diff --git "a/exp/log/log-train-2023-04-26-10-07-06-2" "b/exp/log/log-train-2023-04-26-10-07-06-2" new file mode 100644--- /dev/null +++ "b/exp/log/log-train-2023-04-26-10-07-06-2" @@ -0,0 +1,23976 @@ +2023-04-26 10:07:06,651 INFO [finetune.py:1046] (2/7) Training started +2023-04-26 10:07:06,651 INFO [finetune.py:1056] (2/7) Device: cuda:2 +2023-04-26 10:07:06,654 INFO [finetune.py:1065] (2/7) {'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 10:07:06,654 INFO [finetune.py:1067] (2/7) About to create model +2023-04-26 10:07:07,026 INFO [zipformer.py:405] (2/7) 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 10:07:07,035 INFO [finetune.py:1071] (2/7) Number of model parameters: 70369391 +2023-04-26 10:07:07,035 INFO [finetune.py:626] (2/7) Loading checkpoint from /home/lishaojie/icefall/egs/commonvoice_fr/ASR/pruned_transducer_stateless7_streaming/exp/english_pretrain/pretrained.pt +2023-04-26 10:07:07,188 INFO [finetune.py:647] (2/7) Loading parameters starting with prefix encoder +2023-04-26 10:07:08,682 INFO [finetune.py:1093] (2/7) Using DDP +2023-04-26 10:07:09,245 INFO [commonvoice_fr.py:392] (2/7) About to get train cuts +2023-04-26 10:07:09,248 INFO [commonvoice_fr.py:218] (2/7) Enable MUSAN +2023-04-26 10:07:09,248 INFO [commonvoice_fr.py:219] (2/7) About to get Musan cuts +2023-04-26 10:07:10,764 INFO [commonvoice_fr.py:243] (2/7) Enable SpecAugment +2023-04-26 10:07:10,765 INFO [commonvoice_fr.py:244] (2/7) Time warp factor: 80 +2023-04-26 10:07:10,765 INFO [commonvoice_fr.py:254] (2/7) Num frame mask: 10 +2023-04-26 10:07:10,765 INFO [commonvoice_fr.py:267] (2/7) About to create train dataset +2023-04-26 10:07:10,765 INFO [commonvoice_fr.py:294] (2/7) Using DynamicBucketingSampler. +2023-04-26 10:07:13,497 INFO [commonvoice_fr.py:309] (2/7) About to create train dataloader +2023-04-26 10:07:13,498 INFO [commonvoice_fr.py:399] (2/7) About to get dev cuts +2023-04-26 10:07:13,499 INFO [commonvoice_fr.py:340] (2/7) About to create dev dataset +2023-04-26 10:07:13,901 INFO [commonvoice_fr.py:357] (2/7) About to create dev dataloader +2023-04-26 10:07:13,901 INFO [finetune.py:1289] (2/7) Sanity check -- see if any of the batches in epoch 1 would cause OOM. +2023-04-26 10:11:06,928 INFO [finetune.py:1317] (2/7) Maximum memory allocated so far is 4858MB +2023-04-26 10:11:07,627 INFO [finetune.py:1317] (2/7) Maximum memory allocated so far is 5345MB +2023-04-26 10:11:08,315 INFO [finetune.py:1317] (2/7) Maximum memory allocated so far is 5345MB +2023-04-26 10:11:08,989 INFO [finetune.py:1317] (2/7) Maximum memory allocated so far is 5345MB +2023-04-26 10:11:09,670 INFO [finetune.py:1317] (2/7) Maximum memory allocated so far is 5345MB +2023-04-26 10:11:10,372 INFO [finetune.py:1317] (2/7) Maximum memory allocated so far is 5345MB +2023-04-26 10:11:19,552 INFO [finetune.py:976] (2/7) Epoch 1, batch 0, loss[loss=7.427, simple_loss=6.731, pruned_loss=6.941, over 4760.00 frames. ], tot_loss[loss=7.427, simple_loss=6.731, pruned_loss=6.941, over 4760.00 frames. ], batch size: 28, lr: 2.00e-03, grad_scale: 2.0 +2023-04-26 10:11:19,552 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 10:11:40,202 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 5345MB +2023-04-26 10:11:48,674 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:12:02,477 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.34 vs. limit=2.0 +2023-04-26 10:12:02,988 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 10:12:11,186 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:12:35,019 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.86 vs. limit=2.0 +2023-04-26 10:12:43,282 INFO [finetune.py:976] (2/7) Epoch 1, batch 50, loss[loss=2.569, simple_loss=2.448, pruned_loss=1.238, over 4248.00 frames. ], tot_loss[loss=4.463, simple_loss=4.04, pruned_loss=4.089, over 216829.18 frames. ], batch size: 66, lr: 2.20e-03, grad_scale: 0.00390625 +2023-04-26 10:13:08,807 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=18.28 vs. limit=5.0 +2023-04-26 10:13:19,822 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:13:41,118 WARNING [finetune.py:966] (2/7) Grad scale is small: 6.103515625e-05 +2023-04-26 10:13:41,118 INFO [finetune.py:976] (2/7) Epoch 1, batch 100, loss[loss=2.137, simple_loss=2.032, pruned_loss=1.074, over 4743.00 frames. ], tot_loss[loss=3.584, simple_loss=3.314, pruned_loss=2.64, over 381366.70 frames. ], batch size: 59, lr: 2.40e-03, grad_scale: 0.0001220703125 +2023-04-26 10:14:03,493 INFO [optim.py:369] (2/7) 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] (2/7) Scaling gradients by 0.014711554162204266, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:06,254 INFO [optim.py:451] (2/7) 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] (2/7) Scaling gradients by 0.00018281130178365856, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:23,480 INFO [optim.py:451] (2/7) 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 10:14:24,119 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:14:27,642 INFO [finetune.py:976] (2/7) Epoch 1, batch 150, loss[loss=1.715, simple_loss=1.554, pruned_loss=1.308, over 4930.00 frames. ], tot_loss[loss=2.979, simple_loss=2.762, pruned_loss=2.085, over 507461.57 frames. ], batch size: 33, lr: 2.60e-03, grad_scale: 3.0517578125e-05 +2023-04-26 10:14:28,149 WARNING [optim.py:389] (2/7) Scaling gradients by 0.00022292081848718226, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:28,222 INFO [optim.py:451] (2/7) 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 10:14:40,638 WARNING [optim.py:389] (2/7) Scaling gradients by 0.05655747279524803, model_norm_threshold=15666.9306640625 +2023-04-26 10:14:40,712 INFO [optim.py:451] (2/7) 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 10:14:45,005 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=60.16 vs. limit=5.0 +2023-04-26 10:14:56,526 WARNING [finetune.py:966] (2/7) Grad scale is small: 3.0517578125e-05 +2023-04-26 10:14:56,526 INFO [finetune.py:976] (2/7) Epoch 1, batch 200, loss[loss=1.246, simple_loss=1.073, pruned_loss=1.193, over 4787.00 frames. ], tot_loss[loss=2.443, simple_loss=2.243, pruned_loss=1.791, over 607944.63 frames. ], batch size: 29, lr: 2.80e-03, grad_scale: 6.103515625e-05 +2023-04-26 10:15:07,791 INFO [optim.py:369] (2/7) 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 10:15:12,598 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=15.02 vs. limit=5.0 +2023-04-26 10:15:12,965 WARNING [optim.py:389] (2/7) Scaling gradients by 0.011872046627104282, model_norm_threshold=3679.54541015625 +2023-04-26 10:15:13,040 INFO [optim.py:451] (2/7) 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 10:15:16,161 WARNING [optim.py:389] (2/7) Scaling gradients by 0.08515117317438126, model_norm_threshold=3679.54541015625 +2023-04-26 10:15:16,238 INFO [optim.py:451] (2/7) 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] (2/7) Scaling gradients by 0.04552413150668144, model_norm_threshold=3679.54541015625 +2023-04-26 10:15:16,848 INFO [optim.py:451] (2/7) 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 10:15:25,656 INFO [finetune.py:976] (2/7) Epoch 1, batch 250, loss[loss=1.379, simple_loss=1.169, pruned_loss=1.33, over 4856.00 frames. ], tot_loss[loss=2.1, simple_loss=1.903, pruned_loss=1.63, over 685335.08 frames. ], batch size: 31, lr: 3.00e-03, grad_scale: 6.103515625e-05 +2023-04-26 10:15:50,804 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=296.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:15:52,838 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=300.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:15:53,285 WARNING [finetune.py:966] (2/7) Grad scale is small: 6.103515625e-05 +2023-04-26 10:15:53,285 INFO [finetune.py:976] (2/7) Epoch 1, batch 300, loss[loss=1.188, simple_loss=0.9893, pruned_loss=1.159, over 4810.00 frames. ], tot_loss[loss=1.883, simple_loss=1.683, pruned_loss=1.531, over 742192.50 frames. ], batch size: 25, lr: 3.20e-03, grad_scale: 0.0001220703125 +2023-04-26 10:16:02,530 INFO [optim.py:369] (2/7) 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 10:16:21,504 INFO [finetune.py:976] (2/7) Epoch 1, batch 350, loss[loss=1.261, simple_loss=1.052, pruned_loss=1.17, over 4144.00 frames. ], tot_loss[loss=1.732, simple_loss=1.525, pruned_loss=1.46, over 789175.10 frames. ], batch size: 65, lr: 3.40e-03, grad_scale: 0.0001220703125 +2023-04-26 10:16:24,663 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=357.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:16:42,456 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=387.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:16:56,315 WARNING [finetune.py:966] (2/7) Grad scale is small: 0.0001220703125 +2023-04-26 10:16:56,315 INFO [finetune.py:976] (2/7) Epoch 1, batch 400, loss[loss=1.429, simple_loss=1.149, pruned_loss=1.409, over 4844.00 frames. ], tot_loss[loss=1.615, simple_loss=1.4, pruned_loss=1.405, over 825033.18 frames. ], batch size: 44, lr: 3.60e-03, grad_scale: 0.000244140625 +2023-04-26 10:17:09,478 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3097, 0.7005, 1.1170, 1.4763, 0.9297, 1.4805, 1.6975, 1.5760], + device='cuda:2'), covar=tensor([0.1659, 0.1053, 0.0970, 0.1228, 0.1519, 0.1076, 0.0889, 0.0596], + device='cuda:2'), in_proj_covar=tensor([0.0524, 0.0596, 0.0674, 0.0650, 0.0578, 0.0635, 0.0668, 0.0657], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 10:17:16,619 INFO [optim.py:369] (2/7) 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 10:17:27,880 WARNING [optim.py:389] (2/7) Scaling gradients by 0.02133115753531456, model_norm_threshold=142.37583923339844 +2023-04-26 10:17:27,960 INFO [optim.py:451] (2/7) 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 10:17:41,639 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=439.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:17:52,438 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=448.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:17:53,954 INFO [finetune.py:976] (2/7) Epoch 1, batch 450, loss[loss=1.127, simple_loss=0.8801, pruned_loss=1.144, over 4822.00 frames. ], tot_loss[loss=1.51, simple_loss=1.287, pruned_loss=1.352, over 854618.90 frames. ], batch size: 30, lr: 3.80e-03, grad_scale: 0.000244140625 +2023-04-26 10:18:14,586 WARNING [optim.py:389] (2/7) Scaling gradients by 0.06225070729851723, model_norm_threshold=142.37583923339844 +2023-04-26 10:18:14,660 INFO [optim.py:451] (2/7) 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 10:18:22,847 WARNING [finetune.py:966] (2/7) Grad scale is small: 0.000244140625 +2023-04-26 10:18:22,847 INFO [finetune.py:976] (2/7) Epoch 1, batch 500, loss[loss=1.119, simple_loss=0.8482, pruned_loss=1.159, over 4866.00 frames. ], tot_loss[loss=1.418, simple_loss=1.187, pruned_loss=1.303, over 876767.44 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 0.00048828125 +2023-04-26 10:18:31,734 INFO [optim.py:369] (2/7) 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 10:18:32,903 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=49.59 vs. limit=5.0 +2023-04-26 10:18:47,426 WARNING [optim.py:389] (2/7) Scaling gradients by 0.017591100186109543, model_norm_threshold=62.30100631713867 +2023-04-26 10:18:47,501 INFO [optim.py:451] (2/7) 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 10:18:56,301 WARNING [optim.py:389] (2/7) Scaling gradients by 0.005508477333933115, model_norm_threshold=62.30100631713867 +2023-04-26 10:18:56,374 INFO [optim.py:451] (2/7) 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 10:18:58,474 INFO [finetune.py:976] (2/7) Epoch 1, batch 550, loss[loss=1.114, simple_loss=0.8374, pruned_loss=1.128, over 4833.00 frames. ], tot_loss[loss=1.339, simple_loss=1.1, pruned_loss=1.256, over 895454.83 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 0.00048828125 +2023-04-26 10:19:04,274 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=562.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:19:13,089 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=568.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:19:13,161 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.21 vs. limit=2.0 +2023-04-26 10:19:35,136 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=590.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:19:45,886 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=600.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:19:46,332 WARNING [finetune.py:966] (2/7) Grad scale is small: 0.00048828125 +2023-04-26 10:19:46,332 INFO [finetune.py:976] (2/7) Epoch 1, batch 600, loss[loss=1.219, simple_loss=0.8902, pruned_loss=1.249, over 4807.00 frames. ], tot_loss[loss=1.284, simple_loss=1.033, pruned_loss=1.224, over 910771.18 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 0.0009765625 +2023-04-26 10:20:06,710 INFO [optim.py:369] (2/7) 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 10:20:09,422 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=623.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:20:18,150 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=629.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:20:29,918 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=648.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:20:31,448 INFO [finetune.py:976] (2/7) Epoch 1, batch 650, loss[loss=1.177, simple_loss=0.8558, pruned_loss=1.173, over 4834.00 frames. ], tot_loss[loss=1.249, simple_loss=0.9861, pruned_loss=1.204, over 921919.70 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 0.0009765625 +2023-04-26 10:20:31,536 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=652.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:20:41,809 WARNING [optim.py:389] (2/7) Scaling gradients by 0.07653743773698807, model_norm_threshold=52.5806770324707 +2023-04-26 10:20:41,883 INFO [optim.py:451] (2/7) 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 10:20:59,956 WARNING [finetune.py:966] (2/7) Grad scale is small: 0.0009765625 +2023-04-26 10:20:59,957 INFO [finetune.py:976] (2/7) Epoch 1, batch 700, loss[loss=1.082, simple_loss=0.7724, pruned_loss=1.072, over 4893.00 frames. ], tot_loss[loss=1.215, simple_loss=0.943, pruned_loss=1.176, over 927362.42 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.001953125 +2023-04-26 10:21:09,245 INFO [optim.py:369] (2/7) 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 10:21:20,032 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=739.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:21:22,088 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=743.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:21:28,230 INFO [finetune.py:976] (2/7) Epoch 1, batch 750, loss[loss=1.127, simple_loss=0.7952, pruned_loss=1.099, over 4881.00 frames. ], tot_loss[loss=1.187, simple_loss=0.9067, pruned_loss=1.149, over 934005.17 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 0.001953125 +2023-04-26 10:21:48,061 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=787.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:21:51,310 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.36 vs. limit=2.0 +2023-04-26 10:21:52,803 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3746, 2.1007, 1.6885, 1.6094, 1.3449, 1.2203, 1.4269, 1.5631], + device='cuda:2'), covar=tensor([0.0211, 0.0228, 0.0157, 0.0270, 0.0259, 0.0200, 0.0156, 0.0189], + device='cuda:2'), in_proj_covar=tensor([0.0224, 0.0255, 0.0225, 0.0248, 0.0263, 0.0220, 0.0213, 0.0237], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003, 0.0004], + device='cuda:2') +2023-04-26 10:21:53,226 WARNING [optim.py:389] (2/7) Scaling gradients by 0.039711207151412964, model_norm_threshold=49.79251480102539 +2023-04-26 10:21:53,300 INFO [optim.py:451] (2/7) 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 10:21:55,913 WARNING [finetune.py:966] (2/7) Grad scale is small: 0.001953125 +2023-04-26 10:21:55,913 INFO [finetune.py:976] (2/7) Epoch 1, batch 800, loss[loss=0.9592, simple_loss=0.6775, pruned_loss=0.9064, over 4759.00 frames. ], tot_loss[loss=1.164, simple_loss=0.8766, pruned_loss=1.122, over 940825.78 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 0.00390625 +2023-04-26 10:22:11,120 INFO [optim.py:369] (2/7) 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 10:22:11,733 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1788, 0.9454, 3.9744, 3.6101, 3.9928, 3.9450, 3.6498, 3.7404], + device='cuda:2'), covar=tensor([0.4547, 0.4871, 0.2153, 0.3321, 0.1659, 0.1822, 0.2467, 0.2440], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0314, 0.0454, 0.0463, 0.0371, 0.0432, 0.0352, 0.0401], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 10:22:26,849 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=847.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:22:28,870 INFO [finetune.py:976] (2/7) Epoch 1, batch 850, loss[loss=0.9456, simple_loss=0.6786, pruned_loss=0.8518, over 4933.00 frames. ], tot_loss[loss=1.125, simple_loss=0.8387, pruned_loss=1.075, over 940818.62 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 0.00390625 +2023-04-26 10:22:40,825 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=11.80 vs. limit=5.0 +2023-04-26 10:23:15,920 WARNING [finetune.py:966] (2/7) Grad scale is small: 0.00390625 +2023-04-26 10:23:15,920 INFO [finetune.py:976] (2/7) Epoch 1, batch 900, loss[loss=0.9811, simple_loss=0.6926, pruned_loss=0.8795, over 4889.00 frames. ], tot_loss[loss=1.093, simple_loss=0.8076, pruned_loss=1.031, over 945655.79 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 0.0078125 +2023-04-26 10:23:20,206 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=908.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:23:26,898 INFO [optim.py:369] (2/7) 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 10:23:26,978 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=918.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:23:30,104 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=924.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:23:42,108 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=946.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:23:44,689 INFO [finetune.py:976] (2/7) Epoch 1, batch 950, loss[loss=0.9978, simple_loss=0.6987, pruned_loss=0.8812, over 4900.00 frames. ], tot_loss[loss=1.069, simple_loss=0.7838, pruned_loss=0.9933, over 947572.31 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 0.0078125 +2023-04-26 10:23:45,272 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=952.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:23:49,536 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=15.50 vs. limit=5.0 +2023-04-26 10:24:41,986 INFO [zipformer.py:1188] (2/7) 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] (2/7) Grad scale is small: 0.0078125 +2023-04-26 10:24:42,451 INFO [finetune.py:976] (2/7) Epoch 1, batch 1000, loss[loss=1.008, simple_loss=0.72, pruned_loss=0.851, over 4891.00 frames. ], tot_loss[loss=1.064, simple_loss=0.7751, pruned_loss=0.9717, over 950462.30 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.015625 +2023-04-26 10:25:00,169 INFO [optim.py:369] (2/7) 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 10:25:14,568 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1043.