.5315, 0.6012, 0.5707, 0.5319], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0374, 0.0331, 0.0343, 0.0352, 0.0395, 0.0363, 0.0334], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:30:19,818 INFO [finetune.py:976] (3/7) Epoch 30, batch 450, loss[loss=0.1283, simple_loss=0.2058, pruned_loss=0.02546, over 4753.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2398, pruned_loss=0.04549, over 856291.24 frames. ], batch size: 27, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:30:31,311 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166570.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:39,262 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.427e+02 1.810e+02 2.217e+02 8.020e+02, threshold=3.620e+02, percent-clipped=2.0 +2023-04-28 05:30:47,090 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166592.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:53,784 INFO [finetune.py:976] (3/7) Epoch 30, batch 500, loss[loss=0.1463, simple_loss=0.2204, pruned_loss=0.03608, over 4932.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2373, pruned_loss=0.04505, over 879028.66 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:31:27,594 INFO [finetune.py:976] (3/7) Epoch 30, batch 550, loss[loss=0.1277, simple_loss=0.2016, pruned_loss=0.02687, over 4334.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2337, pruned_loss=0.0439, over 896645.33 frames. ], batch size: 19, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:31:34,193 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166663.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:31:38,963 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3716, 1.3135, 1.3457, 1.5794, 1.5731, 1.3071, 1.0440, 1.4747], + device='cuda:3'), covar=tensor([0.0857, 0.1425, 0.0969, 0.0671, 0.0730, 0.0809, 0.0813, 0.0642], + device='cuda:3'), in_proj_covar=tensor([0.0183, 0.0202, 0.0183, 0.0170, 0.0178, 0.0177, 0.0149, 0.0175], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:31:45,973 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.489e+02 1.745e+02 2.160e+02 4.947e+02, threshold=3.489e+02, percent-clipped=2.0 +2023-04-28 05:31:48,294 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1036, 2.6016, 1.0960, 1.3690, 2.1024, 1.2025, 3.3385, 1.5939], + device='cuda:3'), covar=tensor([0.0684, 0.0630, 0.0760, 0.1235, 0.0460, 0.1013, 0.0242, 0.0646], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 05:31:53,711 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2438, 1.4441, 1.7819, 1.8763, 1.7891, 1.7859, 1.7926, 1.8159], + device='cuda:3'), covar=tensor([0.3978, 0.5235, 0.4364, 0.4247, 0.5414, 0.7031, 0.4754, 0.4481], + device='cuda:3'), in_proj_covar=tensor([0.0344, 0.0374, 0.0330, 0.0343, 0.0352, 0.0395, 0.0362, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:32:01,360 INFO [finetune.py:976] (3/7) Epoch 30, batch 600, loss[loss=0.1521, simple_loss=0.2322, pruned_loss=0.03601, over 4459.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2354, pruned_loss=0.04476, over 911604.37 frames. ], batch size: 19, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:32:05,761 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166710.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 05:32:23,079 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166735.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:32:34,881 INFO [finetune.py:976] (3/7) Epoch 30, batch 650, loss[loss=0.1674, simple_loss=0.233, pruned_loss=0.05086, over 4176.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2376, pruned_loss=0.04549, over 919410.35 frames. ], batch size: 65, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:32:37,944 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166758.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 05:32:47,102 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166764.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:32:58,773 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 05:33:07,157 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.523e+02 1.858e+02 2.285e+02 4.759e+02, threshold=3.715e+02, percent-clipped=1.0 +2023-04-28 05:33:29,705 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166796.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:33:39,477 INFO [finetune.py:976] (3/7) Epoch 30, batch 700, loss[loss=0.1639, simple_loss=0.2406, pruned_loss=0.04356, over 4842.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2393, pruned_loss=0.04566, over 927538.11 frames. ], batch size: 47, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:33:39,586 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166803.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:33:49,712 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5808, 1.0747, 4.3723, 4.1489, 3.7590, 4.2087, 4.0903, 3.8624], + device='cuda:3'), covar=tensor([0.7399, 0.6493, 0.1133, 0.1740, 0.1225, 0.1544, 0.1411, 0.1732], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0311, 0.0409, 0.0411, 0.0353, 0.0416, 0.0318, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:34:03,915 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166825.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:34:22,682 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166837.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:34:23,030 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-28 05:34:37,355 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166851.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:34:43,123 INFO [finetune.py:976] (3/7) Epoch 30, batch 750, loss[loss=0.1209, simple_loss=0.1872, pruned_loss=0.02726, over 4116.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.24, pruned_loss=0.04591, over 933818.60 frames. ], batch size: 18, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:34:51,921 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166865.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:35:11,850 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.441e+02 1.685e+02 2.007e+02 5.927e+02, threshold=3.370e+02, percent-clipped=1.0 +2023-04-28 05:35:22,367 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166887.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:35:44,637 INFO [finetune.py:976] (3/7) Epoch 30, batch 800, loss[loss=0.2017, simple_loss=0.2609, pruned_loss=0.07132, over 4855.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2389, pruned_loss=0.04522, over 936868.98 frames. ], batch size: 44, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:36:17,538 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166929.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:36:26,366 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0802, 2.4810, 0.7746, 1.5220, 1.5267, 1.9020, 1.6264, 0.8805], + device='cuda:3'), covar=tensor([0.1394, 0.1056, 0.1727, 0.1184, 0.1121, 0.0849, 0.1480, 0.1711], + device='cuda:3'), in_proj_covar=tensor([0.0118, 0.0240, 0.0137, 0.0121, 0.0133, 0.0154, 0.0118, 0.0119], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 05:36:49,758 INFO [finetune.py:976] (3/7) Epoch 30, batch 850, loss[loss=0.1855, simple_loss=0.257, pruned_loss=0.05703, over 4809.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2376, pruned_loss=0.04506, over 942079.32 frames. ], batch size: 41, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:36:59,926 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166963.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:37:20,968 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.548e+02 1.839e+02 2.172e+02 6.436e+02, threshold=3.678e+02, percent-clipped=1.0 +2023-04-28 05:37:37,955 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166990.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:37:46,360 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3053, 2.7114, 2.2814, 2.6440, 1.9579, 2.3222, 2.3301, 1.8256], + device='cuda:3'), covar=tensor([0.1621, 0.1129, 0.0860, 0.1087, 0.3122, 0.1202, 0.1560, 0.2163], + device='cuda:3'), in_proj_covar=tensor([0.0284, 0.0303, 0.0218, 0.0276, 0.0312, 0.0256, 0.0250, 0.0264], + device='cuda:3'), out_proj_covar=tensor([1.1298e-04, 1.1928e-04, 8.5521e-05, 1.0827e-04, 1.2560e-04, 1.0038e-04, + 1.0060e-04, 1.0366e-04], device='cuda:3') +2023-04-28 05:37:47,448 INFO [finetune.py:976] (3/7) Epoch 30, batch 900, loss[loss=0.1641, simple_loss=0.2404, pruned_loss=0.04386, over 4858.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2354, pruned_loss=0.04471, over 943110.08 frames. ], batch size: 49, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:37:52,366 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167011.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:38:13,068 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-28 05:38:21,333 INFO [finetune.py:976] (3/7) Epoch 30, batch 950, loss[loss=0.1701, simple_loss=0.2388, pruned_loss=0.05067, over 4822.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.234, pruned_loss=0.04429, over 947665.86 frames. ], batch size: 41, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:38:38,162 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.627e+02 1.847e+02 2.138e+02 3.460e+02, threshold=3.694e+02, percent-clipped=0.0 +2023-04-28 05:38:45,982 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167091.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:38:52,938 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167100.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:38:55,142 INFO [finetune.py:976] (3/7) Epoch 30, batch 1000, loss[loss=0.1692, simple_loss=0.2351, pruned_loss=0.05164, over 4833.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2362, pruned_loss=0.04486, over 947716.47 frames. ], batch size: 30, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:38:55,872 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167104.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:38:56,503 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.1983, 2.9103, 3.0324, 3.6374, 3.5873, 2.9935, 2.7342, 3.3808], + device='cuda:3'), covar=tensor([0.0690, 0.0921, 0.0591, 0.0444, 0.0437, 0.0763, 0.0607, 0.0416], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0204, 0.0184, 0.0171, 0.0179, 0.0179, 0.0150, 0.0177], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:39:06,125 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167120.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:13,540 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3673, 1.4184, 1.2784, 1.6075, 1.5527, 1.7347, 1.3978, 3.0771], + device='cuda:3'), covar=tensor([0.0577, 0.0845, 0.0815, 0.1290, 0.0633, 0.0458, 0.0722, 0.0151], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 05:39:17,578 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167137.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:28,662 INFO [finetune.py:976] (3/7) Epoch 30, batch 1050, loss[loss=0.1729, simple_loss=0.2522, pruned_loss=0.0468, over 4802.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.24, pruned_loss=0.04588, over 949647.89 frames. ], batch size: 51, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:39:34,181 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167161.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:36,582 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:36,605 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:46,135 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.567e+02 1.873e+02 2.278e+02 3.738e+02, threshold=3.746e+02, percent-clipped=2.0 +2023-04-28 05:39:49,155 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167185.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:50,851 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167187.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:54,423 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7025, 1.5642, 1.7942, 2.0104, 2.1249, 1.6243, 1.2786, 1.8908], + device='cuda:3'), covar=tensor([0.0718, 0.1183, 0.0687, 0.0518, 0.0523, 0.0745, 0.0714, 0.0484], + device='cuda:3'), in_proj_covar=tensor([0.0183, 0.0202, 0.0182, 0.0170, 0.0178, 0.0177, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:39:56,828 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8495, 1.1825, 3.2788, 3.0437, 2.8974, 3.1966, 3.2026, 2.9036], + device='cuda:3'), covar=tensor([0.7810, 0.5281, 0.1570, 0.2173, 0.1608, 0.1829, 0.2012, 0.1909], + device='cuda:3'), in_proj_covar=tensor([0.0311, 0.0311, 0.0409, 0.0411, 0.0354, 0.0416, 0.0318, 0.0365], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:40:01,670 INFO [finetune.py:976] (3/7) Epoch 30, batch 1100, loss[loss=0.1827, simple_loss=0.263, pruned_loss=0.05118, over 4912.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.242, pruned_loss=0.04689, over 952574.88 frames. ], batch size: 36, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:40:08,809 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167213.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:40:10,106 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5358, 3.0776, 2.7673, 2.8676, 2.2217, 2.7622, 2.7096, 2.1717], + device='cuda:3'), covar=tensor([0.1904, 0.1243, 0.0674, 0.1135, 0.3064, 0.1090, 0.2063, 0.2466], + device='cuda:3'), in_proj_covar=tensor([0.0281, 0.0301, 0.0217, 0.0275, 0.0309, 0.0254, 0.0249, 0.0262], + device='cuda:3'), out_proj_covar=tensor([1.1210e-04, 1.1847e-04, 8.5098e-05, 1.0771e-04, 1.2452e-04, 9.9679e-05, + 9.9965e-05, 1.0289e-04], device='cuda:3') +2023-04-28 05:40:22,666 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167235.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:40:32,852 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7632, 1.1830, 1.8212, 2.1789, 1.7983, 1.6924, 1.7540, 1.7425], + device='cuda:3'), covar=tensor([0.4403, 0.7033, 0.6244, 0.5347, 0.5925, 0.7931, 0.7749, 0.9162], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0426, 0.0522, 0.0508, 0.0476, 0.0515, 0.0517, 0.0531], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:40:34,492 INFO [finetune.py:976] (3/7) Epoch 30, batch 1150, loss[loss=0.1061, simple_loss=0.1725, pruned_loss=0.01984, over 4271.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2432, pruned_loss=0.04775, over 951481.55 frames. ], batch size: 18, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:40:49,504 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-04-28 05:41:03,290 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.590e+02 1.832e+02 2.168e+02 3.581e+02, threshold=3.663e+02, percent-clipped=0.0 +2023-04-28 05:41:11,483 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167285.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:41:33,522 INFO [finetune.py:976] (3/7) Epoch 30, batch 1200, loss[loss=0.1545, simple_loss=0.2236, pruned_loss=0.04266, over 4315.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.243, pruned_loss=0.04784, over 951903.26 frames. ], batch size: 65, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:42:03,695 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1761, 2.7175, 2.2012, 2.2399, 1.5797, 1.5949, 2.3442, 1.6495], + device='cuda:3'), covar=tensor([0.1786, 0.1518, 0.1504, 0.1717, 0.2438, 0.1982, 0.1008, 0.2111], + device='cuda:3'), in_proj_covar=tensor([0.0200, 0.0211, 0.0171, 0.0205, 0.0202, 0.0188, 0.0157, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 05:42:26,639 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7729, 1.6781, 1.8641, 2.1029, 2.2034, 1.7873, 1.3709, 1.9805], + device='cuda:3'), covar=tensor([0.0757, 0.1208, 0.0713, 0.0539, 0.0561, 0.0882, 0.0728, 0.0522], + device='cuda:3'), in_proj_covar=tensor([0.0181, 0.0201, 0.0181, 0.0169, 0.0177, 0.0176, 0.0148, 0.0174], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:42:37,939 INFO [finetune.py:976] (3/7) Epoch 30, batch 1250, loss[loss=0.1436, simple_loss=0.2173, pruned_loss=0.03493, over 4816.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2401, pruned_loss=0.04651, over 953489.22 frames. ], batch size: 30, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:43:19,229 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 7.674e+01 1.528e+02 1.825e+02 2.243e+02 4.898e+02, threshold=3.650e+02, percent-clipped=3.0 +2023-04-28 05:43:32,135 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167391.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:43:45,346 INFO [finetune.py:976] (3/7) Epoch 30, batch 1300, loss[loss=0.1443, simple_loss=0.2192, pruned_loss=0.03469, over 4788.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2368, pruned_loss=0.04538, over 955273.02 frames. ], batch size: 25, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:44:12,894 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167420.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:44:25,782 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167432.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:44:35,810 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167439.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:44:56,434 INFO [finetune.py:976] (3/7) Epoch 30, batch 1350, loss[loss=0.1413, simple_loss=0.212, pruned_loss=0.03531, over 4796.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2356, pruned_loss=0.04483, over 955357.02 frames. ], batch size: 25, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:44:58,349 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167456.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:45:00,774 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167460.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:45:17,747 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167468.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:45:25,586 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.814e+01 1.503e+02 1.795e+02 2.104e+02 5.763e+02, threshold=3.590e+02, percent-clipped=1.0 +2023-04-28 05:45:29,393 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5601, 1.7813, 1.9769, 1.9914, 1.8575, 1.9116, 1.9974, 1.9530], + device='cuda:3'), covar=tensor([0.3825, 0.5530, 0.4417, 0.4422, 0.5502, 0.7201, 0.5007, 0.4927], + device='cuda:3'), in_proj_covar=tensor([0.0345, 0.0375, 0.0332, 0.0344, 0.0353, 0.0394, 0.0363, 0.0336], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:45:34,225 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167493.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:45:41,215 INFO [finetune.py:976] (3/7) Epoch 30, batch 1400, loss[loss=0.1391, simple_loss=0.2194, pruned_loss=0.02938, over 4757.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2389, pruned_loss=0.04556, over 955566.29 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:46:06,102 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-28 05:46:06,645 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.1795, 2.0804, 1.8013, 1.8406, 2.1917, 1.7783, 2.6119, 1.5530], + device='cuda:3'), covar=tensor([0.3588, 0.1906, 0.4669, 0.2779, 0.1633, 0.2501, 0.1579, 0.4716], + device='cuda:3'), in_proj_covar=tensor([0.0336, 0.0351, 0.0419, 0.0348, 0.0382, 0.0371, 0.0368, 0.0420], + device='cuda:3'), out_proj_covar=tensor([9.9269e-05, 1.0443e-04, 1.2668e-04, 1.0376e-04, 1.1279e-04, 1.1007e-04, + 1.0704e-04, 1.2615e-04], device='cuda:3') +2023-04-28 05:46:14,261 INFO [finetune.py:976] (3/7) Epoch 30, batch 1450, loss[loss=0.1998, simple_loss=0.2753, pruned_loss=0.06215, over 4834.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2414, pruned_loss=0.0463, over 957595.81 frames. ], batch size: 49, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:46:43,887 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1722, 2.6156, 1.0689, 1.4420, 1.9707, 1.2474, 3.4544, 1.7829], + device='cuda:3'), covar=tensor([0.0613, 0.0551, 0.0750, 0.1149, 0.0463, 0.1008, 0.0198, 0.0579], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 05:46:52,055 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.458e+02 1.734e+02 2.083e+02 3.176e+02, threshold=3.469e+02, percent-clipped=0.0 +2023-04-28 05:47:00,606 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167585.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:47:21,141 INFO [finetune.py:976] (3/7) Epoch 30, batch 1500, loss[loss=0.1595, simple_loss=0.2408, pruned_loss=0.03911, over 4853.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2419, pruned_loss=0.0462, over 955452.15 frames. ], batch size: 31, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:47:25,089 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 05:47:51,967 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-28 05:47:56,734 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167633.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:48:20,795 INFO [finetune.py:976] (3/7) Epoch 30, batch 1550, loss[loss=0.1911, simple_loss=0.26, pruned_loss=0.06112, over 4828.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2422, pruned_loss=0.04637, over 955253.56 frames. ], batch size: 49, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:48:40,248 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.357e+01 1.558e+02 1.802e+02 2.131e+02 3.986e+02, threshold=3.604e+02, percent-clipped=1.0 +2023-04-28 05:48:50,788 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6407, 0.7292, 1.5699, 2.0105, 1.7103, 1.5448, 1.6157, 1.5694], + device='cuda:3'), covar=tensor([0.3965, 0.5800, 0.5108, 0.5217, 0.4853, 0.6245, 0.6370, 0.8064], + device='cuda:3'), in_proj_covar=tensor([0.0449, 0.0427, 0.0522, 0.0508, 0.0476, 0.0515, 0.0517, 0.0532], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:48:54,269 INFO [finetune.py:976] (3/7) Epoch 30, batch 1600, loss[loss=0.1678, simple_loss=0.2389, pruned_loss=0.04834, over 4893.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2402, pruned_loss=0.04608, over 956640.29 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:49:05,603 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167719.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:28,178 INFO [finetune.py:976] (3/7) Epoch 30, batch 1650, loss[loss=0.1477, simple_loss=0.2173, pruned_loss=0.03907, over 4852.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.236, pruned_loss=0.04421, over 955422.71 frames. ], batch size: 44, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:49:30,073 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167756.