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:25:16,671 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1079, 0.8794, 0.7881, 1.2980, 1.6794, 1.0161, 0.7163, 1.3494], + device='cuda:2'), covar=tensor([0.0298, 0.0339, 0.0278, 0.0237, 0.0233, 0.0254, 0.0271, 0.0199], + device='cuda:2'), in_proj_covar=tensor([0.0205, 0.0212, 0.0191, 0.0172, 0.0174, 0.0191, 0.0167, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 10:25:18,745 INFO [finetune.py:976] (2/7) Epoch 1, batch 1050, loss[loss=0.9007, simple_loss=0.6291, pruned_loss=0.7626, over 4678.00 frames. ], tot_loss[loss=1.06, simple_loss=0.7677, pruned_loss=0.9509, over 951046.40 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 0.015625 +2023-04-26 10:25:29,951 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-26 10:25:42,483 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1091.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:25:48,290 INFO [finetune.py:976] (2/7) Epoch 1, batch 1100, loss[loss=1.055, simple_loss=0.7415, pruned_loss=0.8697, over 4904.00 frames. ], tot_loss[loss=1.053, simple_loss=0.7588, pruned_loss=0.9282, over 952911.31 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 0.03125 +2023-04-26 10:25:49,007 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=7.99 vs. limit=5.0 +2023-04-26 10:25:57,278 INFO [optim.py:369] (2/7) 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 10:26:08,845 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6769, 2.0555, 1.4025, 1.4012, 1.1248, 1.5178, 1.7328, 1.1948], + device='cuda:2'), covar=tensor([0.0959, 0.0849, 0.0777, 0.1617, 0.1508, 0.1199, 0.0609, 0.1040], + device='cuda:2'), in_proj_covar=tensor([0.0224, 0.0255, 0.0225, 0.0248, 0.0263, 0.0220, 0.0213, 0.0237], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003, 0.0004], + device='cuda:2') +2023-04-26 10:26:17,718 INFO [finetune.py:976] (2/7) Epoch 1, batch 1150, loss[loss=1.006, simple_loss=0.7041, pruned_loss=0.8161, over 4924.00 frames. ], tot_loss[loss=1.048, simple_loss=0.752, pruned_loss=0.9084, over 954655.97 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 0.03125 +2023-04-26 10:26:18,901 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1153.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:26:46,980 INFO [finetune.py:976] (2/7) Epoch 1, batch 1200, loss[loss=0.963, simple_loss=0.6781, pruned_loss=0.7627, over 4887.00 frames. ], tot_loss[loss=1.039, simple_loss=0.7424, pruned_loss=0.8852, over 956220.35 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 0.0625 +2023-04-26 10:26:48,595 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1203.0, num_to_drop=2, layers_to_drop={0, 3} +2023-04-26 10:26:54,327 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1214.0, num_to_drop=2, layers_to_drop={0, 1} +2023-04-26 10:26:56,334 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1218.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:26:59,575 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1224.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:27:19,094 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1246.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:27:20,225 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=9.27 vs. limit=5.0 +2023-04-26 10:27:21,700 INFO [finetune.py:976] (2/7) Epoch 1, batch 1250, loss[loss=1.039, simple_loss=0.723, pruned_loss=0.8178, over 4827.00 frames. ], tot_loss[loss=1.022, simple_loss=0.7273, pruned_loss=0.8574, over 956663.47 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 0.0625 +2023-04-26 10:27:39,833 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=2.84 vs. limit=2.0 +2023-04-26 10:27:40,181 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1266.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:27:49,521 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1272.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:28:15,069 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1062, 1.8055, 2.1833, 2.2279, 1.9175, 1.8912, 2.2495, 2.2151], + device='cuda:2'), covar=tensor([0.0757, 0.2819, 0.2072, 0.1935, 0.2315, 0.2600, 0.1998, 0.1842], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0477, 0.0384, 0.0374, 0.0416, 0.0421, 0.0467, 0.0406], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 10:28:15,538 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1294.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:28:15,668 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-04-26 10:28:24,915 INFO [finetune.py:976] (2/7) Epoch 1, batch 1300, loss[loss=0.9261, simple_loss=0.6368, pruned_loss=0.7242, over 4899.00 frames. ], tot_loss[loss=1.003, simple_loss=0.7105, pruned_loss=0.8297, over 955746.75 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.125 +2023-04-26 10:28:39,897 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=6.60 vs. limit=5.0 +2023-04-26 10:28:40,244 INFO [optim.py:369] (2/7) 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 10:28:59,768 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.19 vs. limit=5.0 +2023-04-26 10:29:00,697 INFO [finetune.py:976] (2/7) Epoch 1, batch 1350, loss[loss=0.9934, simple_loss=0.6863, pruned_loss=0.7613, over 4909.00 frames. ], tot_loss[loss=1, simple_loss=0.7043, pruned_loss=0.8148, over 955824.48 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 0.125 +2023-04-26 10:29:17,796 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=70.95 vs. limit=5.0 +2023-04-26 10:29:52,994 INFO [finetune.py:976] (2/7) Epoch 1, batch 1400, loss[loss=1.131, simple_loss=0.7775, pruned_loss=0.8567, over 4844.00 frames. ], tot_loss[loss=1.016, simple_loss=0.7114, pruned_loss=0.8154, over 955686.97 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 0.25 +2023-04-26 10:30:14,314 INFO [optim.py:369] (2/7) 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 10:30:22,734 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6678, 1.0210, 4.7526, 4.4017, 4.1968, 4.4223, 4.2339, 4.1263], + device='cuda:2'), covar=tensor([1.0188, 0.9990, 0.1202, 0.2382, 0.1438, 0.1731, 0.2265, 0.2122], + device='cuda:2'), in_proj_covar=tensor([0.0351, 0.0320, 0.0458, 0.0468, 0.0375, 0.0437, 0.0356, 0.0406], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 10:30:34,183 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1434.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:30:35,927 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=7.90 vs. limit=5.0 +2023-04-26 10:30:55,950 INFO [finetune.py:976] (2/7) Epoch 1, batch 1450, loss[loss=0.9967, simple_loss=0.6842, pruned_loss=0.7447, over 4924.00 frames. ], tot_loss[loss=1.02, simple_loss=0.7107, pruned_loss=0.807, over 956270.42 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 0.25 +2023-04-26 10:31:38,888 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-04-26 10:31:41,675 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.18 vs. limit=5.0 +2023-04-26 10:31:42,621 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1495.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:31:53,098 INFO [finetune.py:976] (2/7) Epoch 1, batch 1500, loss[loss=1.016, simple_loss=0.7005, pruned_loss=0.7454, over 4903.00 frames. ], tot_loss[loss=1.019, simple_loss=0.7083, pruned_loss=0.7952, over 956212.06 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 0.5 +2023-04-26 10:31:59,749 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1503.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:32:03,095 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1509.0, num_to_drop=2, layers_to_drop={2, 3} +2023-04-26 10:32:13,076 INFO [optim.py:369] (2/7) 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 10:32:31,913 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3669, 3.2118, 2.4890, 3.7416, 3.2401, 3.2137, 1.4080, 3.1637], + device='cuda:2'), covar=tensor([0.2432, 0.1925, 0.4510, 0.3116, 0.2980, 0.2825, 0.7698, 0.3313], + device='cuda:2'), in_proj_covar=tensor([0.0263, 0.0236, 0.0290, 0.0330, 0.0327, 0.0271, 0.0291, 0.0284], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 10:32:48,627 INFO [finetune.py:976] (2/7) Epoch 1, batch 1550, loss[loss=0.9854, simple_loss=0.6834, pruned_loss=0.7106, over 4895.00 frames. ], tot_loss[loss=1.011, simple_loss=0.7026, pruned_loss=0.777, over 955575.55 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 0.5 +2023-04-26 10:32:48,698 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:33:08,374 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1566.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:33:08,451 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.65 vs. limit=5.0 +2023-04-26 10:33:16,983 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.67 vs. limit=2.0 +2023-04-26 10:33:18,600 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-04-26 10:33:50,405 INFO [finetune.py:976] (2/7) Epoch 1, batch 1600, loss[loss=0.8907, simple_loss=0.6262, pruned_loss=0.6279, over 4359.00 frames. ], tot_loss[loss=0.9909, simple_loss=0.6903, pruned_loss=0.7493, over 954659.13 frames. ], batch size: 19, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:34:04,228 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1611.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:34:11,022 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-04-26 10:34:13,524 INFO [optim.py:369] (2/7) 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 10:34:24,070 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1627.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:34:48,569 INFO [finetune.py:976] (2/7) Epoch 1, batch 1650, loss[loss=0.8341, simple_loss=0.6033, pruned_loss=0.5692, over 4904.00 frames. ], tot_loss[loss=0.9624, simple_loss=0.6743, pruned_loss=0.7152, over 953564.44 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:34:59,642 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.29 vs. limit=2.0 +2023-04-26 10:35:08,299 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1672.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:35:12,158 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.5085, 4.2535, 3.0865, 5.0130, 4.3112, 4.4034, 1.7146, 4.2532], + device='cuda:2'), covar=tensor([0.1316, 0.1109, 0.3612, 0.1012, 0.2720, 0.1681, 0.6309, 0.2152], + device='cuda:2'), in_proj_covar=tensor([0.0262, 0.0237, 0.0290, 0.0329, 0.0327, 0.0270, 0.0289, 0.0284], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 10:35:13,316 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8507, 2.2125, 1.2994, 1.7582, 2.5760, 1.7821, 1.8641, 1.8350], + device='cuda:2'), covar=tensor([0.1095, 0.0479, 0.0541, 0.0770, 0.0318, 0.1079, 0.0913, 0.1253], + device='cuda:2'), in_proj_covar=tensor([0.0036, 0.0028, 0.0026, 0.0035, 0.0023, 0.0035, 0.0035, 0.0037], + device='cuda:2'), out_proj_covar=tensor([0.0054, 0.0045, 0.0039, 0.0056, 0.0039, 0.0054, 0.0054, 0.0057], + device='cuda:2') +2023-04-26 10:35:36,513 INFO [finetune.py:976] (2/7) Epoch 1, batch 1700, loss[loss=0.8384, simple_loss=0.6172, pruned_loss=0.5589, over 4826.00 frames. ], tot_loss[loss=0.9285, simple_loss=0.6568, pruned_loss=0.6769, over 953391.87 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:35:59,302 INFO [optim.py:369] (2/7) 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 10:36:24,059 INFO [finetune.py:976] (2/7) Epoch 1, batch 1750, loss[loss=0.7308, simple_loss=0.551, pruned_loss=0.4741, over 4813.00 frames. ], tot_loss[loss=0.8998, simple_loss=0.6445, pruned_loss=0.6425, over 952142.90 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:36:58,582 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1790.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:37:11,175 INFO [finetune.py:976] (2/7) Epoch 1, batch 1800, loss[loss=0.8056, simple_loss=0.6145, pruned_loss=0.5138, over 4176.00 frames. ], tot_loss[loss=0.8761, simple_loss=0.6374, pruned_loss=0.6116, over 953187.66 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:37:20,785 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3527, 3.3896, 1.0901, 1.8313, 1.7301, 2.3911, 1.9175, 0.9863], + device='cuda:2'), covar=tensor([0.1930, 0.0805, 0.2466, 0.1800, 0.1629, 0.1368, 0.2202, 0.2595], + device='cuda:2'), in_proj_covar=tensor([0.0128, 0.0275, 0.0152, 0.0136, 0.0148, 0.0168, 0.0137, 0.0135], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 10:37:21,903 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1809.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:37:28,278 INFO [optim.py:369] (2/7) 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 10:37:40,495 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1838.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:37:47,706 INFO [finetune.py:976] (2/7) Epoch 1, batch 1850, loss[loss=0.5485, simple_loss=0.4519, pruned_loss=0.3277, over 4759.00 frames. ], tot_loss[loss=0.8441, simple_loss=0.6245, pruned_loss=0.5763, over 954214.82 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:37:51,629 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1857.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:37:51,678 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1857.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:37:59,971 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1870.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:38:17,748 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1899.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:38:18,742 INFO [finetune.py:976] (2/7) Epoch 1, batch 1900, loss[loss=0.6606, simple_loss=0.5441, pruned_loss=0.3926, over 4784.00 frames. ], tot_loss[loss=0.8096, simple_loss=0.6092, pruned_loss=0.5409, over 954161.43 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:38:20,022 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-26 10:38:28,761 INFO [optim.py:369] (2/7) 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 10:38:28,886 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1918.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:38:32,529 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1922.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:38:39,000 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1931.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:38:50,101 INFO [finetune.py:976] (2/7) Epoch 1, batch 1950, loss[loss=0.573, simple_loss=0.4838, pruned_loss=0.3325, over 4831.00 frames. ], tot_loss[loss=0.7723, simple_loss=0.5912, pruned_loss=0.5052, over 954724.77 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 1.0 +2023-04-26 10:38:59,756 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1967.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:39:22,528 INFO [finetune.py:976] (2/7) Epoch 1, batch 2000, loss[loss=0.5273, simple_loss=0.4318, pruned_loss=0.3114, over 3025.00 frames. ], tot_loss[loss=0.7323, simple_loss=0.5694, pruned_loss=0.47, over 953154.50 frames. ], batch size: 12, lr: 4.00e-03, grad_scale: 2.0 +2023-04-26 10:39:39,761 INFO [optim.py:369] (2/7) 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 10:40:01,844 INFO [finetune.py:976] (2/7) Epoch 1, batch 2050, loss[loss=0.5573, simple_loss=0.4676, pruned_loss=0.3235, over 4921.00 frames. ], tot_loss[loss=0.6898, simple_loss=0.5465, pruned_loss=0.434, over 953050.78 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 2.0 +2023-04-26 10:40:08,950 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2661, 1.3617, 3.7835, 3.4992, 3.4003, 3.6093, 3.6266, 3.3566], + device='cuda:2'), covar=tensor([0.6798, 0.5692, 0.1099, 0.1731, 0.1262, 0.1911, 0.1540, 0.1410], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0309, 0.0444, 0.0456, 0.0371, 0.0425, 0.0340, 0.0395], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 10:40:37,647 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2090.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:40:49,239 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-26 10:40:50,244 INFO [finetune.py:976] (2/7) Epoch 1, batch 2100, loss[loss=0.5524, simple_loss=0.4949, pruned_loss=0.3049, over 4819.00 frames. ], tot_loss[loss=0.6571, simple_loss=0.531, pruned_loss=0.4051, over 955111.29 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 2.0 +2023-04-26 10:41:11,498 INFO [optim.py:369] (2/7) 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 10:41:31,944 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2138.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:41:41,040 INFO [finetune.py:976] (2/7) Epoch 1, batch 2150, loss[loss=0.5827, simple_loss=0.5235, pruned_loss=0.321, over 4832.00 frames. ], tot_loss[loss=0.6378, simple_loss=0.5251, pruned_loss=0.3858, over 955776.24 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:41:41,808 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.06 vs. limit=5.0 +2023-04-26 10:42:25,868 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 1, batch 2200, loss[loss=0.5857, simple_loss=0.5182, pruned_loss=0.3266, over 4927.00 frames. ], tot_loss[loss=0.6169, simple_loss=0.517, pruned_loss=0.3666, over 955001.22 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:42:37,619 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2213.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:42:38,986 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 10:42:40,428 INFO [optim.py:369] (2/7) 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 10:42:42,828 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2222.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:42:45,139 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2226.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:42:55,404 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6939, 1.5962, 1.8244, 2.1021, 2.1577, 1.6070, 1.3047, 1.7121], + device='cuda:2'), covar=tensor([0.1189, 0.1391, 0.1014, 0.0834, 0.0712, 0.1136, 0.1349, 0.0934], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0208, 0.0188, 0.0167, 0.0167, 0.0183, 0.0162, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 10:43:02,982 INFO [finetune.py:976] (2/7) Epoch 1, batch 2250, loss[loss=0.5677, simple_loss=0.5127, pruned_loss=0.3113, over 4900.00 frames. ], tot_loss[loss=0.5959, simple_loss=0.5075, pruned_loss=0.3485, over 954525.21 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:43:12,383 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2267.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:43:14,135 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2270.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:43:57,382 INFO [finetune.py:976] (2/7) Epoch 1, batch 2300, loss[loss=0.5091, simple_loss=0.4659, pruned_loss=0.2762, over 4863.00 frames. ], tot_loss[loss=0.5725, simple_loss=0.4956, pruned_loss=0.3297, over 954231.06 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:44:17,315 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2315.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:44:19,007 INFO [optim.py:369] (2/7) 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 10:44:43,390 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2340.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:44:51,821 INFO [finetune.py:976] (2/7) Epoch 1, batch 2350, loss[loss=0.3688, simple_loss=0.3618, pruned_loss=0.1879, over 4782.00 frames. ], tot_loss[loss=0.5475, simple_loss=0.4807, pruned_loss=0.311, over 953622.96 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 4.0 +2023-04-26 10:45:40,789 INFO [finetune.py:976] (2/7) Epoch 1, batch 2400, loss[loss=0.4327, simple_loss=0.4171, pruned_loss=0.2242, over 4918.00 frames. ], tot_loss[loss=0.5268, simple_loss=0.4685, pruned_loss=0.2955, over 952766.72 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:45:40,914 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2401.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:45:51,692 INFO [optim.py:369] (2/7) 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 10:46:11,803 INFO [finetune.py:976] (2/7) Epoch 1, batch 2450, loss[loss=0.5036, simple_loss=0.4656, pruned_loss=0.2708, over 4863.00 frames. ], tot_loss[loss=0.5089, simple_loss=0.4577, pruned_loss=0.2824, over 952967.70 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:46:39,137 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2494.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:46:44,083 INFO [finetune.py:976] (2/7) Epoch 1, batch 2500, loss[loss=0.5336, simple_loss=0.5063, pruned_loss=0.2805, over 4808.00 frames. ], tot_loss[loss=0.4972, simple_loss=0.4531, pruned_loss=0.2725, over 955683.79 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:46:52,666 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2513.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:46:56,028 INFO [optim.py:369] (2/7) 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 10:47:00,963 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2526.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:47:10,440 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2542.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:47:16,166 INFO [finetune.py:976] (2/7) Epoch 1, batch 2550, loss[loss=0.5393, simple_loss=0.5056, pruned_loss=0.2865, over 4728.00 frames. ], tot_loss[loss=0.4916, simple_loss=0.4531, pruned_loss=0.2665, over 957263.33 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:47:35,366 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2561.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:47:38,287 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3105, 1.6351, 1.1080, 1.7061, 1.4686, 1.1758, 1.4690, 1.0918], + device='cuda:2'), covar=tensor([0.2095, 0.1606, 0.1865, 0.1427, 0.3328, 0.1757, 0.1992, 0.3103], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0302, 0.0224, 0.0282, 0.0298, 0.0258, 0.0259, 0.0272], + device='cuda:2'), out_proj_covar=tensor([1.1894e-04, 1.2354e-04, 9.1607e-05, 1.1441e-04, 1.2367e-04, 1.0429e-04, + 1.0750e-04, 1.1096e-04], device='cuda:2') +2023-04-26 10:47:50,046 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2574.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:48:06,448 INFO [finetune.py:976] (2/7) Epoch 1, batch 2600, loss[loss=0.4931, simple_loss=0.468, pruned_loss=0.2591, over 4811.00 frames. ], tot_loss[loss=0.4825, simple_loss=0.4485, pruned_loss=0.2593, over 956098.87 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:48:18,465 INFO [optim.py:369] (2/7) 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 10:48:38,254 INFO [finetune.py:976] (2/7) Epoch 1, batch 2650, loss[loss=0.4136, simple_loss=0.4263, pruned_loss=0.2005, over 4811.00 frames. ], tot_loss[loss=0.4742, simple_loss=0.4445, pruned_loss=0.2528, over 952682.16 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:49:01,801 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2685.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:49:14,010 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2696.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 10:49:16,816 INFO [finetune.py:976] (2/7) Epoch 1, batch 2700, loss[loss=0.3748, simple_loss=0.3694, pruned_loss=0.1901, over 4738.00 frames. ], tot_loss[loss=0.462, simple_loss=0.4377, pruned_loss=0.2438, over 953914.10 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:49:39,497 INFO [optim.py:369] (2/7) 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 10:50:04,158 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2746.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 10:50:13,027 INFO [finetune.py:976] (2/7) Epoch 1, batch 2750, loss[loss=0.3782, simple_loss=0.3791, pruned_loss=0.1887, over 4770.00 frames. ], tot_loss[loss=0.4498, simple_loss=0.4295, pruned_loss=0.2356, over 955239.78 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:51:16,384 INFO [finetune.py:976] (2/7) Epoch 1, batch 2800, loss[loss=0.381, simple_loss=0.3869, pruned_loss=0.1875, over 4865.00 frames. ], tot_loss[loss=0.4384, simple_loss=0.4217, pruned_loss=0.228, over 954544.06 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:51:38,598 INFO [optim.py:369] (2/7) 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 10:51:49,422 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2824.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 10:52:12,970 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6708, 1.6558, 1.9010, 1.9684, 1.9929, 1.4993, 1.0178, 1.7752], + device='cuda:2'), covar=tensor([0.1341, 0.1367, 0.0915, 0.0987, 0.0851, 0.1360, 0.1565, 0.0923], + device='cuda:2'), in_proj_covar=tensor([0.0204, 0.0211, 0.0191, 0.0174, 0.0173, 0.0190, 0.0168, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 10:52:22,752 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-04-26 10:52:24,241 INFO [finetune.py:976] (2/7) Epoch 1, batch 2850, loss[loss=0.3824, simple_loss=0.3808, pruned_loss=0.192, over 4760.00 frames. ], tot_loss[loss=0.4324, simple_loss=0.4175, pruned_loss=0.2239, over 954719.59 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:52:26,766 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2478, 1.3974, 1.5595, 1.7824, 1.5000, 1.2623, 1.4134, 1.3327], + device='cuda:2'), covar=tensor([20.9622, 33.3751, 31.1860, 20.0816, 27.3855, 39.6872, 43.8167, 24.5077], + device='cuda:2'), in_proj_covar=tensor([0.0472, 0.0537, 0.0609, 0.0581, 0.0511, 0.0578, 0.0592, 0.0591], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 10:53:07,935 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2885.0, num_to_drop=2, layers_to_drop={1, 2} +2023-04-26 10:53:29,242 INFO [finetune.py:976] (2/7) Epoch 1, batch 2900, loss[loss=0.3673, simple_loss=0.3674, pruned_loss=0.1836, over 4732.00 frames. ], tot_loss[loss=0.4323, simple_loss=0.419, pruned_loss=0.223, over 955148.21 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:53:30,044 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-26 10:53:46,011 INFO [optim.py:369] (2/7) 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 10:53:59,799 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2937.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:54:08,441 INFO [finetune.py:976] (2/7) Epoch 1, batch 2950, loss[loss=0.4092, simple_loss=0.3894, pruned_loss=0.2145, over 4387.00 frames. ], tot_loss[loss=0.4321, simple_loss=0.4214, pruned_loss=0.2216, over 956140.81 frames. ], batch size: 19, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:54:10,503 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 10:54:15,737 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 10:54:23,855 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1748, 2.4201, 1.0518, 1.5281, 1.8114, 1.3479, 3.3994, 1.7868], + device='cuda:2'), covar=tensor([0.0687, 0.0621, 0.0867, 0.1129, 0.0578, 0.0946, 0.0193, 0.0595], + device='cuda:2'), in_proj_covar=tensor([0.0057, 0.0072, 0.0052, 0.0049, 0.0054, 0.0055, 0.0087, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 10:54:37,932 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2996.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 10:54:39,138 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2998.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:54:40,894 INFO [finetune.py:976] (2/7) Epoch 1, batch 3000, loss[loss=0.4637, simple_loss=0.4768, pruned_loss=0.2253, over 4808.00 frames. ], tot_loss[loss=0.4299, simple_loss=0.4212, pruned_loss=0.2194, over 956360.69 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:54:40,895 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 10:54:45,819 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0274, 0.6962, 0.7872, 0.8776, 0.7207, 0.7066, 0.4764, 0.5535], + device='cuda:2'), covar=tensor([ 7.4821, 9.9971, 4.4347, 13.6470, 12.4523, 8.5774, 11.6747, 11.8027], + device='cuda:2'), in_proj_covar=tensor([0.0267, 0.0286, 0.0226, 0.0354, 0.0247, 0.0236, 0.0281, 0.0233], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 10:54:47,770 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6178, 1.3552, 0.6593, 1.1865, 1.5061, 1.4706, 1.4056, 1.3079], + device='cuda:2'), covar=tensor([0.0650, 0.0618, 0.0559, 0.0757, 0.0384, 0.0718, 0.0676, 0.0860], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0026, 0.0024, 0.0031, 0.0021, 0.0031, 0.0030, 0.0033], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0043, 0.0037, 0.0049, 0.0036, 0.0048, 0.0047, 0.0051], + device='cuda:2') +2023-04-26 10:54:48,537 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7211, 1.7995, 1.6637, 1.4838, 1.9593, 1.4760, 2.3584, 1.4491], + device='cuda:2'), covar=tensor([0.3419, 0.1046, 0.2879, 0.1954, 0.1131, 0.1884, 0.0762, 0.3081], + device='cuda:2'), in_proj_covar=tensor([0.0303, 0.0312, 0.0380, 0.0325, 0.0360, 0.0336, 0.0350, 0.0364], + device='cuda:2'), out_proj_covar=tensor([9.3729e-05, 9.6421e-05, 1.1766e-04, 1.0148e-04, 1.1050e-04, 1.0257e-04, + 1.0576e-04, 1.1288e-04], device='cuda:2') +2023-04-26 10:54:51,382 INFO [finetune.py:1010] (2/7) Epoch 1, validation: loss=0.4217, simple_loss=0.4614, pruned_loss=0.191, over 2265189.00 frames. +2023-04-26 10:54:51,383 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 5493MB +2023-04-26 10:55:01,571 INFO [optim.py:369] (2/7) 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 10:55:02,264 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3019.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:55:16,244 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3041.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:55:18,046 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3044.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:55:18,142 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=7.29 vs. limit=5.0 +2023-04-26 10:55:23,634 INFO [finetune.py:976] (2/7) Epoch 1, batch 3050, loss[loss=0.3993, simple_loss=0.4082, pruned_loss=0.1952, over 4810.00 frames. ], tot_loss[loss=0.424, simple_loss=0.4184, pruned_loss=0.215, over 956374.30 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:55:42,237 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3080.0, num_to_drop=2, layers_to_drop={0, 2} +2023-04-26 10:55:56,460 INFO [finetune.py:976] (2/7) Epoch 1, batch 3100, loss[loss=0.3926, simple_loss=0.3841, pruned_loss=0.2006, over 4819.00 frames. ], tot_loss[loss=0.4145, simple_loss=0.4123, pruned_loss=0.2085, over 956975.63 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:56:12,965 INFO [optim.py:369] (2/7) 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 10:56:55,555 INFO [finetune.py:976] (2/7) Epoch 1, batch 3150, loss[loss=0.3513, simple_loss=0.368, pruned_loss=0.1673, over 4903.00 frames. ], tot_loss[loss=0.4062, simple_loss=0.4056, pruned_loss=0.2035, over 957306.02 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:57:32,970 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3180.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 10:57:46,511 INFO [finetune.py:976] (2/7) Epoch 1, batch 3200, loss[loss=0.3737, simple_loss=0.3929, pruned_loss=0.1772, over 4768.00 frames. ], tot_loss[loss=0.3964, simple_loss=0.3981, pruned_loss=0.1974, over 956754.79 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:58:16,458 INFO [optim.py:369] (2/7) 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 10:58:40,550 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6399, 1.7630, 2.0709, 1.9958, 2.1056, 1.5321, 1.0884, 1.8471], + device='cuda:2'), covar=tensor([0.1443, 0.1309, 0.0804, 0.1124, 0.0799, 0.1484, 0.1650, 0.1003], + device='cuda:2'), in_proj_covar=tensor([0.0203, 0.0208, 0.0188, 0.0174, 0.0172, 0.0189, 0.0168, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 10:58:52,915 INFO [finetune.py:976] (2/7) Epoch 1, batch 3250, loss[loss=0.4104, simple_loss=0.4035, pruned_loss=0.2086, over 4822.00 frames. ], tot_loss[loss=0.3942, simple_loss=0.397, pruned_loss=0.1957, over 958551.20 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 10:59:47,339 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3288.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 10:59:50,302 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3293.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:00:00,685 INFO [finetune.py:976] (2/7) Epoch 1, batch 3300, loss[loss=0.4165, simple_loss=0.4292, pruned_loss=0.202, over 4841.00 frames. ], tot_loss[loss=0.3946, simple_loss=0.3988, pruned_loss=0.1952, over 955820.81 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:00:18,059 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5599, 2.6302, 1.4029, 1.7019, 1.0752, 1.2115, 1.3919, 1.0537], + device='cuda:2'), covar=tensor([0.3184, 0.3085, 0.4588, 0.4781, 0.5184, 0.3980, 0.3481, 0.4713], + device='cuda:2'), in_proj_covar=tensor([0.0203, 0.0225, 0.0207, 0.0222, 0.0241, 0.0202, 0.0198, 0.0216], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:00:19,092 INFO [optim.py:369] (2/7) 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 11:00:33,917 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3341.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:00:38,724 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3349.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 11:00:39,793 INFO [finetune.py:976] (2/7) Epoch 1, batch 3350, loss[loss=0.4069, simple_loss=0.4144, pruned_loss=0.1997, over 4791.00 frames. ], tot_loss[loss=0.3947, simple_loss=0.4003, pruned_loss=0.1946, over 957183.83 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:00:57,399 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3375.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:01:03,515 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-26 11:01:05,772 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 1, batch 3400, loss[loss=0.3442, simple_loss=0.3566, pruned_loss=0.1659, over 4762.00 frames. ], tot_loss[loss=0.3936, simple_loss=0.3999, pruned_loss=0.1937, over 956338.51 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:01:14,907 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1995, 1.7901, 1.4689, 1.9113, 1.7875, 2.0320, 1.6954, 4.1159], + device='cuda:2'), covar=tensor([0.0748, 0.0719, 0.0791, 0.1241, 0.0706, 0.0707, 0.0748, 0.0119], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0040, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 11:01:18,488 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0151, 1.9141, 1.6202, 1.7031, 2.0747, 1.6695, 2.4688, 1.4487], + device='cuda:2'), covar=tensor([0.3776, 0.1323, 0.3419, 0.2418, 0.1456, 0.2000, 0.0901, 0.3250], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0321, 0.0392, 0.0333, 0.0367, 0.0343, 0.0360, 0.0373], + device='cuda:2'), out_proj_covar=tensor([9.5720e-05, 9.9167e-05, 1.2140e-04, 1.0426e-04, 1.1251e-04, 1.0482e-04, + 1.0900e-04, 1.1569e-04], device='cuda:2') +2023-04-26 11:01:21,944 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1852, 4.6025, 1.3754, 2.7140, 2.8084, 3.4173, 3.1409, 1.4704], + device='cuda:2'), covar=tensor([0.1219, 0.1051, 0.2121, 0.1131, 0.0846, 0.0923, 0.1025, 0.1830], + device='cuda:2'), in_proj_covar=tensor([0.0122, 0.0269, 0.0148, 0.0131, 0.0143, 0.0165, 0.0129, 0.0132], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:01:23,589 INFO [optim.py:369] (2/7) 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 11:01:45,507 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.12 vs. limit=5.0 +2023-04-26 11:01:57,861 INFO [finetune.py:976] (2/7) Epoch 1, batch 3450, loss[loss=0.3186, simple_loss=0.3528, pruned_loss=0.1422, over 4760.00 frames. ], tot_loss[loss=0.3907, simple_loss=0.3984, pruned_loss=0.1916, over 954499.69 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:02:35,997 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3480.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:02:49,063 INFO [finetune.py:976] (2/7) Epoch 1, batch 3500, loss[loss=0.3698, simple_loss=0.3735, pruned_loss=0.183, over 4931.00 frames. ], tot_loss[loss=0.3832, simple_loss=0.3922, pruned_loss=0.1871, over 954921.05 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:03:00,464 INFO [optim.py:369] (2/7) 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 11:03:07,523 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3528.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:03:24,396 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3546.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:03:27,265 INFO [finetune.py:976] (2/7) Epoch 1, batch 3550, loss[loss=0.3458, simple_loss=0.3772, pruned_loss=0.1572, over 4928.00 frames. ], tot_loss[loss=0.3793, simple_loss=0.3884, pruned_loss=0.1851, over 954716.82 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:03:56,707 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6801, 1.3335, 1.1845, 1.3208, 1.8899, 1.7475, 1.4076, 1.1911], + device='cuda:2'), covar=tensor([0.1229, 0.1483, 0.2221, 0.1685, 0.0586, 0.0971, 0.1453, 0.1779], + device='cuda:2'), in_proj_covar=tensor([0.0333, 0.0345, 0.0346, 0.0308, 0.0357, 0.0379, 0.0328, 0.0365], + device='cuda:2'), out_proj_covar=tensor([7.2897e-05, 7.4364e-05, 7.4779e-05, 6.4525e-05, 7.6260e-05, 8.2953e-05, + 7.1506e-05, 7.8966e-05], device='cuda:2') +2023-04-26 11:04:11,258 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3593.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:04:16,043 INFO [finetune.py:976] (2/7) Epoch 1, batch 3600, loss[loss=0.3732, simple_loss=0.3687, pruned_loss=0.1888, over 4074.00 frames. ], tot_loss[loss=0.3721, simple_loss=0.3829, pruned_loss=0.1807, over 953799.59 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:04:19,852 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3607.0, num_to_drop=2, layers_to_drop={1, 3} +2023-04-26 11:04:20,435 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9387, 1.9578, 1.6583, 1.8321, 2.1615, 1.7238, 2.5570, 1.4692], + device='cuda:2'), covar=tensor([0.4562, 0.1488, 0.4366, 0.2566, 0.1531, 0.2529, 0.0907, 0.3885], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0323, 0.0395, 0.0335, 0.0370, 0.0346, 0.0363, 0.0377], + device='cuda:2'), out_proj_covar=tensor([9.6601e-05, 9.9840e-05, 1.2245e-04, 1.0481e-04, 1.1352e-04, 1.0579e-04, + 1.0980e-04, 1.1685e-04], device='cuda:2') +2023-04-26 11:04:22,293 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2736, 0.8993, 1.0027, 1.2858, 0.9827, 0.9078, 0.7839, 0.8894], + device='cuda:2'), covar=tensor([3.0287, 3.5270, 1.7290, 5.8431, 4.1736, 2.7835, 5.6993, 4.1666], + device='cuda:2'), in_proj_covar=tensor([0.0263, 0.0285, 0.0224, 0.0349, 0.0244, 0.0235, 0.0280, 0.0232], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:04:26,349 INFO [optim.py:369] (2/7) 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 11:04:26,620 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.98 vs. limit=5.0 +2023-04-26 11:04:42,910 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3641.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:04:44,770 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3644.