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:32,997 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167760.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:37,867 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167767.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:47,068 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.494e+02 1.761e+02 2.054e+02 5.504e+02, threshold=3.522e+02, percent-clipped=6.0 +2023-04-28 05:49:47,152 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([3.6348, 3.5625, 2.7026, 4.2572, 3.6612, 3.6610, 1.5842, 3.6244], + device='cuda:3'), covar=tensor([0.1911, 0.1509, 0.3965, 0.1830, 0.3875, 0.2056, 0.6437, 0.2808], + device='cuda:3'), in_proj_covar=tensor([0.0251, 0.0220, 0.0256, 0.0307, 0.0304, 0.0253, 0.0278, 0.0277], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:49:47,203 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167780.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:51,336 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-28 05:49:52,951 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167788.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:02,176 INFO [finetune.py:976] (3/7) Epoch 30, batch 1700, loss[loss=0.1405, simple_loss=0.2278, pruned_loss=0.02661, over 4773.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2332, pruned_loss=0.04329, over 955017.26 frames. ], batch size: 28, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:50:02,851 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167804.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:05,165 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-04-28 05:50:05,285 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167808.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:12,330 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2597, 1.5858, 1.3979, 1.7648, 1.7228, 1.9846, 1.4607, 3.5794], + device='cuda:3'), covar=tensor([0.0640, 0.0802, 0.0820, 0.1220, 0.0645, 0.0530, 0.0745, 0.0149], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 05:50:18,840 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167828.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:35,260 INFO [finetune.py:976] (3/7) Epoch 30, batch 1750, loss[loss=0.1417, simple_loss=0.2044, pruned_loss=0.03954, over 4703.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2334, pruned_loss=0.04321, over 955219.74 frames. ], batch size: 23, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:50:38,436 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.9172, 1.1510, 3.2443, 3.0658, 2.9021, 3.1718, 3.1228, 2.8684], + device='cuda:3'), covar=tensor([0.7637, 0.5631, 0.1521, 0.2137, 0.1464, 0.1927, 0.2449, 0.1773], + device='cuda:3'), in_proj_covar=tensor([0.0310, 0.0310, 0.0408, 0.0411, 0.0351, 0.0415, 0.0316, 0.0364], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:50:53,218 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.596e+01 1.560e+02 1.890e+02 2.217e+02 7.846e+02, threshold=3.780e+02, percent-clipped=0.0 +2023-04-28 05:51:08,164 INFO [finetune.py:976] (3/7) Epoch 30, batch 1800, loss[loss=0.1519, simple_loss=0.2249, pruned_loss=0.03948, over 4890.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2369, pruned_loss=0.04414, over 956274.27 frames. ], batch size: 32, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:51:27,330 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9344, 1.2056, 1.6208, 1.7303, 1.7244, 1.7552, 1.6387, 1.6135], + device='cuda:3'), covar=tensor([0.3603, 0.4599, 0.3714, 0.3744, 0.4823, 0.6442, 0.3774, 0.4163], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0373, 0.0331, 0.0342, 0.0351, 0.0392, 0.0361, 0.0334], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:51:41,397 INFO [finetune.py:976] (3/7) Epoch 30, batch 1850, loss[loss=0.1954, simple_loss=0.2741, pruned_loss=0.05834, over 4801.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2381, pruned_loss=0.04479, over 954725.60 frames. ], batch size: 41, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:52:04,616 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.506e+02 1.821e+02 2.216e+02 4.142e+02, threshold=3.642e+02, percent-clipped=2.0 +2023-04-28 05:52:26,316 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0812, 1.0789, 1.2215, 1.2259, 1.0464, 0.9425, 1.0138, 0.6001], + device='cuda:3'), covar=tensor([0.0515, 0.0598, 0.0463, 0.0500, 0.0684, 0.1227, 0.0511, 0.0609], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0064, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 05:52:26,792 INFO [finetune.py:976] (3/7) Epoch 30, batch 1900, loss[loss=0.1796, simple_loss=0.2468, pruned_loss=0.05624, over 4791.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2405, pruned_loss=0.04582, over 956531.24 frames. ], batch size: 25, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:53:09,302 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168035.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:53:21,015 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.5559, 1.8359, 1.9764, 2.0609, 1.9689, 2.0098, 1.9986, 1.9765], + device='cuda:3'), covar=tensor([0.3703, 0.5532, 0.4581, 0.4539, 0.5150, 0.6614, 0.5027, 0.4885], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0372, 0.0331, 0.0341, 0.0351, 0.0393, 0.0360, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 05:53:31,847 INFO [finetune.py:976] (3/7) Epoch 30, batch 1950, loss[loss=0.1556, simple_loss=0.2296, pruned_loss=0.04078, over 4901.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2387, pruned_loss=0.04462, over 955421.85 frames. ], batch size: 43, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:53:32,597 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1131, 1.6710, 1.9122, 2.1967, 1.9963, 1.6030, 1.3316, 1.8027], + device='cuda:3'), covar=tensor([0.2384, 0.2610, 0.1357, 0.1672, 0.2003, 0.2234, 0.3927, 0.1665], + device='cuda:3'), in_proj_covar=tensor([0.0293, 0.0247, 0.0230, 0.0316, 0.0224, 0.0237, 0.0230, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 05:53:56,263 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168075.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:54:04,670 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.564e+02 1.911e+02 2.576e+02 9.918e+02, threshold=3.821e+02, percent-clipped=1.0 +2023-04-28 05:54:13,533 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168088.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:54:24,497 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=168096.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 05:54:25,688 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-28 05:54:32,406 INFO [finetune.py:976] (3/7) Epoch 30, batch 2000, loss[loss=0.1101, simple_loss=0.1725, pruned_loss=0.02383, over 4775.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2359, pruned_loss=0.04414, over 956270.84 frames. ], batch size: 23, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:54:55,109 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168123.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:55:13,715 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168136.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:55:30,830 INFO [finetune.py:976] (3/7) Epoch 30, batch 2050, loss[loss=0.1424, simple_loss=0.2185, pruned_loss=0.03314, over 4761.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2333, pruned_loss=0.04371, over 956811.65 frames. ], batch size: 27, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:55:40,692 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9093, 2.4540, 1.9834, 2.3600, 1.6697, 2.1153, 1.9771, 1.4355], + device='cuda:3'), covar=tensor([0.2041, 0.1155, 0.0873, 0.1037, 0.3697, 0.1260, 0.1896, 0.2995], + device='cuda:3'), in_proj_covar=tensor([0.0279, 0.0298, 0.0216, 0.0271, 0.0306, 0.0250, 0.0245, 0.0258], + device='cuda:3'), out_proj_covar=tensor([1.1114e-04, 1.1711e-04, 8.4729e-05, 1.0651e-04, 1.2303e-04, 9.8215e-05, + 9.8635e-05, 1.0136e-04], device='cuda:3') +2023-04-28 05:55:43,767 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9253, 2.2395, 1.9104, 1.6702, 1.4447, 1.4646, 1.8735, 1.3626], + device='cuda:3'), covar=tensor([0.1468, 0.1226, 0.1263, 0.1550, 0.1996, 0.1740, 0.0912, 0.1856], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0210, 0.0170, 0.0205, 0.0201, 0.0188, 0.0157, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 05:55:44,975 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6647, 1.4965, 1.7068, 2.0157, 2.0298, 1.5844, 1.3145, 1.7721], + device='cuda:3'), covar=tensor([0.0756, 0.1209, 0.0761, 0.0515, 0.0575, 0.0773, 0.0749, 0.0555], + device='cuda:3'), in_proj_covar=tensor([0.0181, 0.0201, 0.0181, 0.0169, 0.0176, 0.0176, 0.0148, 0.0174], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:55:47,310 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.348e+01 1.497e+02 1.684e+02 2.075e+02 3.378e+02, threshold=3.368e+02, percent-clipped=0.0 +2023-04-28 05:56:04,273 INFO [finetune.py:976] (3/7) Epoch 30, batch 2100, loss[loss=0.1776, simple_loss=0.2441, pruned_loss=0.05557, over 4828.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2339, pruned_loss=0.04453, over 956519.33 frames. ], batch size: 30, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:56:23,097 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168233.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:56:37,578 INFO [finetune.py:976] (3/7) Epoch 30, batch 2150, loss[loss=0.1631, simple_loss=0.2535, pruned_loss=0.03635, over 4790.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2366, pruned_loss=0.04492, over 956322.76 frames. ], batch size: 29, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:56:40,067 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.22 vs. limit=5.0 +2023-04-28 05:56:54,476 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.582e+02 1.868e+02 2.335e+02 3.831e+02, threshold=3.737e+02, percent-clipped=1.0 +2023-04-28 05:57:04,558 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=168294.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:57:09,875 INFO [finetune.py:976] (3/7) Epoch 30, batch 2200, loss[loss=0.1875, simple_loss=0.261, pruned_loss=0.05698, over 4906.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2389, pruned_loss=0.04561, over 954355.52 frames. ], batch size: 46, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:58:09,636 INFO [finetune.py:976] (3/7) Epoch 30, batch 2250, loss[loss=0.1503, simple_loss=0.2309, pruned_loss=0.03487, over 4889.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2404, pruned_loss=0.04621, over 953263.68 frames. ], batch size: 43, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:58:40,749 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168375.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:58:43,712 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.481e+02 1.883e+02 2.203e+02 3.675e+02, threshold=3.765e+02, percent-clipped=0.0 +2023-04-28 05:58:55,049 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168391.