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 11:04:48,954 INFO [finetune.py:976] (2/7) Epoch 1, batch 3650, loss[loss=0.457, simple_loss=0.4452, pruned_loss=0.2344, over 4910.00 frames. ], tot_loss[loss=0.3731, simple_loss=0.3846, pruned_loss=0.1807, over 954259.07 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:04:58,901 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4309, 1.3488, 4.1360, 3.8367, 3.6536, 3.9281, 3.9550, 3.6540], + device='cuda:2'), covar=tensor([0.7052, 0.6125, 0.1086, 0.1653, 0.1239, 0.1598, 0.1293, 0.1496], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0321, 0.0460, 0.0471, 0.0388, 0.0442, 0.0352, 0.0412], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 11:05:04,834 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3675.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:05:05,416 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3676.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:05:16,485 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=6.10 vs. limit=5.0 +2023-04-26 11:05:18,881 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 11:05:19,440 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3585, 1.1735, 1.8332, 1.8323, 1.2388, 1.0141, 1.3994, 0.9233], + device='cuda:2'), covar=tensor([0.1519, 0.1257, 0.0681, 0.0710, 0.1210, 0.1651, 0.1168, 0.1657], + device='cuda:2'), in_proj_covar=tensor([0.0076, 0.0082, 0.0077, 0.0079, 0.0095, 0.0098, 0.0096, 0.0083], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:2') +2023-04-26 11:05:23,031 INFO [finetune.py:976] (2/7) Epoch 1, batch 3700, loss[loss=0.4041, simple_loss=0.427, pruned_loss=0.1907, over 4812.00 frames. ], tot_loss[loss=0.3735, simple_loss=0.3868, pruned_loss=0.1802, over 953029.61 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:05:44,463 INFO [optim.py:369] (2/7) 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 11:05:45,191 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6313, 1.7440, 0.6305, 1.3317, 1.9534, 1.5271, 1.5076, 1.5439], + device='cuda:2'), covar=tensor([0.0617, 0.0511, 0.0563, 0.0677, 0.0348, 0.0683, 0.0621, 0.0776], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0021, 0.0031, 0.0030, 0.0033], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0043, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:2') +2023-04-26 11:05:53,052 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3723.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:06:12,566 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 1, batch 3750, loss[loss=0.4059, simple_loss=0.4028, pruned_loss=0.2045, over 4890.00 frames. ], tot_loss[loss=0.3731, simple_loss=0.3879, pruned_loss=0.1791, over 953356.24 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:07:07,655 INFO [finetune.py:976] (2/7) Epoch 1, batch 3800, loss[loss=0.392, simple_loss=0.4152, pruned_loss=0.1845, over 4852.00 frames. ], tot_loss[loss=0.3716, simple_loss=0.3878, pruned_loss=0.1777, over 952941.49 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:07:09,071 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-26 11:07:29,603 INFO [optim.py:369] (2/7) 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 11:07:44,259 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-26 11:08:14,574 INFO [finetune.py:976] (2/7) Epoch 1, batch 3850, loss[loss=0.3104, simple_loss=0.3442, pruned_loss=0.1383, over 4889.00 frames. ], tot_loss[loss=0.3667, simple_loss=0.3842, pruned_loss=0.1746, over 955984.94 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:09:22,876 INFO [finetune.py:976] (2/7) Epoch 1, batch 3900, loss[loss=0.3233, simple_loss=0.3491, pruned_loss=0.1488, over 4899.00 frames. ], tot_loss[loss=0.361, simple_loss=0.3787, pruned_loss=0.1717, over 955176.52 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:09:29,504 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3902.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 11:09:44,795 INFO [optim.py:369] (2/7) 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 11:09:49,808 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3926.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:09:51,283 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-26 11:10:02,243 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3944.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:10:07,407 INFO [finetune.py:976] (2/7) Epoch 1, batch 3950, loss[loss=0.3321, simple_loss=0.351, pruned_loss=0.1566, over 4829.00 frames. ], tot_loss[loss=0.3526, simple_loss=0.3713, pruned_loss=0.1669, over 953892.13 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:10:15,839 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4743, 1.4653, 1.5832, 1.8032, 1.8998, 1.3834, 1.0283, 1.6625], + device='cuda:2'), covar=tensor([0.1034, 0.1295, 0.0939, 0.0783, 0.0612, 0.1086, 0.1135, 0.0691], + device='cuda:2'), in_proj_covar=tensor([0.0203, 0.0206, 0.0187, 0.0175, 0.0173, 0.0190, 0.0169, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:10:30,549 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3987.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:10:31,729 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1446, 2.4634, 0.8657, 1.5867, 1.6009, 1.9160, 1.7261, 0.8621], + device='cuda:2'), covar=tensor([0.1321, 0.1044, 0.1668, 0.1232, 0.1039, 0.0918, 0.1333, 0.1785], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0273, 0.0150, 0.0132, 0.0143, 0.0167, 0.0130, 0.0134], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:10:34,042 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3992.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:10:42,160 INFO [finetune.py:976] (2/7) Epoch 1, batch 4000, loss[loss=0.3872, simple_loss=0.405, pruned_loss=0.1847, over 4842.00 frames. ], tot_loss[loss=0.3489, simple_loss=0.3684, pruned_loss=0.1648, over 952409.63 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:10:44,044 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.41 vs. limit=5.0 +2023-04-26 11:10:53,096 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 11:10:54,137 INFO [optim.py:369] (2/7) 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 11:11:02,812 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4032.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:11:09,574 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1253, 0.7316, 0.8812, 0.7014, 1.2272, 1.0618, 0.8590, 1.0125], + device='cuda:2'), covar=tensor([0.1376, 0.1838, 0.2393, 0.2038, 0.1015, 0.1409, 0.1719, 0.2030], + device='cuda:2'), in_proj_covar=tensor([0.0332, 0.0344, 0.0345, 0.0310, 0.0355, 0.0378, 0.0326, 0.0364], + device='cuda:2'), out_proj_covar=tensor([7.2638e-05, 7.4204e-05, 7.4408e-05, 6.4944e-05, 7.5702e-05, 8.2873e-05, + 7.1130e-05, 7.8761e-05], device='cuda:2') +2023-04-26 11:11:15,866 INFO [finetune.py:976] (2/7) Epoch 1, batch 4050, loss[loss=0.3368, simple_loss=0.3709, pruned_loss=0.1514, over 4912.00 frames. ], tot_loss[loss=0.3487, simple_loss=0.3693, pruned_loss=0.1641, over 951059.12 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:11:45,035 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-26 11:11:49,883 INFO [finetune.py:976] (2/7) Epoch 1, batch 4100, loss[loss=0.3278, simple_loss=0.3444, pruned_loss=0.1556, over 4731.00 frames. ], tot_loss[loss=0.35, simple_loss=0.3718, pruned_loss=0.1641, over 951155.08 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-04-26 11:12:08,256 INFO [optim.py:369] (2/7) 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 11:12:34,740 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4141.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:12:41,352 INFO [finetune.py:976] (2/7) Epoch 1, batch 4150, loss[loss=0.3586, simple_loss=0.3902, pruned_loss=0.1635, over 4903.00 frames. ], tot_loss[loss=0.3493, simple_loss=0.3723, pruned_loss=0.1631, over 952807.07 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:12:41,536 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.09 vs. limit=5.0 +2023-04-26 11:12:52,521 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-26 11:13:15,084 INFO [finetune.py:976] (2/7) Epoch 1, batch 4200, loss[loss=0.3251, simple_loss=0.3525, pruned_loss=0.1489, over 4800.00 frames. ], tot_loss[loss=0.3452, simple_loss=0.3698, pruned_loss=0.1603, over 954152.73 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:13:16,209 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4202.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:13:16,246 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4202.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:13:26,499 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8107, 1.9945, 1.6048, 1.6043, 1.8898, 1.5680, 2.4183, 1.2805], + device='cuda:2'), covar=tensor([0.3601, 0.1095, 0.3544, 0.2094, 0.1700, 0.2213, 0.0944, 0.3517], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0328, 0.0401, 0.0339, 0.0374, 0.0348, 0.0368, 0.0381], + device='cuda:2'), out_proj_covar=tensor([9.7700e-05, 1.0133e-04, 1.2417e-04, 1.0592e-04, 1.1464e-04, 1.0618e-04, + 1.1131e-04, 1.1783e-04], device='cuda:2') +2023-04-26 11:13:28,642 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4250.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:14:05,986 INFO [finetune.py:976] (2/7) Epoch 1, batch 4250, loss[loss=0.4066, simple_loss=0.4027, pruned_loss=0.2052, over 4856.00 frames. ], tot_loss[loss=0.3382, simple_loss=0.3636, pruned_loss=0.1564, over 955250.79 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:14:51,196 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4282.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:14:59,281 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4288.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:15:07,277 INFO [finetune.py:976] (2/7) Epoch 1, batch 4300, loss[loss=0.3315, simple_loss=0.349, pruned_loss=0.157, over 4823.00 frames. ], tot_loss[loss=0.3323, simple_loss=0.3583, pruned_loss=0.1531, over 955617.66 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:15:19,787 INFO [optim.py:369] (2/7) 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 11:15:29,482 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4332.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:15:39,946 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4349.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:15:41,062 INFO [finetune.py:976] (2/7) Epoch 1, batch 4350, loss[loss=0.3521, simple_loss=0.3636, pruned_loss=0.1703, over 4909.00 frames. ], tot_loss[loss=0.3269, simple_loss=0.353, pruned_loss=0.1504, over 955500.81 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:15:41,163 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0327, 1.6545, 1.4604, 1.8947, 1.7596, 2.0119, 1.6547, 3.6328], + device='cuda:2'), covar=tensor([0.0787, 0.0765, 0.0846, 0.1282, 0.0688, 0.0593, 0.0726, 0.0136], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0040, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 11:16:02,408 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4380.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:16:15,320 INFO [finetune.py:976] (2/7) Epoch 1, batch 4400, loss[loss=0.2938, simple_loss=0.3165, pruned_loss=0.1355, over 3988.00 frames. ], tot_loss[loss=0.3276, simple_loss=0.354, pruned_loss=0.1506, over 955807.07 frames. ], batch size: 17, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:16:26,707 INFO [optim.py:369] (2/7) 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 11:16:45,938 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9294, 1.4777, 4.6922, 4.4040, 4.0782, 4.4226, 4.3249, 4.0818], + device='cuda:2'), covar=tensor([0.6291, 0.6484, 0.0976, 0.1511, 0.1059, 0.1444, 0.1175, 0.1573], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0322, 0.0458, 0.0466, 0.0386, 0.0442, 0.0349, 0.0411], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 11:16:48,901 INFO [finetune.py:976] (2/7) Epoch 1, batch 4450, loss[loss=0.3123, simple_loss=0.3433, pruned_loss=0.1407, over 4836.00 frames. ], tot_loss[loss=0.3311, simple_loss=0.3586, pruned_loss=0.1518, over 957489.01 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:17:19,614 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 1, batch 4500, loss[loss=0.374, simple_loss=0.3932, pruned_loss=0.1773, over 4913.00 frames. ], tot_loss[loss=0.332, simple_loss=0.3603, pruned_loss=0.1519, over 957215.97 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:17:32,585 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 11:17:32,938 INFO [optim.py:369] (2/7) 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 11:17:50,729 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 11:18:15,785 INFO [finetune.py:976] (2/7) Epoch 1, batch 4550, loss[loss=0.2583, simple_loss=0.2806, pruned_loss=0.118, over 4199.00 frames. ], tot_loss[loss=0.3333, simple_loss=0.3623, pruned_loss=0.1521, over 957354.14 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:19:01,282 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4582.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:19:13,034 INFO [finetune.py:976] (2/7) Epoch 1, batch 4600, loss[loss=0.2714, simple_loss=0.3115, pruned_loss=0.1157, over 4767.00 frames. ], tot_loss[loss=0.3315, simple_loss=0.3608, pruned_loss=0.1511, over 956416.68 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:19:19,948 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5507, 1.9075, 1.7000, 2.1452, 1.9978, 2.2666, 1.8065, 3.7219], + device='cuda:2'), covar=tensor([0.0654, 0.0698, 0.0771, 0.1196, 0.0617, 0.0520, 0.0698, 0.0184], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0041, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 11:19:23,471 INFO [optim.py:369] (2/7) 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 11:19:33,261 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4630.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:19:42,842 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4644.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:19:47,128 INFO [finetune.py:976] (2/7) Epoch 1, batch 4650, loss[loss=0.2603, simple_loss=0.2945, pruned_loss=0.1131, over 4728.00 frames. ], tot_loss[loss=0.3254, simple_loss=0.355, pruned_loss=0.1479, over 956678.47 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:20:13,215 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-26 11:20:42,012 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.73 vs. limit=2.0 +2023-04-26 11:20:42,427 INFO [finetune.py:976] (2/7) Epoch 1, batch 4700, loss[loss=0.2442, simple_loss=0.2948, pruned_loss=0.09683, over 4932.00 frames. ], tot_loss[loss=0.3186, simple_loss=0.3493, pruned_loss=0.144, over 958506.85 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:21:03,816 INFO [optim.py:369] (2/7) 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 11:21:39,265 INFO [finetune.py:976] (2/7) Epoch 1, batch 4750, loss[loss=0.3173, simple_loss=0.3353, pruned_loss=0.1496, over 4752.00 frames. ], tot_loss[loss=0.3132, simple_loss=0.3446, pruned_loss=0.1409, over 958755.31 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:21:43,100 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-04-26 11:21:46,771 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6737, 1.2186, 1.3289, 1.1417, 1.7646, 1.6402, 1.2392, 1.2850], + device='cuda:2'), covar=tensor([0.1301, 0.1770, 0.2149, 0.2087, 0.0926, 0.1319, 0.1971, 0.2015], + device='cuda:2'), in_proj_covar=tensor([0.0332, 0.0345, 0.0349, 0.0314, 0.0356, 0.0381, 0.0327, 0.0364], + device='cuda:2'), out_proj_covar=tensor([7.2753e-05, 7.4269e-05, 7.5333e-05, 6.5656e-05, 7.5916e-05, 8.3462e-05, + 7.1405e-05, 7.8786e-05], device='cuda:2') +2023-04-26 11:22:11,025 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 1, batch 4800, loss[loss=0.3738, simple_loss=0.3723, pruned_loss=0.1876, over 4740.00 frames. ], tot_loss[loss=0.3161, simple_loss=0.3473, pruned_loss=0.1425, over 956174.46 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:22:24,009 INFO [optim.py:369] (2/7) 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 11:22:43,635 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4845.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:22:47,741 INFO [finetune.py:976] (2/7) Epoch 1, batch 4850, loss[loss=0.3228, simple_loss=0.3671, pruned_loss=0.1393, over 4813.00 frames. ], tot_loss[loss=0.3202, simple_loss=0.352, pruned_loss=0.1442, over 955206.63 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:23:08,730 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3906, 1.4094, 1.3647, 1.1742, 1.4217, 1.0904, 1.8051, 1.2148], + device='cuda:2'), covar=tensor([0.4013, 0.1563, 0.4640, 0.2325, 0.1570, 0.2193, 0.1306, 0.3858], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0340, 0.0417, 0.0353, 0.0388, 0.0361, 0.0384, 0.0394], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:23:22,879 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-26 11:23:26,202 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0262, 2.1273, 1.6896, 1.8637, 2.1257, 1.5965, 2.6572, 1.4317], + device='cuda:2'), covar=tensor([0.4871, 0.1548, 0.5166, 0.2860, 0.2094, 0.2791, 0.1232, 0.4016], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0340, 0.0419, 0.0353, 0.0389, 0.0361, 0.0384, 0.0395], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:23:33,605 INFO [finetune.py:976] (2/7) Epoch 1, batch 4900, loss[loss=0.2833, simple_loss=0.3333, pruned_loss=0.1166, over 4770.00 frames. ], tot_loss[loss=0.3204, simple_loss=0.3523, pruned_loss=0.1443, over 953330.17 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:23:36,235 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0106, 0.9202, 1.2346, 1.1473, 1.0089, 0.8360, 0.9209, 0.5703], + device='cuda:2'), covar=tensor([0.0887, 0.0904, 0.0699, 0.0757, 0.1011, 0.1673, 0.0729, 0.1416], + device='cuda:2'), in_proj_covar=tensor([0.0074, 0.0080, 0.0077, 0.0077, 0.0092, 0.0096, 0.0093, 0.0082], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:2') +2023-04-26 11:23:46,616 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1035, 1.4361, 1.3242, 1.7151, 1.5143, 1.9317, 1.3563, 3.1651], + device='cuda:2'), covar=tensor([0.0759, 0.0754, 0.0827, 0.1302, 0.0685, 0.0641, 0.0754, 0.0198], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0040, 0.0041, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 11:23:49,558 INFO [optim.py:369] (2/7) 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 11:24:07,371 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9047, 1.0481, 1.4590, 2.0279, 1.5068, 1.1488, 0.9468, 1.3388], + device='cuda:2'), covar=tensor([0.4705, 0.6014, 0.3007, 0.6641, 0.5782, 0.4286, 0.8207, 0.4847], + device='cuda:2'), in_proj_covar=tensor([0.0253, 0.0272, 0.0213, 0.0331, 0.0231, 0.0225, 0.0266, 0.0219], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:24:09,955 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.05 vs. limit=5.0 +2023-04-26 11:24:29,555 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4944.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:24:30,127 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.46 vs. limit=2.0 +2023-04-26 11:24:41,249 INFO [finetune.py:976] (2/7) Epoch 1, batch 4950, loss[loss=0.3125, simple_loss=0.3522, pruned_loss=0.1364, over 4811.00 frames. ], tot_loss[loss=0.3198, simple_loss=0.3529, pruned_loss=0.1434, over 953959.63 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:25:18,997 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4992.