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 05:59:12,294 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7222, 1.3030, 1.8056, 2.2429, 1.8853, 1.7213, 1.7716, 1.7356], + device='cuda:3'), covar=tensor([0.4108, 0.6385, 0.5822, 0.4852, 0.5027, 0.7195, 0.7342, 0.9546], + device='cuda:3'), in_proj_covar=tensor([0.0449, 0.0426, 0.0521, 0.0508, 0.0475, 0.0515, 0.0516, 0.0530], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 05:59:13,986 INFO [finetune.py:976] (3/7) Epoch 30, batch 2300, loss[loss=0.1635, simple_loss=0.2425, pruned_loss=0.04225, over 4821.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2406, pruned_loss=0.04619, over 951891.64 frames. ], batch size: 30, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:59:17,295 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-28 05:59:38,058 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:59:38,105 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:00:16,726 INFO [finetune.py:976] (3/7) Epoch 30, batch 2350, loss[loss=0.1239, simple_loss=0.2096, pruned_loss=0.01917, over 4835.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.239, pruned_loss=0.04562, over 951307.84 frames. ], batch size: 30, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:00:37,862 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168471.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:00:48,857 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.790e+01 1.485e+02 1.765e+02 2.221e+02 4.572e+02, threshold=3.529e+02, percent-clipped=1.0 +2023-04-28 06:01:21,063 INFO [finetune.py:976] (3/7) Epoch 30, batch 2400, loss[loss=0.1775, simple_loss=0.2472, pruned_loss=0.05395, over 4927.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2356, pruned_loss=0.04461, over 953200.64 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:02:28,241 INFO [finetune.py:976] (3/7) Epoch 30, batch 2450, loss[loss=0.1731, simple_loss=0.2385, pruned_loss=0.0538, over 4832.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2329, pruned_loss=0.04397, over 956288.90 frames. ], batch size: 30, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:03:10,382 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.524e+02 1.793e+02 2.156e+02 5.076e+02, threshold=3.587e+02, percent-clipped=1.0 +2023-04-28 06:03:21,584 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168589.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:03:36,320 INFO [finetune.py:976] (3/7) Epoch 30, batch 2500, loss[loss=0.1699, simple_loss=0.2447, pruned_loss=0.04758, over 4764.00 frames. ], tot_loss[loss=0.162, simple_loss=0.235, pruned_loss=0.0445, over 954652.62 frames. ], batch size: 27, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:04:49,840 INFO [finetune.py:976] (3/7) Epoch 30, batch 2550, loss[loss=0.1742, simple_loss=0.2435, pruned_loss=0.05248, over 4861.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2383, pruned_loss=0.04501, over 953330.30 frames. ], batch size: 31, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:05:23,464 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.608e+02 1.886e+02 2.283e+02 4.830e+02, threshold=3.772e+02, percent-clipped=5.0 +2023-04-28 06:05:31,396 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168691.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:05:38,669 INFO [finetune.py:976] (3/7) Epoch 30, batch 2600, loss[loss=0.1989, simple_loss=0.2808, pruned_loss=0.0585, over 4906.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2403, pruned_loss=0.04557, over 954175.94 frames. ], batch size: 36, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:05:53,354 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9936, 1.1954, 4.6843, 4.4068, 4.1248, 4.4480, 4.1875, 4.1405], + device='cuda:3'), covar=tensor([0.7226, 0.6571, 0.1202, 0.2077, 0.1208, 0.1724, 0.2282, 0.1705], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0312, 0.0410, 0.0414, 0.0353, 0.0419, 0.0317, 0.0366], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:06:26,240 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168739.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:06:40,212 INFO [finetune.py:976] (3/7) Epoch 30, batch 2650, loss[loss=0.168, simple_loss=0.2364, pruned_loss=0.04979, over 4878.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2404, pruned_loss=0.04534, over 955915.83 frames. ], batch size: 32, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:07:17,594 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.537e+02 1.777e+02 2.075e+02 3.537e+02, threshold=3.553e+02, percent-clipped=0.0 +2023-04-28 06:07:44,074 INFO [finetune.py:976] (3/7) Epoch 30, batch 2700, loss[loss=0.1486, simple_loss=0.2282, pruned_loss=0.03449, over 4756.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.239, pruned_loss=0.04439, over 955463.83 frames. ], batch size: 27, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:08:02,164 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0730, 2.5720, 1.1453, 1.3898, 2.0735, 1.2778, 3.3847, 1.6865], + device='cuda:3'), covar=tensor([0.0731, 0.0732, 0.0820, 0.1220, 0.0460, 0.1010, 0.0192, 0.0627], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0046, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 06:08:10,795 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-28 06:08:17,674 INFO [finetune.py:976] (3/7) Epoch 30, batch 2750, loss[loss=0.1584, simple_loss=0.2362, pruned_loss=0.0403, over 4927.00 frames. ], tot_loss[loss=0.162, simple_loss=0.236, pruned_loss=0.04402, over 956254.52 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:08:27,443 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-28 06:08:28,479 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.4088, 1.5972, 1.8009, 1.8945, 1.8650, 1.8712, 1.8103, 1.8721], + device='cuda:3'), covar=tensor([0.3888, 0.5457, 0.4316, 0.4535, 0.5234, 0.6761, 0.4962, 0.4532], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0373, 0.0330, 0.0342, 0.0350, 0.0393, 0.0361, 0.0333], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 06:08:35,307 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.781e+01 1.528e+02 1.886e+02 2.380e+02 6.138e+02, threshold=3.773e+02, percent-clipped=2.0 +2023-04-28 06:08:40,441 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-28 06:08:41,775 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168889.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:08:50,767 INFO [finetune.py:976] (3/7) Epoch 30, batch 2800, loss[loss=0.1246, simple_loss=0.1957, pruned_loss=0.02673, over 4798.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.2335, pruned_loss=0.04348, over 957728.01 frames. ], batch size: 26, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:09:12,961 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168937.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:09:24,481 INFO [finetune.py:976] (3/7) Epoch 30, batch 2850, loss[loss=0.1682, simple_loss=0.2335, pruned_loss=0.05141, over 4825.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2326, pruned_loss=0.0438, over 955295.26 frames. ], batch size: 30, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:09:33,160 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168967.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:09:34,348 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168969.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 06:09:41,978 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.589e+01 1.497e+02 1.758e+02 2.128e+02 6.037e+02, threshold=3.517e+02, percent-clipped=2.0 +2023-04-28 06:09:58,542 INFO [finetune.py:976] (3/7) Epoch 30, batch 2900, loss[loss=0.1818, simple_loss=0.2554, pruned_loss=0.05407, over 4921.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2349, pruned_loss=0.04409, over 956454.99 frames. ], batch size: 36, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:10:14,814 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169028.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:10:16,027 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169030.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 06:10:19,554 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6371, 2.7522, 2.2329, 2.4750, 2.7759, 2.3467, 3.7213, 2.1440], + device='cuda:3'), covar=tensor([0.3874, 0.2107, 0.4463, 0.3293, 0.1920, 0.2658, 0.1497, 0.4242], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0358, 0.0428, 0.0354, 0.0387, 0.0378, 0.0374, 0.0427], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:10:31,849 INFO [finetune.py:976] (3/7) Epoch 30, batch 2950, loss[loss=0.1811, simple_loss=0.2529, pruned_loss=0.0546, over 4893.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2373, pruned_loss=0.04425, over 955543.02 frames. ], batch size: 32, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:10:44,284 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-28 06:10:49,917 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.313e+01 1.646e+02 1.860e+02 2.214e+02 4.123e+02, threshold=3.720e+02, percent-clipped=2.0 +2023-04-28 06:10:50,623 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169082.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:11:05,769 INFO [finetune.py:976] (3/7) Epoch 30, batch 3000, loss[loss=0.1595, simple_loss=0.2309, pruned_loss=0.04399, over 4768.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2389, pruned_loss=0.04505, over 954299.46 frames. ], batch size: 26, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:11:05,769 INFO [finetune.py:1001] (3/7) Computing validation loss +2023-04-28 06:11:08,372 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7484, 1.6168, 1.8344, 2.0867, 2.1687, 1.6674, 1.4659, 1.9563], + device='cuda:3'), covar=tensor([0.0711, 0.1183, 0.0712, 0.0542, 0.0539, 0.0815, 0.0729, 0.0511], + device='cuda:3'), in_proj_covar=tensor([0.0187, 0.0207, 0.0187, 0.0174, 0.0181, 0.0182, 0.0152, 0.0180], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:11:10,655 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3049, 1.8503, 2.1913, 2.3925, 2.2035, 1.8132, 1.2655, 1.9093], + device='cuda:3'), covar=tensor([0.3297, 0.2997, 0.1452, 0.1921, 0.2416, 0.2714, 0.3930, 0.2035], + device='cuda:3'), in_proj_covar=tensor([0.0292, 0.0246, 0.0229, 0.0315, 0.0223, 0.0236, 0.0230, 0.0186], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 06:11:16,543 INFO [finetune.py:1010] (3/7) Epoch 30, validation: loss=0.1534, simple_loss=0.2215, pruned_loss=0.04259, over 2265189.00 frames. +2023-04-28 06:11:16,543 INFO [finetune.py:1011] (3/7) Maximum memory allocated so far is 6447MB +2023-04-28 06:11:16,666 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3755, 1.2579, 1.6001, 1.5514, 1.2157, 1.1476, 1.2112, 0.7200], + device='cuda:3'), covar=tensor([0.0483, 0.0587, 0.0322, 0.0529, 0.0729, 0.1058, 0.0543, 0.0523], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0068, 0.0065, 0.0069, 0.0076, 0.0095, 0.0073, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 06:11:24,701 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169116.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:11:52,067 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169143.