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:25:26,587 INFO [finetune.py:976] (2/7) Epoch 1, batch 5000, loss[loss=0.3023, simple_loss=0.3127, pruned_loss=0.1459, over 4222.00 frames. ], tot_loss[loss=0.3155, simple_loss=0.3494, pruned_loss=0.1409, over 953084.04 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:25:37,462 INFO [optim.py:369] (2/7) 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 11:25:53,459 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9983, 3.9941, 3.2136, 4.5349, 3.7299, 3.9502, 1.9511, 3.9358], + device='cuda:2'), covar=tensor([0.1467, 0.0907, 0.3015, 0.0888, 0.2572, 0.1582, 0.4276, 0.1703], + device='cuda:2'), in_proj_covar=tensor([0.0262, 0.0232, 0.0282, 0.0326, 0.0322, 0.0271, 0.0287, 0.0287], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:26:10,645 INFO [finetune.py:976] (2/7) Epoch 1, batch 5050, loss[loss=0.3001, simple_loss=0.3359, pruned_loss=0.1322, over 4774.00 frames. ], tot_loss[loss=0.3115, simple_loss=0.3451, pruned_loss=0.1389, over 954133.30 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:26:35,462 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0590, 0.6936, 0.8002, 0.6070, 1.1695, 0.9289, 0.7829, 0.9479], + device='cuda:2'), covar=tensor([0.1564, 0.2144, 0.2903, 0.2218, 0.1040, 0.1751, 0.2016, 0.2426], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0344, 0.0349, 0.0315, 0.0356, 0.0380, 0.0328, 0.0364], + device='cuda:2'), out_proj_covar=tensor([7.2540e-05, 7.4149e-05, 7.5382e-05, 6.5889e-05, 7.5990e-05, 8.3250e-05, + 7.1428e-05, 7.8638e-05], device='cuda:2') +2023-04-26 11:27:11,310 INFO [finetune.py:976] (2/7) Epoch 1, batch 5100, loss[loss=0.3203, simple_loss=0.3237, pruned_loss=0.1584, over 3972.00 frames. ], tot_loss[loss=0.3051, simple_loss=0.3394, pruned_loss=0.1354, over 953759.47 frames. ], batch size: 17, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:27:40,805 INFO [optim.py:369] (2/7) 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 11:28:17,921 INFO [finetune.py:976] (2/7) Epoch 1, batch 5150, loss[loss=0.2773, simple_loss=0.3264, pruned_loss=0.1141, over 4917.00 frames. ], tot_loss[loss=0.3061, simple_loss=0.3395, pruned_loss=0.1363, over 952166.94 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:29:07,358 INFO [finetune.py:976] (2/7) Epoch 1, batch 5200, loss[loss=0.3498, simple_loss=0.3847, pruned_loss=0.1574, over 4823.00 frames. ], tot_loss[loss=0.3086, simple_loss=0.3433, pruned_loss=0.137, over 953620.36 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:29:21,206 INFO [optim.py:369] (2/7) 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 11:29:39,872 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2803, 2.1942, 1.8011, 2.0016, 2.3137, 1.8402, 2.8678, 1.5933], + device='cuda:2'), covar=tensor([0.4616, 0.1723, 0.4718, 0.3158, 0.2097, 0.2686, 0.1639, 0.4099], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0341, 0.0421, 0.0355, 0.0391, 0.0362, 0.0387, 0.0397], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:29:41,592 INFO [finetune.py:976] (2/7) Epoch 1, batch 5250, loss[loss=0.2775, simple_loss=0.3266, pruned_loss=0.1142, over 4763.00 frames. ], tot_loss[loss=0.3084, simple_loss=0.3444, pruned_loss=0.1362, over 953715.98 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:29:41,682 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2493, 1.1152, 5.2368, 4.8810, 4.6512, 4.9254, 4.5933, 4.5943], + device='cuda:2'), covar=tensor([0.6181, 0.6454, 0.0880, 0.1728, 0.0897, 0.1243, 0.1214, 0.1574], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0318, 0.0451, 0.0462, 0.0380, 0.0436, 0.0344, 0.0406], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 11:30:05,897 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0803, 1.4694, 1.3050, 1.6856, 1.4957, 1.9430, 1.3702, 3.3591], + device='cuda:2'), covar=tensor([0.0811, 0.0891, 0.0931, 0.1541, 0.0766, 0.0612, 0.0876, 0.0192], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0040, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 11:30:07,140 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1086, 1.6180, 1.4141, 1.8423, 1.7066, 2.1306, 1.5507, 3.6728], + device='cuda:2'), covar=tensor([0.0800, 0.0815, 0.0846, 0.1317, 0.0684, 0.0555, 0.0763, 0.0150], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0046, 0.0041, 0.0040, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 11:30:15,015 INFO [finetune.py:976] (2/7) Epoch 1, batch 5300, loss[loss=0.3061, simple_loss=0.3482, pruned_loss=0.132, over 4887.00 frames. ], tot_loss[loss=0.3098, simple_loss=0.3464, pruned_loss=0.1366, over 955569.16 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:30:27,273 INFO [optim.py:369] (2/7) 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 11:30:49,040 INFO [finetune.py:976] (2/7) Epoch 1, batch 5350, loss[loss=0.2447, simple_loss=0.2906, pruned_loss=0.09947, over 4794.00 frames. ], tot_loss[loss=0.3074, simple_loss=0.3447, pruned_loss=0.135, over 954545.38 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:31:16,465 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3330, 3.2136, 0.7990, 1.8156, 1.7928, 2.3080, 2.0873, 0.9644], + device='cuda:2'), covar=tensor([0.1403, 0.0888, 0.2078, 0.1292, 0.1116, 0.1065, 0.1350, 0.1918], + device='cuda:2'), in_proj_covar=tensor([0.0124, 0.0272, 0.0151, 0.0132, 0.0144, 0.0166, 0.0129, 0.0133], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:31:35,564 INFO [finetune.py:976] (2/7) Epoch 1, batch 5400, loss[loss=0.2835, simple_loss=0.3306, pruned_loss=0.1182, over 4822.00 frames. ], tot_loss[loss=0.3055, simple_loss=0.3425, pruned_loss=0.1343, over 954140.98 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:31:52,845 INFO [optim.py:369] (2/7) 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 11:32:06,481 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8535, 0.5838, 0.8012, 1.1246, 1.0457, 0.8669, 0.8321, 0.8191], + device='cuda:2'), covar=tensor([10.4504, 15.8368, 14.6644, 14.3475, 14.1159, 16.1996, 16.4789, 10.2133], + device='cuda:2'), in_proj_covar=tensor([0.0430, 0.0494, 0.0569, 0.0544, 0.0456, 0.0516, 0.0522, 0.0531], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:32:15,535 INFO [finetune.py:976] (2/7) Epoch 1, batch 5450, loss[loss=0.2688, simple_loss=0.3124, pruned_loss=0.1126, over 4825.00 frames. ], tot_loss[loss=0.2996, simple_loss=0.3369, pruned_loss=0.1312, over 954739.44 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:32:22,389 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3541, 1.3328, 1.3455, 1.3641, 1.4845, 1.5921, 1.4827, 1.4592], + device='cuda:2'), covar=tensor([ 3.6012, 10.0312, 6.4644, 5.4689, 5.6567, 8.5399, 8.5126, 6.7947], + device='cuda:2'), in_proj_covar=tensor([0.0260, 0.0363, 0.0293, 0.0292, 0.0320, 0.0337, 0.0352, 0.0321], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:32:37,139 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2702, 1.3953, 1.4518, 1.7152, 2.1536, 1.8948, 1.7381, 1.5531], + device='cuda:2'), covar=tensor([0.1607, 0.2232, 0.2569, 0.3316, 0.1484, 0.2304, 0.2341, 0.2086], + device='cuda:2'), in_proj_covar=tensor([0.0333, 0.0348, 0.0350, 0.0318, 0.0359, 0.0382, 0.0330, 0.0365], + device='cuda:2'), out_proj_covar=tensor([7.3025e-05, 7.5009e-05, 7.5623e-05, 6.6602e-05, 7.6672e-05, 8.3640e-05, + 7.1870e-05, 7.8991e-05], device='cuda:2') +2023-04-26 11:32:40,131 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8395, 0.6832, 0.8806, 1.1515, 1.0203, 0.8894, 0.8962, 0.8936], + device='cuda:2'), covar=tensor([ 8.4511, 14.0667, 13.8249, 12.8100, 10.1949, 15.1920, 15.6101, 10.2384], + device='cuda:2'), in_proj_covar=tensor([0.0430, 0.0495, 0.0569, 0.0545, 0.0456, 0.0516, 0.0522, 0.0531], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:32:49,240 INFO [finetune.py:976] (2/7) Epoch 1, batch 5500, loss[loss=0.3415, simple_loss=0.367, pruned_loss=0.158, over 4905.00 frames. ], tot_loss[loss=0.2972, simple_loss=0.3341, pruned_loss=0.1301, over 955046.35 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:32:49,389 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4367, 1.3321, 1.3525, 1.3684, 1.5006, 1.5891, 1.3954, 1.3332], + device='cuda:2'), covar=tensor([ 4.1180, 10.4830, 6.8957, 5.4646, 5.7106, 8.4977, 9.3770, 7.6255], + device='cuda:2'), in_proj_covar=tensor([0.0261, 0.0363, 0.0293, 0.0292, 0.0320, 0.0338, 0.0352, 0.0321], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:32:59,728 INFO [optim.py:369] (2/7) 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 11:33:46,363 INFO [finetune.py:976] (2/7) Epoch 1, batch 5550, loss[loss=0.3466, simple_loss=0.3845, pruned_loss=0.1544, over 4822.00 frames. ], tot_loss[loss=0.2996, simple_loss=0.3361, pruned_loss=0.1315, over 956104.27 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:34:21,069 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.67 vs. limit=2.0 +2023-04-26 11:34:53,183 INFO [finetune.py:976] (2/7) Epoch 1, batch 5600, loss[loss=0.3814, simple_loss=0.4063, pruned_loss=0.1782, over 4172.00 frames. ], tot_loss[loss=0.3037, simple_loss=0.3414, pruned_loss=0.133, over 955863.11 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:35:13,911 INFO [optim.py:369] (2/7) 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 11:35:24,472 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7959, 2.7132, 1.5622, 1.7620, 1.2485, 1.3390, 1.7486, 1.1683], + device='cuda:2'), covar=tensor([0.2845, 0.2512, 0.3434, 0.3908, 0.4533, 0.3461, 0.2631, 0.3740], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0222, 0.0200, 0.0218, 0.0237, 0.0198, 0.0194, 0.0210], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:35:42,019 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5955, 1.3021, 1.1704, 1.2289, 1.8442, 1.6057, 1.3690, 1.2123], + device='cuda:2'), covar=tensor([0.1608, 0.1899, 0.2204, 0.2052, 0.0901, 0.2118, 0.2207, 0.1895], + device='cuda:2'), in_proj_covar=tensor([0.0334, 0.0347, 0.0350, 0.0318, 0.0358, 0.0383, 0.0330, 0.0364], + device='cuda:2'), out_proj_covar=tensor([7.3152e-05, 7.4860e-05, 7.5698e-05, 6.6611e-05, 7.6308e-05, 8.3819e-05, + 7.1860e-05, 7.8766e-05], device='cuda:2') +2023-04-26 11:35:42,702 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 11:35:44,926 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5650.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:35:45,418 INFO [finetune.py:976] (2/7) Epoch 1, batch 5650, loss[loss=0.2885, simple_loss=0.3389, pruned_loss=0.1191, over 4898.00 frames. ], tot_loss[loss=0.3068, simple_loss=0.3448, pruned_loss=0.1344, over 956731.57 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:36:10,552 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9564, 2.6098, 1.7725, 1.7124, 1.4554, 1.5541, 1.8102, 1.3861], + device='cuda:2'), covar=tensor([0.2246, 0.2196, 0.3012, 0.3343, 0.3692, 0.2788, 0.2100, 0.3098], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0222, 0.0199, 0.0217, 0.0236, 0.0198, 0.0193, 0.0210], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:36:11,273 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 11:36:15,235 INFO [finetune.py:976] (2/7) Epoch 1, batch 5700, loss[loss=0.2956, simple_loss=0.3136, pruned_loss=0.1388, over 4312.00 frames. ], tot_loss[loss=0.3016, simple_loss=0.3381, pruned_loss=0.1325, over 936349.11 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:36:21,270 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5711.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:36:31,942 INFO [optim.py:369] (2/7) 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 11:37:01,400 INFO [finetune.py:976] (2/7) Epoch 2, batch 0, loss[loss=0.2519, simple_loss=0.2996, pruned_loss=0.1021, over 4807.00 frames. ], tot_loss[loss=0.2519, simple_loss=0.2996, pruned_loss=0.1021, over 4807.00 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:37:01,400 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 11:37:23,940 INFO [finetune.py:1010] (2/7) Epoch 2, validation: loss=0.2101, simple_loss=0.2777, pruned_loss=0.0712, over 2265189.00 frames. +2023-04-26 11:37:23,941 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6069MB +2023-04-26 11:37:46,572 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5755.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:37:53,836 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2812, 2.9638, 0.9395, 1.5923, 2.1536, 1.4346, 4.1334, 1.7834], + device='cuda:2'), covar=tensor([0.0707, 0.0821, 0.1071, 0.1293, 0.0597, 0.1066, 0.0210, 0.0700], + device='cuda:2'), in_proj_covar=tensor([0.0057, 0.0073, 0.0054, 0.0051, 0.0056, 0.0056, 0.0088, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0012, 0.0008], + device='cuda:2') +2023-04-26 11:38:02,075 INFO [finetune.py:976] (2/7) Epoch 2, batch 50, loss[loss=0.316, simple_loss=0.3622, pruned_loss=0.1349, over 4798.00 frames. ], tot_loss[loss=0.3052, simple_loss=0.3439, pruned_loss=0.1332, over 217226.50 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:38:27,676 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5816.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:38:29,284 INFO [optim.py:369] (2/7) 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 11:38:35,780 INFO [finetune.py:976] (2/7) Epoch 2, batch 100, loss[loss=0.266, simple_loss=0.3145, pruned_loss=0.1087, over 4827.00 frames. ], tot_loss[loss=0.2962, simple_loss=0.334, pruned_loss=0.1292, over 381988.38 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:38:50,938 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 11:38:55,148 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5858.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:39:09,186 INFO [finetune.py:976] (2/7) Epoch 2, batch 150, loss[loss=0.3263, simple_loss=0.3494, pruned_loss=0.1516, over 4817.00 frames. ], tot_loss[loss=0.2875, simple_loss=0.3262, pruned_loss=0.1244, over 509567.82 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:39:18,938 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7355, 1.3870, 1.2455, 1.3456, 2.0561, 1.7260, 1.4022, 1.2803], + device='cuda:2'), covar=tensor([0.1440, 0.1605, 0.2422, 0.2012, 0.0704, 0.1641, 0.2000, 0.2063], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0344, 0.0348, 0.0316, 0.0355, 0.0379, 0.0327, 0.0362], + device='cuda:2'), out_proj_covar=tensor([7.2488e-05, 7.4208e-05, 7.5277e-05, 6.6285e-05, 7.5643e-05, 8.3080e-05, + 7.1362e-05, 7.8175e-05], device='cuda:2') +2023-04-26 11:39:31,958 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6580, 1.8734, 1.1098, 1.3440, 2.1226, 1.5418, 1.4237, 1.5494], + device='cuda:2'), covar=tensor([0.0628, 0.0468, 0.0468, 0.0647, 0.0328, 0.0626, 0.0623, 0.0769], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0043, 0.0038, 0.0049, 0.0037, 0.0047, 0.0046, 0.0051], + device='cuda:2') +2023-04-26 11:39:34,932 INFO [optim.py:369] (2/7) 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 11:39:35,686 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5919.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:39:39,728 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9796, 2.5010, 1.9914, 2.3376, 1.9317, 2.0143, 2.4272, 1.7790], + device='cuda:2'), covar=tensor([0.2195, 0.1256, 0.1211, 0.1405, 0.2363, 0.1255, 0.1776, 0.2476], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0330, 0.0241, 0.0304, 0.0309, 0.0280, 0.0276, 0.0297], + device='cuda:2'), out_proj_covar=tensor([1.2622e-04, 1.3545e-04, 9.8840e-05, 1.2343e-04, 1.2772e-04, 1.1320e-04, + 1.1445e-04, 1.2111e-04], device='cuda:2') +2023-04-26 11:39:48,162 INFO [finetune.py:976] (2/7) Epoch 2, batch 200, loss[loss=0.3378, simple_loss=0.3635, pruned_loss=0.1561, over 4710.00 frames. ], tot_loss[loss=0.284, simple_loss=0.3232, pruned_loss=0.1224, over 608202.64 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:40:52,013 INFO [finetune.py:976] (2/7) Epoch 2, batch 250, loss[loss=0.2718, simple_loss=0.3381, pruned_loss=0.1028, over 4822.00 frames. ], tot_loss[loss=0.2888, simple_loss=0.3278, pruned_loss=0.1249, over 686227.83 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:41:06,884 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5990.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:41:07,032 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-26 11:41:09,893 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 11:41:18,234 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6006.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:41:25,494 INFO [optim.py:369] (2/7) 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 11:41:33,099 INFO [finetune.py:976] (2/7) Epoch 2, batch 300, loss[loss=0.3127, simple_loss=0.351, pruned_loss=0.1372, over 4893.00 frames. ], tot_loss[loss=0.2932, simple_loss=0.3322, pruned_loss=0.1271, over 745981.47 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:41:35,033 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1607, 1.7812, 1.5386, 2.1205, 2.0642, 2.1904, 1.6471, 4.4387], + device='cuda:2'), covar=tensor([0.0746, 0.0807, 0.0855, 0.1249, 0.0671, 0.0627, 0.0774, 0.0152], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0047, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 11:41:49,305 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6043.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:41:54,625 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6051.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:42:22,355 INFO [finetune.py:976] (2/7) Epoch 2, batch 350, loss[loss=0.2623, simple_loss=0.3013, pruned_loss=0.1117, over 4203.00 frames. ], tot_loss[loss=0.2954, simple_loss=0.3349, pruned_loss=0.128, over 789764.40 frames. ], batch size: 17, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:42:47,488 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6111.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:43:02,075 INFO [optim.py:369] (2/7) 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 11:43:14,258 INFO [finetune.py:976] (2/7) Epoch 2, batch 400, loss[loss=0.2854, simple_loss=0.3341, pruned_loss=0.1184, over 4919.00 frames. ], tot_loss[loss=0.2958, simple_loss=0.3366, pruned_loss=0.1275, over 825265.91 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-04-26 11:43:20,191 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6135.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:43:33,298 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7556, 1.2573, 1.2865, 1.2326, 1.9049, 1.6630, 1.3771, 1.2869], + device='cuda:2'), covar=tensor([0.1171, 0.1710, 0.2146, 0.2038, 0.0772, 0.1758, 0.2105, 0.1968], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0344, 0.0347, 0.0316, 0.0354, 0.0379, 0.0326, 0.0359], + device='cuda:2'), out_proj_covar=tensor([7.1799e-05, 7.4079e-05, 7.4900e-05, 6.6232e-05, 7.5364e-05, 8.2942e-05, + 7.1097e-05, 7.7679e-05], device='cuda:2') +2023-04-26 11:43:43,651 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6170.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:43:48,407 INFO [finetune.py:976] (2/7) Epoch 2, batch 450, loss[loss=0.2592, simple_loss=0.306, pruned_loss=0.1062, over 4933.00 frames. ], tot_loss[loss=0.2934, simple_loss=0.3344, pruned_loss=0.1262, over 855263.72 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:44:00,948 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6196.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:06,759 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6203.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:11,620 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6214.