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:12:04,396 INFO [finetune.py:976] (3/7) Epoch 30, batch 3050, loss[loss=0.1631, simple_loss=0.2447, pruned_loss=0.04074, over 4923.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2404, pruned_loss=0.04521, over 954753.14 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:12:13,892 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169159.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:12:36,626 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169177.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:12:42,081 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.505e+02 1.782e+02 2.187e+02 4.062e+02, threshold=3.563e+02, percent-clipped=1.0 +2023-04-28 06:13:04,822 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-28 06:13:06,207 INFO [finetune.py:976] (3/7) Epoch 30, batch 3100, loss[loss=0.1677, simple_loss=0.2272, pruned_loss=0.05405, over 4905.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.238, pruned_loss=0.04429, over 953312.66 frames. ], batch size: 36, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:13:15,602 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4082, 1.4360, 3.7304, 3.4445, 3.3852, 3.5191, 3.6340, 3.2877], + device='cuda:3'), covar=tensor([0.6761, 0.5076, 0.1465, 0.2435, 0.1238, 0.1895, 0.1249, 0.1943], + device='cuda:3'), in_proj_covar=tensor([0.0309, 0.0307, 0.0405, 0.0408, 0.0348, 0.0414, 0.0314, 0.0363], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:13:29,644 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169220.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:13:39,363 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9083, 1.9945, 1.1479, 1.5752, 2.1576, 1.7074, 1.5916, 1.7811], + device='cuda:3'), covar=tensor([0.0448, 0.0334, 0.0271, 0.0522, 0.0236, 0.0472, 0.0471, 0.0510], + device='cuda:3'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 06:13:50,611 INFO [finetune.py:976] (3/7) Epoch 30, batch 3150, loss[loss=0.1749, simple_loss=0.2529, pruned_loss=0.04843, over 4889.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2353, pruned_loss=0.04353, over 955211.38 frames. ], batch size: 35, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:14:10,081 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.051e+01 1.353e+02 1.739e+02 2.048e+02 6.582e+02, threshold=3.478e+02, percent-clipped=1.0 +2023-04-28 06:14:14,484 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8301, 1.2626, 1.8698, 2.3126, 1.8952, 1.7530, 1.8231, 1.7757], + device='cuda:3'), covar=tensor([0.4276, 0.6744, 0.6026, 0.5075, 0.5717, 0.7545, 0.7610, 0.8809], + device='cuda:3'), in_proj_covar=tensor([0.0446, 0.0426, 0.0519, 0.0506, 0.0473, 0.0514, 0.0513, 0.0529], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:14:23,996 INFO [finetune.py:976] (3/7) Epoch 30, batch 3200, loss[loss=0.1339, simple_loss=0.2149, pruned_loss=0.02642, over 4902.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2323, pruned_loss=0.04297, over 954694.86 frames. ], batch size: 32, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:14:38,827 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169323.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:14:40,576 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169325.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 06:14:42,426 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8473, 1.5221, 1.7776, 2.1203, 2.2047, 1.7004, 1.6274, 1.9712], + device='cuda:3'), covar=tensor([0.0747, 0.1303, 0.0748, 0.0559, 0.0576, 0.0851, 0.0683, 0.0530], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0205, 0.0185, 0.0172, 0.0179, 0.0180, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:14:50,920 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.7897, 1.6324, 1.8388, 2.1180, 2.1819, 1.6979, 1.4495, 2.0325], + device='cuda:3'), covar=tensor([0.0832, 0.1331, 0.0884, 0.0590, 0.0581, 0.0872, 0.0754, 0.0529], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0205, 0.0185, 0.0172, 0.0179, 0.0180, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:14:57,968 INFO [finetune.py:976] (3/7) Epoch 30, batch 3250, loss[loss=0.1855, simple_loss=0.2633, pruned_loss=0.05381, over 4739.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.2346, pruned_loss=0.04416, over 955060.09 frames. ], batch size: 54, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:15:18,032 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.661e+01 1.547e+02 1.939e+02 2.313e+02 4.246e+02, threshold=3.878e+02, percent-clipped=2.0 +2023-04-28 06:15:32,127 INFO [finetune.py:976] (3/7) Epoch 30, batch 3300, loss[loss=0.1776, simple_loss=0.2633, pruned_loss=0.04598, over 4811.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2404, pruned_loss=0.04624, over 956453.36 frames. ], batch size: 51, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:15:56,494 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169438.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:16:05,664 INFO [finetune.py:976] (3/7) Epoch 30, batch 3350, loss[loss=0.1313, simple_loss=0.2007, pruned_loss=0.03091, over 4821.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2406, pruned_loss=0.04601, over 955993.19 frames. ], batch size: 25, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:16:18,729 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169472.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:16:26,057 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.565e+02 1.911e+02 2.262e+02 9.005e+02, threshold=3.822e+02, percent-clipped=1.0 +2023-04-28 06:16:33,526 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9275, 1.4105, 1.7626, 1.7179, 1.7318, 1.4420, 0.8373, 1.4324], + device='cuda:3'), covar=tensor([0.2983, 0.3054, 0.1585, 0.1954, 0.2297, 0.2475, 0.4189, 0.1928], + device='cuda:3'), in_proj_covar=tensor([0.0295, 0.0248, 0.0231, 0.0318, 0.0225, 0.0238, 0.0232, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 06:16:39,411 INFO [finetune.py:976] (3/7) Epoch 30, batch 3400, loss[loss=0.1864, simple_loss=0.2613, pruned_loss=0.05576, over 4849.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2405, pruned_loss=0.04568, over 956314.96 frames. ], batch size: 44, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:16:47,268 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169515.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:16:51,891 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2047, 1.4862, 1.3473, 1.7715, 1.5955, 1.7015, 1.4322, 3.0854], + device='cuda:3'), covar=tensor([0.0665, 0.0823, 0.0833, 0.1244, 0.0676, 0.0477, 0.0750, 0.0158], + device='cuda:3'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 06:17:11,816 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-28 06:17:12,721 INFO [finetune.py:976] (3/7) Epoch 30, batch 3450, loss[loss=0.1874, simple_loss=0.2481, pruned_loss=0.0633, over 4815.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2388, pruned_loss=0.04455, over 956577.38 frames. ], batch size: 33, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:17:36,962 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.504e+02 1.780e+02 2.155e+02 3.271e+02, threshold=3.560e+02, percent-clipped=0.0 +2023-04-28 06:18:07,733 INFO [finetune.py:976] (3/7) Epoch 30, batch 3500, loss[loss=0.2141, simple_loss=0.2795, pruned_loss=0.0743, over 4852.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2378, pruned_loss=0.04488, over 957682.64 frames. ], batch size: 47, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:18:29,320 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169621.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:18:31,027 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169623.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:18:32,244 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169625.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 06:19:13,502 INFO [finetune.py:976] (3/7) Epoch 30, batch 3550, loss[loss=0.1713, simple_loss=0.2448, pruned_loss=0.04886, over 4898.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2354, pruned_loss=0.0443, over 958052.13 frames. ], batch size: 35, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:19:30,803 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169671.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:19:32,048 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169673.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 06:19:32,668 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0012, 1.7078, 1.9851, 2.3156, 2.3027, 1.8922, 1.6453, 2.1420], + device='cuda:3'), covar=tensor([0.0717, 0.1147, 0.0723, 0.0542, 0.0588, 0.0829, 0.0725, 0.0499], + device='cuda:3'), in_proj_covar=tensor([0.0185, 0.0205, 0.0185, 0.0173, 0.0180, 0.0181, 0.0152, 0.0179], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:19:42,308 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.126e+01 1.430e+02 1.728e+02 2.038e+02 3.209e+02, threshold=3.455e+02, percent-clipped=0.0 +2023-04-28 06:19:43,557 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169682.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:20:01,238 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3965, 2.8551, 0.9692, 1.4603, 2.0034, 1.3093, 3.7137, 1.6463], + device='cuda:3'), covar=tensor([0.0679, 0.1007, 0.0931, 0.1287, 0.0570, 0.1059, 0.0219, 0.0673], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0046, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 06:20:12,672 INFO [finetune.py:976] (3/7) Epoch 30, batch 3600, loss[loss=0.09953, simple_loss=0.1722, pruned_loss=0.01341, over 4761.00 frames. ], tot_loss[loss=0.1603, simple_loss=0.2332, pruned_loss=0.04372, over 957658.58 frames. ], batch size: 28, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:20:45,855 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169729.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:20:47,108 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6835, 1.7750, 1.6385, 2.1135, 1.9868, 2.0841, 1.6210, 4.5126], + device='cuda:3'), covar=tensor([0.0505, 0.0807, 0.0795, 0.1224, 0.0614, 0.0482, 0.0755, 0.0119], + device='cuda:3'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:3'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:3') +2023-04-28 06:20:57,495 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169738.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:21:10,903 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169747.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:21:20,084 INFO [finetune.py:976] (3/7) Epoch 30, batch 3650, loss[loss=0.1903, simple_loss=0.2711, pruned_loss=0.05477, over 4803.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2348, pruned_loss=0.04359, over 956063.93 frames. ], batch size: 45, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:21:43,298 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169772.