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:13,493 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8491, 1.8633, 2.1667, 2.3028, 2.3300, 1.6585, 1.2940, 1.9007], + device='cuda:2'), covar=tensor([0.1349, 0.1204, 0.0699, 0.0854, 0.0707, 0.1352, 0.1497, 0.0857], + device='cuda:2'), in_proj_covar=tensor([0.0210, 0.0211, 0.0190, 0.0182, 0.0179, 0.0197, 0.0176, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:44:15,816 INFO [optim.py:369] (2/7) 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 11:44:21,880 INFO [finetune.py:976] (2/7) Epoch 2, batch 500, loss[loss=0.2685, simple_loss=0.3074, pruned_loss=0.1148, over 4907.00 frames. ], tot_loss[loss=0.2903, simple_loss=0.3311, pruned_loss=0.1247, over 879612.69 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:44:23,803 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6231.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:23,942 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-04-26 11:44:43,179 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3703, 2.3374, 2.7941, 2.9754, 2.7830, 2.1180, 1.7597, 2.4237], + device='cuda:2'), covar=tensor([0.1350, 0.1079, 0.0623, 0.0856, 0.0740, 0.1285, 0.1414, 0.0851], + device='cuda:2'), in_proj_covar=tensor([0.0210, 0.0211, 0.0190, 0.0182, 0.0179, 0.0197, 0.0176, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:44:46,842 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6264.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:48,030 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6266.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:51,698 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6272.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:44:55,245 INFO [finetune.py:976] (2/7) Epoch 2, batch 550, loss[loss=0.3213, simple_loss=0.3476, pruned_loss=0.1474, over 4832.00 frames. ], tot_loss[loss=0.2876, simple_loss=0.3279, pruned_loss=0.1237, over 894625.06 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:44:59,710 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0930, 1.1743, 1.2423, 1.3118, 1.3814, 1.4805, 1.3724, 1.3909], + device='cuda:2'), covar=tensor([3.0232, 7.4706, 5.1664, 4.3506, 4.7093, 7.6127, 6.4698, 5.5631], + device='cuda:2'), in_proj_covar=tensor([0.0266, 0.0369, 0.0296, 0.0297, 0.0325, 0.0347, 0.0357, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:45:15,010 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6306.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:45:28,235 INFO [optim.py:369] (2/7) 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 11:45:37,840 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 600, loss[loss=0.2777, simple_loss=0.332, pruned_loss=0.1117, over 4827.00 frames. ], tot_loss[loss=0.288, simple_loss=0.3277, pruned_loss=0.1242, over 908322.20 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:46:00,718 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6346.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:46:12,540 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6354.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:46:39,286 INFO [finetune.py:976] (2/7) Epoch 2, batch 650, loss[loss=0.2829, simple_loss=0.3214, pruned_loss=0.1222, over 4758.00 frames. ], tot_loss[loss=0.2924, simple_loss=0.3325, pruned_loss=0.1262, over 919542.28 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:46:53,632 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6399.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:02,622 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 11:47:13,364 INFO [finetune.py:976] (2/7) Epoch 2, batch 700, loss[loss=0.2511, simple_loss=0.2986, pruned_loss=0.1018, over 4744.00 frames. ], tot_loss[loss=0.293, simple_loss=0.3338, pruned_loss=0.1261, over 926787.15 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:47:30,803 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7520, 1.4735, 4.4180, 4.0818, 3.9691, 4.1877, 4.2260, 3.9475], + device='cuda:2'), covar=tensor([0.6539, 0.6115, 0.1159, 0.1916, 0.1020, 0.1640, 0.1060, 0.1491], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0318, 0.0455, 0.0463, 0.0381, 0.0441, 0.0348, 0.0404], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 11:47:46,401 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6459.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:48,748 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6462.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:51,736 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:47:58,777 INFO [finetune.py:976] (2/7) Epoch 2, batch 750, loss[loss=0.3334, simple_loss=0.3666, pruned_loss=0.1501, over 4834.00 frames. ], tot_loss[loss=0.293, simple_loss=0.3343, pruned_loss=0.1259, over 933855.47 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:48:18,132 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6491.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:48:50,921 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6514.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:48:53,698 INFO [optim.py:369] (2/7) 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 11:49:03,653 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6523.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:49:05,427 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 800, loss[loss=0.3236, simple_loss=0.358, pruned_loss=0.1446, over 4895.00 frames. ], tot_loss[loss=0.2916, simple_loss=0.3336, pruned_loss=0.1248, over 939101.11 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:49:06,736 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6528.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:27,196 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6559.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:29,547 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6562.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:49:33,147 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 850, loss[loss=0.2774, simple_loss=0.3164, pruned_loss=0.1192, over 4901.00 frames. ], tot_loss[loss=0.2877, simple_loss=0.3302, pruned_loss=0.1226, over 943324.58 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:50:18,614 INFO [optim.py:369] (2/7) 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 11:50:21,171 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6622.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:50:25,264 INFO [finetune.py:976] (2/7) Epoch 2, batch 900, loss[loss=0.2806, simple_loss=0.3247, pruned_loss=0.1183, over 4829.00 frames. ], tot_loss[loss=0.2845, simple_loss=0.3266, pruned_loss=0.1212, over 944911.02 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:50:25,977 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6629.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:50:36,404 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6646.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:50:37,614 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0590, 1.2639, 5.2507, 4.8754, 4.5271, 4.9631, 4.6484, 4.6259], + device='cuda:2'), covar=tensor([0.6693, 0.6938, 0.1187, 0.2128, 0.1173, 0.1536, 0.1282, 0.1624], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0317, 0.0452, 0.0461, 0.0379, 0.0439, 0.0347, 0.0403], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 11:50:42,585 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.90 vs. limit=5.0 +2023-04-26 11:50:58,406 INFO [finetune.py:976] (2/7) Epoch 2, batch 950, loss[loss=0.3218, simple_loss=0.3579, pruned_loss=0.1429, over 4919.00 frames. ], tot_loss[loss=0.2824, simple_loss=0.3238, pruned_loss=0.1205, over 947849.28 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:51:06,480 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6690.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:51:08,889 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6694.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:51:11,962 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 11:51:48,517 INFO [finetune.py:976] (2/7) Epoch 2, batch 1000, loss[loss=0.2265, simple_loss=0.2726, pruned_loss=0.09019, over 4765.00 frames. ], tot_loss[loss=0.2854, simple_loss=0.327, pruned_loss=0.1219, over 951109.17 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:52:11,813 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6747.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:52:12,486 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7039, 1.2673, 1.2358, 1.2606, 1.9790, 1.6421, 1.2617, 1.2011], + device='cuda:2'), covar=tensor([0.1499, 0.1818, 0.2044, 0.1951, 0.0791, 0.1726, 0.2352, 0.2045], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0343, 0.0346, 0.0315, 0.0353, 0.0377, 0.0326, 0.0359], + device='cuda:2'), out_proj_covar=tensor([7.1845e-05, 7.3783e-05, 7.4694e-05, 6.6087e-05, 7.5231e-05, 8.2557e-05, + 7.1018e-05, 7.7436e-05], device='cuda:2') +2023-04-26 11:52:55,093 INFO [finetune.py:976] (2/7) Epoch 2, batch 1050, loss[loss=0.3025, simple_loss=0.342, pruned_loss=0.1315, over 4895.00 frames. ], tot_loss[loss=0.2891, simple_loss=0.3308, pruned_loss=0.1237, over 953597.18 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:53:00,706 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4655, 1.8407, 1.2382, 1.0662, 1.1024, 1.0965, 1.2387, 1.0281], + device='cuda:2'), covar=tensor([0.2426, 0.2301, 0.3056, 0.3528, 0.4202, 0.3000, 0.2365, 0.3309], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0224, 0.0199, 0.0219, 0.0237, 0.0199, 0.0194, 0.0211], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:53:03,724 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6791.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:53:07,963 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9484, 1.2444, 4.7878, 4.4537, 4.2509, 4.5135, 4.3405, 4.1554], + device='cuda:2'), covar=tensor([0.6701, 0.6230, 0.1034, 0.1799, 0.1076, 0.0974, 0.1633, 0.1440], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0317, 0.0455, 0.0461, 0.0380, 0.0439, 0.0347, 0.0404], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 11:53:33,223 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6818.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 11:53:42,745 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6823.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:53:44,639 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6826.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:53:51,480 INFO [finetune.py:976] (2/7) Epoch 2, batch 1100, loss[loss=0.3276, simple_loss=0.3562, pruned_loss=0.1495, over 4816.00 frames. ], tot_loss[loss=0.2897, simple_loss=0.3322, pruned_loss=0.1236, over 954517.65 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:53:51,636 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9387, 0.6773, 0.9139, 1.2307, 1.1976, 0.9039, 0.9134, 0.9134], + device='cuda:2'), covar=tensor([ 5.7404, 9.0538, 9.7819, 9.9988, 6.4019, 10.3824, 10.7234, 7.1247], + device='cuda:2'), in_proj_covar=tensor([0.0439, 0.0504, 0.0585, 0.0569, 0.0469, 0.0525, 0.0532, 0.0541], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:54:04,439 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6839.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:16,584 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:22,454 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6867.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:27,636 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6874.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:30,498 INFO [finetune.py:976] (2/7) Epoch 2, batch 1150, loss[loss=0.293, simple_loss=0.3377, pruned_loss=0.1242, over 4814.00 frames. ], tot_loss[loss=0.2889, simple_loss=0.3318, pruned_loss=0.123, over 955961.67 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:54:30,617 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6878.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:54:40,720 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0073, 1.2864, 1.2500, 1.2687, 1.3304, 1.3735, 1.4115, 1.3594], + device='cuda:2'), covar=tensor([3.3096, 6.8662, 4.8041, 4.2426, 4.6440, 7.8258, 6.6325, 5.6537], + device='cuda:2'), in_proj_covar=tensor([0.0271, 0.0373, 0.0299, 0.0300, 0.0328, 0.0354, 0.0361, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:54:42,046 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 11:54:45,646 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9264, 1.0921, 1.1673, 1.2227, 1.2876, 1.3479, 1.2542, 1.2916], + device='cuda:2'), covar=tensor([3.3141, 6.9405, 4.8389, 4.0073, 4.6224, 7.4129, 6.2928, 5.0248], + device='cuda:2'), in_proj_covar=tensor([0.0271, 0.0374, 0.0300, 0.0300, 0.0329, 0.0354, 0.0361, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:54:48,160 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 11:54:49,251 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6907.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:50,553 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6909.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:54,164 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6915.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:54:56,913 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6922.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:55:04,516 INFO [finetune.py:976] (2/7) Epoch 2, batch 1200, loss[loss=0.2274, simple_loss=0.2785, pruned_loss=0.08816, over 4804.00 frames. ], tot_loss[loss=0.2867, simple_loss=0.3293, pruned_loss=0.1221, over 955726.45 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:55:11,751 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6939.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:55:22,790 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 11:55:31,685 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:55:31,757 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:55:38,016 INFO [finetune.py:976] (2/7) Epoch 2, batch 1250, loss[loss=0.2794, simple_loss=0.3194, pruned_loss=0.1198, over 4853.00 frames. ], tot_loss[loss=0.2835, simple_loss=0.3256, pruned_loss=0.1207, over 955203.87 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 32.0 +2023-04-26 11:55:43,399 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6985.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:55:52,573 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9183, 1.1798, 1.6225, 2.1157, 1.5230, 1.2196, 0.9783, 1.4374], + device='cuda:2'), covar=tensor([0.6589, 0.8474, 0.3838, 0.8268, 0.8949, 0.6310, 1.1873, 0.8662], + device='cuda:2'), in_proj_covar=tensor([0.0263, 0.0278, 0.0221, 0.0345, 0.0236, 0.0234, 0.0273, 0.0222], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:56:04,204 INFO [optim.py:369] (2/7) 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 11:56:07,965 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-26 11:56:11,779 INFO [finetune.py:976] (2/7) Epoch 2, batch 1300, loss[loss=0.217, simple_loss=0.2754, pruned_loss=0.07935, over 4911.00 frames. ], tot_loss[loss=0.2791, simple_loss=0.3214, pruned_loss=0.1184, over 955956.31 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:56:25,540 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7048.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:56:26,800 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9814, 0.4834, 0.7011, 1.2998, 1.2039, 0.9282, 0.8649, 0.8276], + device='cuda:2'), covar=tensor([ 5.7424, 8.2581, 8.9376, 10.4212, 6.7822, 9.2762, 8.5072, 6.3081], + device='cuda:2'), in_proj_covar=tensor([0.0443, 0.0506, 0.0590, 0.0575, 0.0473, 0.0529, 0.0535, 0.0545], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:56:55,868 INFO [finetune.py:976] (2/7) Epoch 2, batch 1350, loss[loss=0.2923, simple_loss=0.3559, pruned_loss=0.1144, over 4805.00 frames. ], tot_loss[loss=0.2783, simple_loss=0.3209, pruned_loss=0.1179, over 954928.93 frames. ], batch size: 41, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:56:56,574 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7075, 4.0504, 0.9739, 2.1272, 2.3081, 2.5959, 2.6672, 0.9837], + device='cuda:2'), covar=tensor([0.1369, 0.0919, 0.2058, 0.1317, 0.1013, 0.1149, 0.1306, 0.2320], + device='cuda:2'), in_proj_covar=tensor([0.0124, 0.0272, 0.0151, 0.0133, 0.0144, 0.0165, 0.0130, 0.0133], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 11:57:34,487 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7109.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 11:57:39,840 INFO [optim.py:369] (2/7) 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 11:57:39,967 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7118.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:57:48,844 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 1400, loss[loss=0.2481, simple_loss=0.3121, pruned_loss=0.09202, over 4805.00 frames. ], tot_loss[loss=0.282, simple_loss=0.3252, pruned_loss=0.1194, over 956096.67 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:58:36,325 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7163.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:58:38,105 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7166.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:58:47,202 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7171.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:58:51,417 INFO [finetune.py:976] (2/7) Epoch 2, batch 1450, loss[loss=0.2407, simple_loss=0.2877, pruned_loss=0.09685, over 4757.00 frames. ], tot_loss[loss=0.2821, simple_loss=0.3262, pruned_loss=0.119, over 954360.88 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:59:19,360 INFO [optim.py:369] (2/7) 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 11:59:23,166 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7224.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 11:59:25,501 INFO [finetune.py:976] (2/7) Epoch 2, batch 1500, loss[loss=0.2926, simple_loss=0.3379, pruned_loss=0.1236, over 4820.00 frames. ], tot_loss[loss=0.2842, simple_loss=0.3282, pruned_loss=0.1201, over 955634.82 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 11:59:26,852 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0299, 1.2929, 1.6084, 2.3170, 1.6175, 1.2739, 0.9609, 1.5848], + device='cuda:2'), covar=tensor([0.7109, 0.8672, 0.4129, 0.8615, 0.9423, 0.6469, 1.2760, 0.9559], + device='cuda:2'), in_proj_covar=tensor([0.0264, 0.0279, 0.0221, 0.0346, 0.0237, 0.0234, 0.0273, 0.0222], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 11:59:29,695 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7234.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 11:59:35,466 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0175, 0.6763, 0.8817, 1.2999, 1.2084, 0.9280, 0.9312, 0.8924], + device='cuda:2'), covar=tensor([5.2053, 7.7733, 8.0018, 8.9606, 5.5635, 9.5503, 8.8493, 6.5243], + device='cuda:2'), in_proj_covar=tensor([0.0445, 0.0509, 0.0593, 0.0579, 0.0475, 0.0531, 0.0538, 0.0548], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 11:59:51,524 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 11:59:57,730 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 1550, loss[loss=0.2696, simple_loss=0.3258, pruned_loss=0.1067, over 4865.00 frames. ], tot_loss[loss=0.2843, simple_loss=0.3284, pruned_loss=0.12, over 953272.81 frames. ], batch size: 34, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:00:04,502 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7285.