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:21:49,830 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.8378, 2.1961, 2.3036, 2.3783, 2.2362, 2.2874, 2.3768, 2.2886], + device='cuda:3'), covar=tensor([0.3195, 0.4340, 0.3865, 0.3879, 0.4359, 0.5612, 0.4518, 0.4079], + device='cuda:3'), in_proj_covar=tensor([0.0343, 0.0375, 0.0331, 0.0343, 0.0351, 0.0394, 0.0362, 0.0335], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:3') +2023-04-28 06:21:53,880 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.801e+01 1.645e+02 1.930e+02 2.224e+02 4.571e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-04-28 06:22:03,119 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169786.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:03,163 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169786.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:11,320 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169790.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:25,957 INFO [finetune.py:976] (3/7) Epoch 30, batch 3700, loss[loss=0.1467, simple_loss=0.2236, pruned_loss=0.0349, over 4786.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2392, pruned_loss=0.04483, over 954938.21 frames. ], batch size: 28, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:22:26,937 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-28 06:22:35,264 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169808.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:44,963 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169815.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:47,943 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169820.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:23:11,051 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-28 06:23:22,789 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169847.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:23:31,475 INFO [finetune.py:976] (3/7) Epoch 30, batch 3750, loss[loss=0.1804, simple_loss=0.2527, pruned_loss=0.05404, over 4860.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2407, pruned_loss=0.04565, over 953222.66 frames. ], batch size: 31, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:23:42,220 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169863.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:24:04,806 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.513e+02 1.804e+02 2.185e+02 4.010e+02, threshold=3.607e+02, percent-clipped=1.0 +2023-04-28 06:24:14,129 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-28 06:24:14,491 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8607, 2.4397, 1.9295, 1.8531, 1.3395, 1.4073, 2.0201, 1.3191], + device='cuda:3'), covar=tensor([0.1647, 0.1243, 0.1370, 0.1577, 0.2202, 0.1925, 0.0924, 0.1972], + device='cuda:3'), in_proj_covar=tensor([0.0198, 0.0208, 0.0169, 0.0203, 0.0200, 0.0187, 0.0156, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 06:24:16,929 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4914, 1.3025, 1.7115, 1.6352, 1.3747, 1.2790, 1.3074, 0.8554], + device='cuda:3'), covar=tensor([0.0491, 0.0770, 0.0374, 0.0536, 0.0694, 0.1052, 0.0519, 0.0596], + device='cuda:3'), in_proj_covar=tensor([0.0070, 0.0067, 0.0064, 0.0068, 0.0075, 0.0094, 0.0072, 0.0061], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 06:24:36,630 INFO [finetune.py:976] (3/7) Epoch 30, batch 3800, loss[loss=0.1868, simple_loss=0.2636, pruned_loss=0.05495, over 4905.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2415, pruned_loss=0.0459, over 955730.31 frames. ], batch size: 37, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:25:05,040 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.1593, 1.7783, 2.2483, 2.5879, 2.1733, 2.0348, 2.1608, 2.1648], + device='cuda:3'), covar=tensor([0.4526, 0.7504, 0.7079, 0.5414, 0.6226, 0.8153, 0.8653, 0.9775], + device='cuda:3'), in_proj_covar=tensor([0.0448, 0.0429, 0.0524, 0.0509, 0.0477, 0.0517, 0.0517, 0.0533], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:25:18,401 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6530, 1.5052, 1.7538, 1.9834, 2.0760, 1.6167, 1.4225, 1.8762], + device='cuda:3'), covar=tensor([0.0798, 0.1288, 0.0723, 0.0557, 0.0558, 0.0858, 0.0707, 0.0528], + device='cuda:3'), in_proj_covar=tensor([0.0184, 0.0205, 0.0185, 0.0173, 0.0179, 0.0180, 0.0151, 0.0178], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:25:40,862 INFO [finetune.py:976] (3/7) Epoch 30, batch 3850, loss[loss=0.1707, simple_loss=0.2357, pruned_loss=0.05285, over 4822.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2401, pruned_loss=0.0452, over 955823.92 frames. ], batch size: 40, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:26:12,313 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169977.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:26:14,680 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.720e+01 1.456e+02 1.655e+02 1.971e+02 4.065e+02, threshold=3.309e+02, percent-clipped=1.0 +2023-04-28 06:26:45,748 INFO [finetune.py:976] (3/7) Epoch 30, batch 3900, loss[loss=0.1308, simple_loss=0.2092, pruned_loss=0.02619, over 4773.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2378, pruned_loss=0.04464, over 956334.97 frames. ], batch size: 26, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:27:14,566 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 06:27:26,115 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-04-28 06:27:43,925 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=170045.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:27:48,655 INFO [finetune.py:976] (3/7) Epoch 30, batch 3950, loss[loss=0.1371, simple_loss=0.2098, pruned_loss=0.03218, over 4823.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2348, pruned_loss=0.0436, over 956321.01 frames. ], batch size: 30, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:27:53,830 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-04-28 06:28:24,771 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.459e+02 1.752e+02 2.144e+02 4.109e+02, threshold=3.505e+02, percent-clipped=2.0 +2023-04-28 06:28:27,245 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170085.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:28:46,926 INFO [finetune.py:976] (3/7) Epoch 30, batch 4000, loss[loss=0.1712, simple_loss=0.2533, pruned_loss=0.04455, over 4822.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2335, pruned_loss=0.04315, over 955307.43 frames. ], batch size: 39, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:28:47,013 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170103.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:28:53,796 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5822, 1.5573, 0.8798, 1.3509, 1.5230, 1.4596, 1.3941, 1.4697], + device='cuda:3'), covar=tensor([0.0452, 0.0308, 0.0372, 0.0481, 0.0329, 0.0436, 0.0415, 0.0490], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0027], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 06:28:53,821 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=170106.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:29:39,694 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170142.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:29:50,075 INFO [finetune.py:976] (3/7) Epoch 30, batch 4050, loss[loss=0.1717, simple_loss=0.2577, pruned_loss=0.04284, over 4808.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2372, pruned_loss=0.04447, over 952889.04 frames. ], batch size: 45, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:30:17,102 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-28 06:30:22,175 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.558e+02 1.902e+02 2.221e+02 4.560e+02, threshold=3.805e+02, percent-clipped=3.0 +2023-04-28 06:30:36,566 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-28 06:30:36,789 INFO [finetune.py:976] (3/7) Epoch 30, batch 4100, loss[loss=0.1308, simple_loss=0.2081, pruned_loss=0.02672, over 4927.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2392, pruned_loss=0.04517, over 952098.51 frames. ], batch size: 33, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:30:58,761 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.9149, 1.4152, 1.5631, 1.6047, 2.0209, 1.6823, 1.3969, 1.4787], + device='cuda:3'), covar=tensor([0.1510, 0.1459, 0.1706, 0.1231, 0.0917, 0.1427, 0.1917, 0.2045], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0309, 0.0352, 0.0288, 0.0325, 0.0308, 0.0301, 0.0377], + device='cuda:3'), out_proj_covar=tensor([6.3984e-05, 6.2943e-05, 7.3543e-05, 5.7450e-05, 6.6065e-05, 6.3751e-05, + 6.2092e-05, 7.9485e-05], device='cuda:3') +2023-04-28 06:31:10,087 INFO [finetune.py:976] (3/7) Epoch 30, batch 4150, loss[loss=0.1794, simple_loss=0.261, pruned_loss=0.04888, over 4919.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2409, pruned_loss=0.04563, over 952734.48 frames. ], batch size: 41, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:31:26,665 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170277.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:31:29,027 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.851e+01 1.521e+02 1.781e+02 2.076e+02 3.722e+02, threshold=3.562e+02, percent-clipped=0.0 +2023-04-28 06:31:42,966 INFO [finetune.py:976] (3/7) Epoch 30, batch 4200, loss[loss=0.1781, simple_loss=0.2484, pruned_loss=0.05389, over 4821.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2421, pruned_loss=0.04628, over 954711.08 frames. ], batch size: 38, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:31:58,819 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170325.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:32:15,661 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.2263, 2.7060, 1.1075, 1.5501, 2.0154, 1.3227, 3.5968, 1.9632], + device='cuda:3'), covar=tensor([0.0584, 0.0581, 0.0746, 0.1117, 0.0463, 0.0914, 0.0164, 0.0533], + device='cuda:3'), in_proj_covar=tensor([0.0050, 0.0064, 0.0046, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 06:32:16,768 INFO [finetune.py:976] (3/7) Epoch 30, batch 4250, loss[loss=0.1266, simple_loss=0.2046, pruned_loss=0.02432, over 4772.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2396, pruned_loss=0.04526, over 956054.33 frames. ], batch size: 26, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:32:18,774 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 06:32:36,242 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.843e+01 1.455e+02 1.729e+02 2.061e+02 4.081e+02, threshold=3.459e+02, percent-clipped=1.0 +2023-04-28 06:32:38,798 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170385.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:32:48,585 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170401.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:32:50,173 INFO [finetune.py:976] (3/7) Epoch 30, batch 4300, loss[loss=0.1547, simple_loss=0.2266, pruned_loss=0.04142, over 4764.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2363, pruned_loss=0.04436, over 956153.44 frames. ], batch size: 26, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:32:50,289 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170403.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:10,858 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170433.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:16,816 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170442.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:22,276 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170451.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:23,432 INFO [finetune.py:976] (3/7) Epoch 30, batch 4350, loss[loss=0.157, simple_loss=0.2272, pruned_loss=0.0434, over 4906.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2329, pruned_loss=0.0433, over 958392.12 frames. ], batch size: 46, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:33:28,875 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=170461.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:52,313 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.526e+02 1.784e+02 2.388e+02 5.022e+02, threshold=3.568e+02, percent-clipped=3.0 +2023-04-28 06:34:04,609 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170490.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:34:23,439 INFO [finetune.py:976] (3/7) Epoch 30, batch 4400, loss[loss=0.1859, simple_loss=0.2548, pruned_loss=0.05853, over 4916.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2343, pruned_loss=0.04417, over 957587.12 frames. ], batch size: 38, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:34:47,085 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=170522.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 06:35:05,434 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.4238, 1.6693, 1.3696, 1.2212, 1.1590, 1.1445, 1.3760, 1.0990], + device='cuda:3'), covar=tensor([0.1460, 0.1227, 0.1408, 0.1638, 0.2105, 0.1871, 0.0910, 0.1909], + device='cuda:3'), in_proj_covar=tensor([0.0199, 0.0209, 0.0170, 0.0204, 0.0201, 0.0187, 0.0157, 0.0187], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 06:35:08,310 INFO [finetune.py:976] (3/7) Epoch 30, batch 4450, loss[loss=0.1542, simple_loss=0.2406, pruned_loss=0.03384, over 4812.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2394, pruned_loss=0.04568, over 955575.87 frames. ], batch size: 38, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:35:18,639 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=170569.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:35:26,201 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.572e+02 1.800e+02 2.072e+02 3.635e+02, threshold=3.600e+02, percent-clipped=1.0 +2023-04-28 06:35:38,125 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-28 06:35:41,961 INFO [finetune.py:976] (3/7) Epoch 30, batch 4500, loss[loss=0.1632, simple_loss=0.2411, pruned_loss=0.04263, over 4867.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2408, pruned_loss=0.04594, over 956297.65 frames. ], batch size: 31, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:35:45,761 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0996, 1.5617, 1.9222, 2.3966, 1.9266, 1.5816, 1.2778, 1.7778], + device='cuda:3'), covar=tensor([0.3142, 0.3160, 0.1777, 0.1934, 0.2605, 0.2600, 0.4143, 0.1947], + device='cuda:3'), in_proj_covar=tensor([0.0296, 0.0248, 0.0232, 0.0318, 0.0226, 0.0238, 0.0232, 0.0188], + device='cuda:3'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:3') +2023-04-28 06:35:55,453 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-28 06:35:59,494 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=170630.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:36:15,352 INFO [finetune.py:976] (3/7) Epoch 30, batch 4550, loss[loss=0.1639, simple_loss=0.2434, pruned_loss=0.04223, over 4897.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2408, pruned_loss=0.04572, over 956306.23 frames. ], batch size: 43, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:36:24,377 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0025, 1.5237, 4.0552, 3.8327, 3.5841, 3.8328, 3.6683, 3.6267], + device='cuda:3'), covar=tensor([0.6505, 0.5197, 0.1073, 0.1663, 0.1086, 0.1717, 0.3675, 0.1376], + device='cuda:3'), in_proj_covar=tensor([0.0308, 0.0306, 0.0402, 0.0407, 0.0346, 0.0412, 0.0315, 0.0361], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:36:25,047 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.2906, 2.1924, 1.9014, 1.8893, 2.4010, 1.9235, 2.8784, 1.7658], + device='cuda:3'), covar=tensor([0.3462, 0.2095, 0.4539, 0.3044, 0.1588, 0.2636, 0.1307, 0.4379], + device='cuda:3'), in_proj_covar=tensor([0.0341, 0.0358, 0.0429, 0.0354, 0.0388, 0.0377, 0.0372, 0.0425], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:36:33,303 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.995e+01 1.537e+02 1.792e+02 2.273e+02 3.619e+02, threshold=3.584e+02, percent-clipped=1.0 +2023-04-28 06:36:47,829 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170701.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:36:48,972 INFO [finetune.py:976] (3/7) Epoch 30, batch 4600, loss[loss=0.1884, simple_loss=0.2586, pruned_loss=0.05913, over 4791.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.24, pruned_loss=0.04521, over 954869.64 frames. ], batch size: 26, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:37:10,101 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.3713, 1.9191, 1.6655, 2.3048, 2.5114, 2.0542, 1.9582, 1.7394], + device='cuda:3'), covar=tensor([0.1692, 0.1557, 0.1749, 0.1427, 0.1057, 0.1729, 0.2196, 0.2104], + device='cuda:3'), in_proj_covar=tensor([0.0315, 0.0307, 0.0350, 0.0286, 0.0323, 0.0305, 0.0300, 0.0377], + device='cuda:3'), out_proj_covar=tensor([6.3933e-05, 6.2649e-05, 7.3093e-05, 5.7044e-05, 6.5776e-05, 6.3235e-05, + 6.1846e-05, 7.9499e-05], device='cuda:3') +2023-04-28 06:37:19,720 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170749.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:37:22,626 INFO [finetune.py:976] (3/7) Epoch 30, batch 4650, loss[loss=0.1244, simple_loss=0.1901, pruned_loss=0.02931, over 4490.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2377, pruned_loss=0.04459, over 954563.90 frames. ], batch size: 19, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:37:39,988 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.849e+01 1.520e+02 1.798e+02 2.217e+02 4.405e+02, threshold=3.597e+02, percent-clipped=2.0 +2023-04-28 06:37:55,525 INFO [finetune.py:976] (3/7) Epoch 30, batch 4700, loss[loss=0.1249, simple_loss=0.2017, pruned_loss=0.02399, over 4938.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2336, pruned_loss=0.04373, over 952773.19 frames. ], batch size: 33, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:38:05,046 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170817.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 06:38:17,404 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=3.91 vs. limit=5.0 +2023-04-28 06:38:29,291 INFO [finetune.py:976] (3/7) Epoch 30, batch 4750, loss[loss=0.1444, simple_loss=0.2187, pruned_loss=0.03503, over 4845.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2313, pruned_loss=0.04337, over 952035.86 frames. ], batch size: 47, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:38:47,639 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.520e+02 1.829e+02 2.099e+02 3.984e+02, threshold=3.658e+02, percent-clipped=1.0 +2023-04-28 06:38:59,651 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.0910, 2.5538, 1.0747, 1.4331, 1.9619, 1.2217, 3.4078, 1.7604], + device='cuda:3'), covar=tensor([0.0683, 0.0688, 0.0805, 0.1198, 0.0486, 0.1003, 0.0193, 0.0612], + device='cuda:3'), in_proj_covar=tensor([0.0051, 0.0064, 0.0046, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:3'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:3') +2023-04-28 06:39:16,798 INFO [finetune.py:976] (3/7) Epoch 30, batch 4800, loss[loss=0.1979, simple_loss=0.2634, pruned_loss=0.0662, over 4206.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.235, pruned_loss=0.04512, over 951769.09 frames. ], batch size: 65, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:39:45,897 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170925.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:39:47,776 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=170928.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:40:21,162 INFO [finetune.py:976] (3/7) Epoch 30, batch 4850, loss[loss=0.173, simple_loss=0.257, pruned_loss=0.04449, over 4770.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2372, pruned_loss=0.04507, over 951575.57 frames. ], batch size: 54, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:40:52,179 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 8.978e+01 1.534e+02 1.869e+02 2.233e+02 4.849e+02, threshold=3.739e+02, percent-clipped=1.0 +2023-04-28 06:40:59,543 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=170989.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:41:19,849 INFO [finetune.py:976] (3/7) Epoch 30, batch 4900, loss[loss=0.1345, simple_loss=0.2018, pruned_loss=0.03364, over 4470.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2384, pruned_loss=0.04514, over 950663.06 frames. ], batch size: 19, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:42:04,420 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6243, 1.4606, 1.6448, 1.9165, 2.0569, 1.5843, 1.3013, 1.8025], + device='cuda:3'), covar=tensor([0.0853, 0.1269, 0.0871, 0.0603, 0.0590, 0.0873, 0.0786, 0.0564], + device='cuda:3'), in_proj_covar=tensor([0.0181, 0.0201, 0.0182, 0.0171, 0.0176, 0.0177, 0.0149, 0.0175], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:42:25,534 INFO [finetune.py:976] (3/7) Epoch 30, batch 4950, loss[loss=0.173, simple_loss=0.25, pruned_loss=0.04802, over 4812.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2403, pruned_loss=0.0453, over 953449.36 frames. ], batch size: 45, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:42:35,108 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6766, 1.6674, 0.9831, 1.3827, 1.6992, 1.5460, 1.4088, 1.5114], + device='cuda:3'), covar=tensor([0.0448, 0.0323, 0.0302, 0.0500, 0.0272, 0.0444, 0.0448, 0.0499], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0027], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 06:42:43,495 INFO [scaling.py:679] (3/7) Whitening: num_groups=1, num_channels=384, metric=5.47 vs. limit=5.0 +2023-04-28 06:42:45,070 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.630e+02 1.843e+02 2.166e+02 3.655e+02, threshold=3.685e+02, percent-clipped=0.0 +2023-04-28 06:42:52,363 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171092.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:42:59,426 INFO [finetune.py:976] (3/7) Epoch 30, batch 5000, loss[loss=0.1267, simple_loss=0.2014, pruned_loss=0.02597, over 4697.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2396, pruned_loss=0.04503, over 952823.20 frames. ], batch size: 23, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:43:09,446 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171117.