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:00:11,042 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0319, 1.0866, 1.2130, 1.2975, 1.3926, 1.4819, 1.2373, 1.2927], + device='cuda:2'), covar=tensor([2.4097, 5.5980, 4.1501, 3.6143, 3.8847, 6.1675, 5.5163, 4.6371], + device='cuda:2'), in_proj_covar=tensor([0.0272, 0.0374, 0.0299, 0.0300, 0.0329, 0.0357, 0.0361, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:00:27,652 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 2, batch 1600, loss[loss=0.2594, simple_loss=0.3022, pruned_loss=0.1083, over 4859.00 frames. ], tot_loss[loss=0.2797, simple_loss=0.324, pruned_loss=0.1177, over 953281.68 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:00:36,973 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7333.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:00:38,910 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7336.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:01:07,695 INFO [finetune.py:976] (2/7) Epoch 2, batch 1650, loss[loss=0.2432, simple_loss=0.2829, pruned_loss=0.1018, over 4903.00 frames. ], tot_loss[loss=0.2762, simple_loss=0.3203, pruned_loss=0.116, over 954976.98 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:01:25,935 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7404.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 12:01:28,254 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7407.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:01:34,835 INFO [optim.py:369] (2/7) 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 12:01:37,976 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-26 12:01:40,986 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1156, 0.9351, 1.1182, 1.1046, 1.0208, 0.8151, 0.8006, 0.3683], + device='cuda:2'), covar=tensor([0.0541, 0.0821, 0.0793, 0.0646, 0.0752, 0.1476, 0.0756, 0.1569], + device='cuda:2'), in_proj_covar=tensor([0.0072, 0.0080, 0.0076, 0.0075, 0.0088, 0.0095, 0.0090, 0.0081], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 12:01:41,458 INFO [finetune.py:976] (2/7) Epoch 2, batch 1700, loss[loss=0.3266, simple_loss=0.3418, pruned_loss=0.1557, over 4281.00 frames. ], tot_loss[loss=0.2737, simple_loss=0.3173, pruned_loss=0.1151, over 954730.85 frames. ], batch size: 18, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:02:03,139 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-04-26 12:02:09,795 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7468.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:02:22,155 INFO [finetune.py:976] (2/7) Epoch 2, batch 1750, loss[loss=0.3233, simple_loss=0.375, pruned_loss=0.1358, over 4808.00 frames. ], tot_loss[loss=0.2755, simple_loss=0.3192, pruned_loss=0.1159, over 954219.84 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:03:11,384 INFO [optim.py:369] (2/7) 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 12:03:12,084 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7519.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:03:23,644 INFO [finetune.py:976] (2/7) Epoch 2, batch 1800, loss[loss=0.2714, simple_loss=0.3294, pruned_loss=0.1067, over 4934.00 frames. ], tot_loss[loss=0.2788, simple_loss=0.3236, pruned_loss=0.117, over 956012.86 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:03:33,045 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7534.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:03:45,103 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2358, 3.0122, 1.0552, 1.4164, 1.9784, 1.3311, 3.8334, 1.8573], + device='cuda:2'), covar=tensor([0.0729, 0.0835, 0.1154, 0.1265, 0.0646, 0.1079, 0.0181, 0.0671], + device='cuda:2'), in_proj_covar=tensor([0.0057, 0.0073, 0.0054, 0.0050, 0.0056, 0.0056, 0.0087, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 12:04:17,663 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:04:31,079 INFO [finetune.py:976] (2/7) Epoch 2, batch 1850, loss[loss=0.3175, simple_loss=0.3443, pruned_loss=0.1453, over 4878.00 frames. ], tot_loss[loss=0.2819, simple_loss=0.3258, pruned_loss=0.1189, over 954040.26 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:04:34,064 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7582.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:05:01,019 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-26 12:05:07,099 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:05:14,346 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 1900, loss[loss=0.3243, simple_loss=0.3503, pruned_loss=0.1491, over 4299.00 frames. ], tot_loss[loss=0.2808, simple_loss=0.3253, pruned_loss=0.1181, over 952906.58 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:05:18,630 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7631.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:05:39,443 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8390, 2.1737, 1.2264, 1.4983, 2.2758, 1.8283, 1.6472, 1.8211], + device='cuda:2'), covar=tensor([0.0555, 0.0407, 0.0416, 0.0626, 0.0296, 0.0557, 0.0545, 0.0696], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0033], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0044, 0.0039, 0.0049, 0.0037, 0.0048, 0.0047, 0.0051], + device='cuda:2') +2023-04-26 12:05:50,037 INFO [finetune.py:976] (2/7) Epoch 2, batch 1950, loss[loss=0.2648, simple_loss=0.3052, pruned_loss=0.1123, over 4774.00 frames. ], tot_loss[loss=0.2777, simple_loss=0.3229, pruned_loss=0.1163, over 953775.99 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:05:53,289 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-04-26 12:05:54,914 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7685.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:06:06,523 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7704.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:06:12,240 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4701, 1.4866, 4.1760, 3.8744, 3.7146, 3.9473, 3.9646, 3.7134], + device='cuda:2'), covar=tensor([0.6406, 0.5271, 0.0986, 0.1697, 0.1054, 0.1479, 0.1228, 0.1500], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0315, 0.0449, 0.0458, 0.0379, 0.0433, 0.0344, 0.0400], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 12:06:13,997 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8882, 2.4047, 1.9651, 2.3196, 1.7319, 1.9469, 2.0276, 1.6604], + device='cuda:2'), covar=tensor([0.2048, 0.1396, 0.1119, 0.1349, 0.3187, 0.1424, 0.2038, 0.2754], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0334, 0.0245, 0.0310, 0.0313, 0.0284, 0.0279, 0.0301], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:06:17,408 INFO [optim.py:369] (2/7) 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 12:06:23,982 INFO [finetune.py:976] (2/7) Epoch 2, batch 2000, loss[loss=0.2864, simple_loss=0.321, pruned_loss=0.1259, over 4816.00 frames. ], tot_loss[loss=0.2738, simple_loss=0.3187, pruned_loss=0.1145, over 954872.02 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:06:33,719 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7743.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:06:36,708 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9015, 2.8829, 2.3599, 3.3063, 2.9440, 2.9093, 1.2235, 2.8451], + device='cuda:2'), covar=tensor([0.2217, 0.1577, 0.3072, 0.2611, 0.2993, 0.2099, 0.5935, 0.2680], + device='cuda:2'), in_proj_covar=tensor([0.0260, 0.0232, 0.0278, 0.0328, 0.0323, 0.0272, 0.0286, 0.0287], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:06:39,111 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7752.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:06:47,290 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:06:57,780 INFO [finetune.py:976] (2/7) Epoch 2, batch 2050, loss[loss=0.1905, simple_loss=0.2525, pruned_loss=0.06428, over 4818.00 frames. ], tot_loss[loss=0.2694, simple_loss=0.3141, pruned_loss=0.1124, over 956405.51 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:07:13,089 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7802.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:07:14,335 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7804.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:07:24,635 INFO [optim.py:369] (2/7) 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 12:07:25,367 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7819.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:07:31,147 INFO [finetune.py:976] (2/7) Epoch 2, batch 2100, loss[loss=0.2683, simple_loss=0.3018, pruned_loss=0.1174, over 4813.00 frames. ], tot_loss[loss=0.2687, simple_loss=0.3131, pruned_loss=0.1122, over 957096.24 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:07:54,050 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7863.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:07:56,932 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7867.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:08:15,063 INFO [finetune.py:976] (2/7) Epoch 2, batch 2150, loss[loss=0.2456, simple_loss=0.2925, pruned_loss=0.09934, over 4744.00 frames. ], tot_loss[loss=0.2712, simple_loss=0.316, pruned_loss=0.1132, over 955623.91 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:08:50,097 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2438, 1.4994, 1.4662, 1.8870, 2.2301, 1.8411, 1.7058, 1.6003], + device='cuda:2'), covar=tensor([0.1756, 0.2288, 0.2038, 0.2406, 0.1519, 0.2119, 0.2812, 0.2054], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0341, 0.0346, 0.0317, 0.0354, 0.0373, 0.0325, 0.0357], + device='cuda:2'), out_proj_covar=tensor([7.1584e-05, 7.3420e-05, 7.4763e-05, 6.6427e-05, 7.5492e-05, 8.1715e-05, + 7.0861e-05, 7.7152e-05], device='cuda:2') +2023-04-26 12:09:02,670 INFO [optim.py:369] (2/7) 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 12:09:13,241 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-26 12:09:20,978 INFO [finetune.py:976] (2/7) Epoch 2, batch 2200, loss[loss=0.3062, simple_loss=0.3399, pruned_loss=0.1363, over 4823.00 frames. ], tot_loss[loss=0.2736, simple_loss=0.319, pruned_loss=0.1141, over 955431.79 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:09:22,952 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 2250, loss[loss=0.3085, simple_loss=0.3555, pruned_loss=0.1307, over 4895.00 frames. ], tot_loss[loss=0.2746, simple_loss=0.3203, pruned_loss=0.1145, over 952648.91 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:10:22,019 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7979.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:10:23,124 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7980.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:10:47,701 INFO [optim.py:369] (2/7) 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 12:10:55,296 INFO [finetune.py:976] (2/7) Epoch 2, batch 2300, loss[loss=0.2675, simple_loss=0.3094, pruned_loss=0.1128, over 4778.00 frames. ], tot_loss[loss=0.2752, simple_loss=0.3212, pruned_loss=0.1146, over 953788.82 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:11:18,728 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8063.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:11:28,370 INFO [finetune.py:976] (2/7) Epoch 2, batch 2350, loss[loss=0.2569, simple_loss=0.31, pruned_loss=0.1019, over 4903.00 frames. ], tot_loss[loss=0.2705, simple_loss=0.3171, pruned_loss=0.112, over 952364.02 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:11:42,664 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8099.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:11:45,176 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9961, 2.0760, 1.9512, 1.8121, 2.1520, 1.6160, 2.8477, 1.4102], + device='cuda:2'), covar=tensor([0.3970, 0.1579, 0.4282, 0.2877, 0.1828, 0.2716, 0.1130, 0.4517], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0351, 0.0434, 0.0368, 0.0403, 0.0373, 0.0398, 0.0408], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:11:50,393 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8111.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:11:53,612 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 12:11:54,656 INFO [optim.py:369] (2/7) 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 12:12:01,654 INFO [finetune.py:976] (2/7) Epoch 2, batch 2400, loss[loss=0.2278, simple_loss=0.2738, pruned_loss=0.09088, over 4827.00 frames. ], tot_loss[loss=0.267, simple_loss=0.3136, pruned_loss=0.1102, over 955108.18 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:12:21,849 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 2450, loss[loss=0.2752, simple_loss=0.3152, pruned_loss=0.1176, over 4823.00 frames. ], tot_loss[loss=0.2636, simple_loss=0.3099, pruned_loss=0.1086, over 957163.82 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:13:01,853 INFO [optim.py:369] (2/7) 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 12:13:07,516 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1488, 0.9438, 1.1304, 1.4459, 1.4228, 1.1506, 1.1642, 1.0784], + device='cuda:2'), covar=tensor([3.3160, 4.8691, 5.2566, 5.7376, 3.5389, 5.9851, 6.0734, 4.2382], + device='cuda:2'), in_proj_covar=tensor([0.0445, 0.0509, 0.0594, 0.0584, 0.0477, 0.0527, 0.0536, 0.0547], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:13:07,998 INFO [finetune.py:976] (2/7) Epoch 2, batch 2500, loss[loss=0.319, simple_loss=0.3538, pruned_loss=0.1421, over 4171.00 frames. ], tot_loss[loss=0.266, simple_loss=0.3122, pruned_loss=0.1099, over 955147.66 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:13:38,236 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1400, 2.7035, 1.1253, 1.3801, 2.1093, 1.3482, 3.3870, 1.6808], + device='cuda:2'), covar=tensor([0.0691, 0.0813, 0.0859, 0.1239, 0.0499, 0.0992, 0.0193, 0.0689], + device='cuda:2'), in_proj_covar=tensor([0.0057, 0.0074, 0.0055, 0.0051, 0.0056, 0.0057, 0.0087, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 12:13:49,850 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:14:03,522 INFO [finetune.py:976] (2/7) Epoch 2, batch 2550, loss[loss=0.2228, simple_loss=0.2509, pruned_loss=0.0974, over 4019.00 frames. ], tot_loss[loss=0.2694, simple_loss=0.3163, pruned_loss=0.1113, over 951302.01 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:14:04,880 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8280.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:14:35,119 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4610, 3.6418, 0.9578, 1.8598, 1.9179, 2.5240, 2.2243, 0.9709], + device='cuda:2'), covar=tensor([0.1491, 0.1038, 0.2214, 0.1438, 0.1130, 0.1134, 0.1510, 0.2065], + device='cuda:2'), in_proj_covar=tensor([0.0125, 0.0275, 0.0152, 0.0133, 0.0145, 0.0167, 0.0130, 0.0135], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 12:15:02,137 INFO [optim.py:369] (2/7) 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 12:15:13,570 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8326.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:15:20,189 INFO [finetune.py:976] (2/7) Epoch 2, batch 2600, loss[loss=0.3167, simple_loss=0.3478, pruned_loss=0.1428, over 4144.00 frames. ], tot_loss[loss=0.2711, simple_loss=0.3185, pruned_loss=0.1118, over 951360.62 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:15:20,255 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8328.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:15:44,966 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1235, 1.5151, 1.3715, 1.7253, 1.6015, 2.0313, 1.4143, 3.3956], + device='cuda:2'), covar=tensor([0.0766, 0.0766, 0.0798, 0.1277, 0.0651, 0.0571, 0.0772, 0.0154], + device='cuda:2'), in_proj_covar=tensor([0.0041, 0.0041, 0.0042, 0.0047, 0.0042, 0.0041, 0.0041, 0.0066], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 12:16:16,234 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 12:16:23,382 INFO [finetune.py:976] (2/7) Epoch 2, batch 2650, loss[loss=0.3093, simple_loss=0.3487, pruned_loss=0.135, over 4809.00 frames. ], tot_loss[loss=0.2711, simple_loss=0.3184, pruned_loss=0.1119, over 950792.52 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:16:45,818 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-26 12:16:49,114 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8399.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:17:13,448 INFO [optim.py:369] (2/7) 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 12:17:19,410 INFO [finetune.py:976] (2/7) Epoch 2, batch 2700, loss[loss=0.2735, simple_loss=0.3146, pruned_loss=0.1162, over 4742.00 frames. ], tot_loss[loss=0.2674, simple_loss=0.3158, pruned_loss=0.1095, over 950028.84 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:17:26,366 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-26 12:17:30,319 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1185, 0.5232, 0.7867, 1.3926, 1.3185, 0.9911, 0.9423, 0.9261], + device='cuda:2'), covar=tensor([4.1332, 6.1995, 6.9152, 7.0274, 4.4812, 7.1120, 7.0187, 4.8353], + device='cuda:2'), in_proj_covar=tensor([0.0445, 0.0509, 0.0595, 0.0586, 0.0478, 0.0527, 0.0535, 0.0547], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:17:32,539 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8447.