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 06:43:20,770 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.0547, 2.0353, 1.8057, 1.6936, 2.1541, 1.7637, 2.6613, 1.6032], + device='cuda:3'), covar=tensor([0.3582, 0.2097, 0.4374, 0.3114, 0.1585, 0.2491, 0.1140, 0.4458], + device='cuda:3'), in_proj_covar=tensor([0.0340, 0.0356, 0.0426, 0.0351, 0.0385, 0.0374, 0.0371, 0.0423], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:43:33,015 INFO [finetune.py:976] (3/7) Epoch 30, batch 5050, loss[loss=0.1375, simple_loss=0.2142, pruned_loss=0.03042, over 4879.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2376, pruned_loss=0.04473, over 954340.83 frames. ], batch size: 31, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:43:33,143 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171153.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:43:41,329 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=171165.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:43:52,462 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.935e+01 1.498e+02 1.792e+02 2.110e+02 3.798e+02, threshold=3.585e+02, percent-clipped=1.0 +2023-04-28 06:44:02,210 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171196.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:44:06,507 INFO [finetune.py:976] (3/7) Epoch 30, batch 5100, loss[loss=0.1473, simple_loss=0.2127, pruned_loss=0.04099, over 4853.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.235, pruned_loss=0.04422, over 954973.80 frames. ], batch size: 47, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:44:21,421 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171225.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:44:24,167 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.3971, 1.3936, 1.6474, 1.7044, 1.3363, 1.0410, 1.3134, 0.7079], + device='cuda:3'), covar=tensor([0.0613, 0.0477, 0.0384, 0.0543, 0.0712, 0.1306, 0.0597, 0.0662], + device='cuda:3'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0076, 0.0095, 0.0072, 0.0062], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 06:44:40,212 INFO [finetune.py:976] (3/7) Epoch 30, batch 5150, loss[loss=0.1856, simple_loss=0.2531, pruned_loss=0.05908, over 4917.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2345, pruned_loss=0.04442, over 954580.39 frames. ], batch size: 36, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:44:42,809 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171257.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:44:53,820 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=171273.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:45:05,610 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.555e+02 1.961e+02 2.367e+02 3.631e+02, threshold=3.922e+02, percent-clipped=1.0 +2023-04-28 06:45:07,335 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=171284.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:45:35,546 INFO [finetune.py:976] (3/7) Epoch 30, batch 5200, loss[loss=0.1546, simple_loss=0.218, pruned_loss=0.04564, over 3329.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2378, pruned_loss=0.0451, over 951721.55 frames. ], batch size: 14, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:46:31,230 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171345.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:46:36,592 INFO [finetune.py:976] (3/7) Epoch 30, batch 5250, loss[loss=0.1733, simple_loss=0.2382, pruned_loss=0.0542, over 3982.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2384, pruned_loss=0.04477, over 951293.31 frames. ], batch size: 17, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:46:59,748 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([0.8233, 1.1681, 3.2799, 3.0222, 2.9122, 3.2050, 3.2039, 2.8843], + device='cuda:3'), covar=tensor([0.7861, 0.5811, 0.1597, 0.2496, 0.1598, 0.2115, 0.1758, 0.1906], + device='cuda:3'), in_proj_covar=tensor([0.0314, 0.0310, 0.0409, 0.0414, 0.0351, 0.0420, 0.0319, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:47:08,718 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.5351, 1.2143, 4.3659, 4.0600, 3.7805, 4.2061, 4.0260, 3.8837], + device='cuda:3'), covar=tensor([0.7427, 0.6061, 0.1005, 0.1755, 0.1190, 0.1307, 0.1972, 0.1477], + device='cuda:3'), in_proj_covar=tensor([0.0313, 0.0310, 0.0409, 0.0413, 0.0350, 0.0419, 0.0319, 0.0367], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:47:18,735 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.554e+02 1.778e+02 2.227e+02 5.005e+02, threshold=3.556e+02, percent-clipped=1.0 +2023-04-28 06:47:43,263 INFO [finetune.py:976] (3/7) Epoch 30, batch 5300, loss[loss=0.1107, simple_loss=0.189, pruned_loss=0.0162, over 4800.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2398, pruned_loss=0.04458, over 949546.04 frames. ], batch size: 25, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:47:50,521 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171406.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:48:41,619 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=171448.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:48:44,664 INFO [finetune.py:976] (3/7) Epoch 30, batch 5350, loss[loss=0.1809, simple_loss=0.2537, pruned_loss=0.05401, over 4895.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2409, pruned_loss=0.0451, over 950084.52 frames. ], batch size: 35, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:48:54,917 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([2.6226, 2.4806, 2.5401, 3.1673, 3.1080, 2.4188, 2.2815, 2.6750], + device='cuda:3'), covar=tensor([0.0814, 0.0952, 0.0616, 0.0471, 0.0465, 0.0903, 0.0655, 0.0503], + device='cuda:3'), in_proj_covar=tensor([0.0182, 0.0202, 0.0183, 0.0171, 0.0177, 0.0178, 0.0149, 0.0176], + device='cuda:3'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:3') +2023-04-28 06:49:03,279 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.512e+02 1.839e+02 2.201e+02 3.854e+02, threshold=3.679e+02, percent-clipped=1.0 +2023-04-28 06:49:15,018 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.6842, 1.4614, 0.6758, 1.3623, 1.4796, 1.5393, 1.4329, 1.4421], + device='cuda:3'), covar=tensor([0.0484, 0.0381, 0.0350, 0.0540, 0.0275, 0.0475, 0.0467, 0.0562], + device='cuda:3'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:3'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0054, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:3') +2023-04-28 06:49:15,618 INFO [zipformer.py:1188] (3/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171498.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:49:18,573 INFO [finetune.py:976] (3/7) Epoch 30, batch 5400, loss[loss=0.1911, simple_loss=0.2495, pruned_loss=0.06631, over 4869.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2388, pruned_loss=0.04473, over 949264.94 frames. ], batch size: 31, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:49:51,781 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=171552.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:49:52,349 INFO [finetune.py:976] (3/7) Epoch 30, batch 5450, loss[loss=0.1774, simple_loss=0.2412, pruned_loss=0.0568, over 4916.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2364, pruned_loss=0.04448, over 951406.69 frames. ], batch size: 37, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:49:56,127 INFO [zipformer.py:1188] (3/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171559.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:50:10,994 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 9.901e+01 1.444e+02 1.753e+02 2.019e+02 4.846e+02, threshold=3.505e+02, percent-clipped=2.0 +2023-04-28 06:50:11,414 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-28 06:50:12,312 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171584.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:50:14,380 INFO [scaling.py:679] (3/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-28 06:50:25,849 INFO [finetune.py:976] (3/7) Epoch 30, batch 5500, loss[loss=0.1409, simple_loss=0.2067, pruned_loss=0.0375, over 4797.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2353, pruned_loss=0.04451, over 951343.55 frames. ], batch size: 29, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:50:31,401 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8265, 2.0801, 1.7597, 1.5526, 1.3462, 1.3749, 1.7235, 1.2506], + device='cuda:3'), covar=tensor([0.1736, 0.1330, 0.1445, 0.1744, 0.2385, 0.1965, 0.1078, 0.2168], + device='cuda:3'), in_proj_covar=tensor([0.0200, 0.0210, 0.0171, 0.0205, 0.0202, 0.0188, 0.0157, 0.0189], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:3') +2023-04-28 06:50:33,195 INFO [zipformer.py:2441] (3/7) attn_weights_entropy = tensor([1.8560, 4.1015, 0.8186, 2.2221, 2.3962, 2.6461, 2.4186, 0.9590], + device='cuda:3'), covar=tensor([0.1246, 0.0762, 0.2012, 0.1097, 0.0904, 0.1011, 0.1297, 0.1956], + device='cuda:3'), in_proj_covar=tensor([0.0116, 0.0235, 0.0135, 0.0120, 0.0130, 0.0151, 0.0116, 0.0116], + device='cuda:3'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:3') +2023-04-28 06:50:45,015 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=171632.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:50:59,649 INFO [finetune.py:976] (3/7) Epoch 30, batch 5550, loss[loss=0.1817, simple_loss=0.2482, pruned_loss=0.0576, over 4827.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2366, pruned_loss=0.04481, over 951712.94 frames. ], batch size: 33, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:51:38,770 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.632e+02 1.883e+02 2.258e+02 4.653e+02, threshold=3.766e+02, percent-clipped=3.0 +2023-04-28 06:52:00,709 INFO [zipformer.py:1188] (3/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=171701.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:52:01,826 INFO [finetune.py:976] (3/7) Epoch 30, batch 5600, loss[loss=0.1626, simple_loss=0.2318, pruned_loss=0.04668, over 4829.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2398, pruned_loss=0.04529, over 952078.88 frames. ], batch size: 25, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:52:55,570 INFO [zipformer.py:1188] (3/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171748.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:52:58,357 INFO [finetune.py:976] (3/7) Epoch 30, batch 5650, loss[loss=0.2143, simple_loss=0.2896, pruned_loss=0.06951, over 4860.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2428, pruned_loss=0.04593, over 953025.40 frames. ], batch size: 44, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:53:37,490 INFO [optim.py:369] (3/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.430e+02 1.735e+02 1.948e+02 3.086e+02, threshold=3.470e+02, percent-clipped=0.0 +2023-04-28 06:53:50,971 INFO [zipformer.py:1188] (3/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=171796.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:54:00,720 INFO [finetune.py:976] (3/7) Epoch 30, batch 5700, loss[loss=0.1673, simple_loss=0.2274, pruned_loss=0.0536, over 4275.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2386, pruned_loss=0.04572, over 935621.76 frames. ], batch size: 18, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:54:33,270 INFO [finetune.py:1241] (3/7) Done!