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:17:40,752 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:17:53,352 INFO [finetune.py:976] (2/7) Epoch 2, batch 2750, loss[loss=0.2411, simple_loss=0.2953, pruned_loss=0.09339, over 4796.00 frames. ], tot_loss[loss=0.2642, simple_loss=0.3126, pruned_loss=0.1079, over 952515.74 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:18:11,238 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-26 12:18:12,775 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:18:21,016 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 2, batch 2800, loss[loss=0.1971, simple_loss=0.2513, pruned_loss=0.0714, over 4801.00 frames. ], tot_loss[loss=0.2606, simple_loss=0.3084, pruned_loss=0.1065, over 952471.05 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:18:26,649 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1224, 1.4765, 1.3432, 1.7917, 1.5539, 1.9331, 1.3991, 3.6857], + device='cuda:2'), covar=tensor([0.0887, 0.1102, 0.1048, 0.1536, 0.0913, 0.0744, 0.1081, 0.0210], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0066], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 12:19:00,279 INFO [finetune.py:976] (2/7) Epoch 2, batch 2850, loss[loss=0.2483, simple_loss=0.3127, pruned_loss=0.09193, over 4860.00 frames. ], tot_loss[loss=0.2606, simple_loss=0.3078, pruned_loss=0.1067, over 952752.11 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:19:05,694 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1990, 1.4231, 1.5002, 1.5187, 1.5460, 1.1771, 0.7248, 1.3684], + device='cuda:2'), covar=tensor([0.1202, 0.1413, 0.0860, 0.0841, 0.0793, 0.1146, 0.1427, 0.0758], + device='cuda:2'), in_proj_covar=tensor([0.0211, 0.0210, 0.0190, 0.0183, 0.0180, 0.0199, 0.0176, 0.0194], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:19:45,299 INFO [optim.py:369] (2/7) 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 12:19:52,104 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 2900, loss[loss=0.2496, simple_loss=0.314, pruned_loss=0.09257, over 4849.00 frames. ], tot_loss[loss=0.2628, simple_loss=0.3107, pruned_loss=0.1075, over 954991.73 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:20:03,732 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6707, 2.3895, 1.5880, 1.4308, 1.2178, 1.2920, 1.5850, 1.1649], + device='cuda:2'), covar=tensor([0.2490, 0.2122, 0.2657, 0.3235, 0.3998, 0.3123, 0.2153, 0.3255], + device='cuda:2'), in_proj_covar=tensor([0.0201, 0.0227, 0.0198, 0.0220, 0.0236, 0.0198, 0.0193, 0.0211], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 12:20:26,450 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5170, 2.2110, 1.4405, 1.3822, 1.1280, 1.1877, 1.4275, 1.0582], + device='cuda:2'), covar=tensor([0.2549, 0.2164, 0.2854, 0.3284, 0.3933, 0.2859, 0.2169, 0.3264], + device='cuda:2'), in_proj_covar=tensor([0.0201, 0.0228, 0.0199, 0.0221, 0.0237, 0.0199, 0.0194, 0.0212], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 12:20:57,049 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2436, 1.4957, 1.3652, 2.0441, 1.8051, 2.1443, 1.5092, 4.3956], + device='cuda:2'), covar=tensor([0.0784, 0.0865, 0.0869, 0.1299, 0.0720, 0.0655, 0.0824, 0.0172], + device='cuda:2'), in_proj_covar=tensor([0.0041, 0.0041, 0.0042, 0.0047, 0.0042, 0.0041, 0.0041, 0.0066], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 12:21:07,311 INFO [finetune.py:976] (2/7) Epoch 2, batch 2950, loss[loss=0.2677, simple_loss=0.3046, pruned_loss=0.1154, over 4714.00 frames. ], tot_loss[loss=0.2672, simple_loss=0.3155, pruned_loss=0.1095, over 954371.92 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:21:22,377 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7405, 2.0986, 1.6089, 2.1000, 1.5371, 1.6789, 1.9091, 1.3635], + device='cuda:2'), covar=tensor([0.2131, 0.1521, 0.1305, 0.1298, 0.3449, 0.1574, 0.1859, 0.3050], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0338, 0.0248, 0.0313, 0.0318, 0.0287, 0.0281, 0.0305], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:21:30,316 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 12:21:52,171 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.13 vs. limit=5.0 +2023-04-26 12:22:03,515 INFO [optim.py:369] (2/7) 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 12:22:15,450 INFO [finetune.py:976] (2/7) Epoch 2, batch 3000, loss[loss=0.2661, simple_loss=0.325, pruned_loss=0.1036, over 4906.00 frames. ], tot_loss[loss=0.2687, simple_loss=0.3172, pruned_loss=0.1101, over 955371.66 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:22:15,450 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 12:22:23,966 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5462, 1.0592, 1.2846, 1.1553, 1.6882, 1.4357, 1.0930, 1.2954], + device='cuda:2'), covar=tensor([0.1756, 0.1734, 0.2493, 0.1894, 0.1158, 0.1572, 0.2090, 0.2322], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0342, 0.0351, 0.0318, 0.0355, 0.0376, 0.0326, 0.0360], + device='cuda:2'), out_proj_covar=tensor([7.1600e-05, 7.3557e-05, 7.5896e-05, 6.6750e-05, 7.5650e-05, 8.2187e-05, + 7.0993e-05, 7.7677e-05], device='cuda:2') +2023-04-26 12:22:25,545 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5052, 1.0388, 1.2443, 1.1016, 1.6880, 1.3931, 1.0319, 1.2586], + device='cuda:2'), covar=tensor([0.1556, 0.1536, 0.2435, 0.1871, 0.1004, 0.1445, 0.1999, 0.2032], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0342, 0.0351, 0.0318, 0.0355, 0.0376, 0.0326, 0.0360], + device='cuda:2'), out_proj_covar=tensor([7.1600e-05, 7.3557e-05, 7.5896e-05, 6.6750e-05, 7.5650e-05, 8.2187e-05, + 7.0993e-05, 7.7677e-05], device='cuda:2') +2023-04-26 12:22:32,547 INFO [finetune.py:1010] (2/7) Epoch 2, validation: loss=0.1863, simple_loss=0.2571, pruned_loss=0.0578, over 2265189.00 frames. +2023-04-26 12:22:32,548 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6094MB +2023-04-26 12:23:09,068 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8760.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:23:10,915 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0382, 2.7496, 2.0518, 2.5463, 2.1207, 2.1778, 2.4790, 1.8120], + device='cuda:2'), covar=tensor([0.3182, 0.2294, 0.1555, 0.2147, 0.3270, 0.2119, 0.2751, 0.3812], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0335, 0.0246, 0.0311, 0.0315, 0.0286, 0.0279, 0.0302], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:23:21,913 INFO [finetune.py:976] (2/7) Epoch 2, batch 3050, loss[loss=0.2804, simple_loss=0.3313, pruned_loss=0.1148, over 4909.00 frames. ], tot_loss[loss=0.2682, simple_loss=0.3171, pruned_loss=0.1097, over 956786.41 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:23:32,294 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3047, 2.8966, 1.0041, 1.3853, 2.1197, 1.3629, 3.8521, 1.7195], + device='cuda:2'), covar=tensor([0.0685, 0.1039, 0.1044, 0.1224, 0.0590, 0.0990, 0.0235, 0.0650], + device='cuda:2'), in_proj_covar=tensor([0.0057, 0.0073, 0.0054, 0.0051, 0.0056, 0.0056, 0.0087, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 12:23:42,734 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9668, 1.6525, 2.2011, 2.3938, 1.6701, 1.3418, 1.9613, 1.1680], + device='cuda:2'), covar=tensor([0.1048, 0.1167, 0.0812, 0.1085, 0.1299, 0.1668, 0.0966, 0.1442], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0079, 0.0075, 0.0074, 0.0087, 0.0095, 0.0089, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 12:23:49,150 INFO [optim.py:369] (2/7) 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 12:23:51,021 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8821.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:23:56,176 INFO [finetune.py:976] (2/7) Epoch 2, batch 3100, loss[loss=0.277, simple_loss=0.3128, pruned_loss=0.1206, over 4829.00 frames. ], tot_loss[loss=0.2662, simple_loss=0.3147, pruned_loss=0.1088, over 958316.37 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:24:04,018 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0793, 0.4780, 0.8743, 1.4698, 1.3700, 1.0420, 0.9933, 0.8973], + device='cuda:2'), covar=tensor([3.3118, 5.0170, 5.3903, 6.1263, 3.5789, 5.5941, 5.4703, 4.1502], + device='cuda:2'), in_proj_covar=tensor([0.0446, 0.0507, 0.0594, 0.0588, 0.0477, 0.0526, 0.0534, 0.0546], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:24:19,539 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1887, 2.6660, 0.9227, 1.3687, 1.9617, 1.2265, 3.6093, 1.7785], + device='cuda:2'), covar=tensor([0.0704, 0.0770, 0.0978, 0.1323, 0.0615, 0.1077, 0.0220, 0.0691], + device='cuda:2'), in_proj_covar=tensor([0.0057, 0.0074, 0.0055, 0.0051, 0.0056, 0.0057, 0.0087, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 12:24:29,315 INFO [finetune.py:976] (2/7) Epoch 2, batch 3150, loss[loss=0.2472, simple_loss=0.3032, pruned_loss=0.09558, over 4928.00 frames. ], tot_loss[loss=0.2628, simple_loss=0.3108, pruned_loss=0.1074, over 957370.59 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:24:54,711 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-26 12:24:56,257 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8921.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:01,754 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8927.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:02,221 INFO [finetune.py:976] (2/7) Epoch 2, batch 3200, loss[loss=0.2955, simple_loss=0.3183, pruned_loss=0.1363, over 4759.00 frames. ], tot_loss[loss=0.2566, simple_loss=0.3047, pruned_loss=0.1043, over 956169.30 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:25:21,285 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1547, 1.4499, 1.3436, 1.4086, 1.4479, 1.5636, 1.4254, 1.4340], + device='cuda:2'), covar=tensor([2.0538, 4.3942, 3.3958, 2.7520, 3.0754, 4.8338, 4.2548, 3.6961], + device='cuda:2'), in_proj_covar=tensor([0.0280, 0.0380, 0.0304, 0.0306, 0.0334, 0.0369, 0.0366, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:25:30,260 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8969.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:41,812 INFO [finetune.py:976] (2/7) Epoch 2, batch 3250, loss[loss=0.2974, simple_loss=0.3431, pruned_loss=0.1259, over 4832.00 frames. ], tot_loss[loss=0.2586, simple_loss=0.3065, pruned_loss=0.1053, over 955767.28 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:25:53,845 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8986.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:25:55,594 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8988.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:26:25,606 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3349, 1.1475, 1.4933, 1.6343, 1.4742, 1.2899, 1.3655, 1.3881], + device='cuda:2'), covar=tensor([ 5.7203, 8.0369, 10.2424, 9.1214, 5.7723, 9.9402, 10.5847, 7.2926], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0509, 0.0598, 0.0591, 0.0479, 0.0528, 0.0536, 0.0549], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:26:27,429 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:26:48,554 INFO [finetune.py:976] (2/7) Epoch 2, batch 3300, loss[loss=0.2536, simple_loss=0.3123, pruned_loss=0.09749, over 4801.00 frames. ], tot_loss[loss=0.263, simple_loss=0.3111, pruned_loss=0.1075, over 952394.70 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:27:08,777 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9047.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:27:26,231 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9074.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:27:28,045 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6408, 1.1703, 1.4500, 1.6761, 1.3005, 1.1261, 0.7194, 1.1990], + device='cuda:2'), covar=tensor([0.5970, 0.7358, 0.3504, 0.6005, 0.7562, 0.5646, 1.0404, 0.7185], + device='cuda:2'), in_proj_covar=tensor([0.0270, 0.0280, 0.0224, 0.0350, 0.0238, 0.0237, 0.0273, 0.0221], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:27:28,526 INFO [finetune.py:976] (2/7) Epoch 2, batch 3350, loss[loss=0.3283, simple_loss=0.3662, pruned_loss=0.1452, over 4822.00 frames. ], tot_loss[loss=0.2662, simple_loss=0.3143, pruned_loss=0.109, over 952510.02 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:28:00,090 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:28:17,539 INFO [finetune.py:976] (2/7) Epoch 2, batch 3400, loss[loss=0.2446, simple_loss=0.2938, pruned_loss=0.09771, over 4826.00 frames. ], tot_loss[loss=0.2662, simple_loss=0.3144, pruned_loss=0.109, over 952821.75 frames. ], batch size: 30, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:29:01,798 INFO [finetune.py:976] (2/7) Epoch 2, batch 3450, loss[loss=0.2604, simple_loss=0.306, pruned_loss=0.1074, over 4822.00 frames. ], tot_loss[loss=0.265, simple_loss=0.3134, pruned_loss=0.1083, over 953184.18 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:29:01,922 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8735, 2.6778, 1.8384, 2.0317, 1.4295, 1.5631, 1.8754, 1.5150], + device='cuda:2'), covar=tensor([0.1846, 0.1813, 0.2287, 0.2416, 0.3170, 0.2383, 0.1653, 0.2492], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0227, 0.0198, 0.0219, 0.0236, 0.0198, 0.0193, 0.0211], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 12:29:29,510 INFO [optim.py:369] (2/7) 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 12:29:35,077 INFO [finetune.py:976] (2/7) Epoch 2, batch 3500, loss[loss=0.2923, simple_loss=0.3301, pruned_loss=0.1272, over 4865.00 frames. ], tot_loss[loss=0.2598, simple_loss=0.3082, pruned_loss=0.1057, over 951396.48 frames. ], batch size: 34, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:30:08,949 INFO [finetune.py:976] (2/7) Epoch 2, batch 3550, loss[loss=0.2416, simple_loss=0.27, pruned_loss=0.1066, over 3942.00 frames. ], tot_loss[loss=0.2562, simple_loss=0.3044, pruned_loss=0.104, over 951817.71 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:30:12,101 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:30:48,985 INFO [finetune.py:976] (2/7) Epoch 2, batch 3600, loss[loss=0.1882, simple_loss=0.2539, pruned_loss=0.0612, over 4777.00 frames. ], tot_loss[loss=0.2529, simple_loss=0.3012, pruned_loss=0.1023, over 952970.27 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:31:01,647 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4122, 1.1738, 1.6924, 1.5821, 1.2322, 1.0573, 1.3542, 0.8430], + device='cuda:2'), covar=tensor([0.0957, 0.1050, 0.0622, 0.0963, 0.1203, 0.1459, 0.0829, 0.1382], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0079, 0.0075, 0.0073, 0.0086, 0.0094, 0.0089, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 12:31:03,417 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9342.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:31:44,877 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9369.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:31:48,042 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6641, 1.6368, 1.8999, 2.0997, 2.1070, 1.5726, 1.2302, 1.7961], + device='cuda:2'), covar=tensor([0.1209, 0.1307, 0.0707, 0.0766, 0.0653, 0.1279, 0.1292, 0.0823], + device='cuda:2'), in_proj_covar=tensor([0.0212, 0.0210, 0.0189, 0.0185, 0.0181, 0.0199, 0.0178, 0.0195], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:31:55,587 INFO [finetune.py:976] (2/7) Epoch 2, batch 3650, loss[loss=0.2311, simple_loss=0.2658, pruned_loss=0.09822, over 3995.00 frames. ], tot_loss[loss=0.2563, simple_loss=0.3044, pruned_loss=0.1041, over 950895.24 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:32:08,883 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5840, 1.1717, 4.2631, 4.0091, 3.7395, 3.9454, 3.9160, 3.7645], + device='cuda:2'), covar=tensor([0.6741, 0.6046, 0.0909, 0.1559, 0.1057, 0.1881, 0.1878, 0.1421], + device='cuda:2'), in_proj_covar=tensor([0.0333, 0.0314, 0.0449, 0.0454, 0.0378, 0.0432, 0.0342, 0.0399], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 12:32:25,430 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9414.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:32:27,076 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9416.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:32:28,803 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 2, batch 3700, loss[loss=0.2348, simple_loss=0.2961, pruned_loss=0.08673, over 4868.00 frames. ], tot_loss[loss=0.2609, simple_loss=0.3093, pruned_loss=0.1063, over 949587.99 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:32:42,847 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5233, 1.1793, 4.2451, 3.9501, 3.7475, 3.9717, 3.9287, 3.7749], + device='cuda:2'), covar=tensor([0.6677, 0.5871, 0.0969, 0.1639, 0.1006, 0.1502, 0.1673, 0.1427], + device='cuda:2'), in_proj_covar=tensor([0.0334, 0.0314, 0.0450, 0.0454, 0.0378, 0.0431, 0.0343, 0.0400], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 12:32:44,621 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9829, 1.8489, 1.5890, 1.4983, 1.9096, 1.6916, 2.1511, 1.4175], + device='cuda:2'), covar=tensor([0.2863, 0.1095, 0.3381, 0.2115, 0.1181, 0.1527, 0.1412, 0.3410], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0348, 0.0433, 0.0367, 0.0402, 0.0373, 0.0397, 0.0410], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:32:58,812 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9464.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:33:06,996 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9475.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:33:08,731 INFO [finetune.py:976] (2/7) Epoch 2, batch 3750, loss[loss=0.2557, simple_loss=0.3169, pruned_loss=0.09727, over 4816.00 frames. ], tot_loss[loss=0.2632, simple_loss=0.312, pruned_loss=0.1072, over 950885.05 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:33:47,551 INFO [optim.py:369] (2/7) 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 12:33:59,191 INFO [finetune.py:976] (2/7) Epoch 2, batch 3800, loss[loss=0.2657, simple_loss=0.2948, pruned_loss=0.1183, over 4745.00 frames. ], tot_loss[loss=0.2638, simple_loss=0.3128, pruned_loss=0.1074, over 951616.75 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:34:11,358 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5442, 1.3289, 0.7118, 1.2436, 1.4959, 1.4323, 1.3394, 1.4126], + device='cuda:2'), covar=tensor([0.0563, 0.0436, 0.0496, 0.0618, 0.0340, 0.0583, 0.0594, 0.0641], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 12:34:22,767 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 12:34:55,619 INFO [finetune.py:976] (2/7) Epoch 2, batch 3850, loss[loss=0.2412, simple_loss=0.2996, pruned_loss=0.09138, over 4906.00 frames. ], tot_loss[loss=0.2612, simple_loss=0.3104, pruned_loss=0.106, over 951591.55 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:34:58,835 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9583.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:34:59,485 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9584.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:35:01,986 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1461, 1.2967, 1.2317, 1.4148, 1.3709, 1.5799, 1.3861, 1.3878], + device='cuda:2'), covar=tensor([1.9304, 3.9393, 2.9997, 2.4429, 2.8924, 4.7709, 3.9163, 3.2950], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0384, 0.0306, 0.0309, 0.0337, 0.0377, 0.0369, 0.0335], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:35:11,352 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-26 12:35:22,631 INFO [optim.py:369] (2/7) 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 12:35:29,158 INFO [finetune.py:976] (2/7) Epoch 2, batch 3900, loss[loss=0.2548, simple_loss=0.318, pruned_loss=0.09574, over 4910.00 frames. ], tot_loss[loss=0.2566, simple_loss=0.3059, pruned_loss=0.1036, over 952846.15 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:35:38,162 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9631.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:35:50,773 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9642.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:35:52,027 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8398, 2.2701, 1.8126, 2.1988, 1.7030, 1.8401, 1.9353, 1.5281], + device='cuda:2'), covar=tensor([0.2337, 0.1373, 0.1230, 0.1471, 0.3437, 0.1591, 0.2055, 0.2945], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0337, 0.0248, 0.0313, 0.0317, 0.0289, 0.0280, 0.0304], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:35:52,632 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9645.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:36:12,999 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9658.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:36:25,010 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5952, 0.8330, 1.0764, 1.2540, 1.2913, 1.4634, 1.0927, 1.1489], + device='cuda:2'), covar=tensor([1.8200, 3.3698, 2.7936, 2.3402, 2.6163, 4.0477, 3.4641, 2.8591], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0385, 0.0306, 0.0309, 0.0337, 0.0377, 0.0369, 0.0335], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:36:26,740 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9669.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:36:38,403 INFO [finetune.py:976] (2/7) Epoch 2, batch 3950, loss[loss=0.2801, simple_loss=0.3164, pruned_loss=0.1219, over 4814.00 frames. ], tot_loss[loss=0.2538, simple_loss=0.3024, pruned_loss=0.1026, over 951636.94 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:36:55,195 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9690.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:37:06,996 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8685, 1.8606, 1.8128, 1.4978, 2.0396, 1.5224, 2.6496, 1.5036], + device='cuda:2'), covar=tensor([0.4620, 0.1714, 0.4804, 0.3533, 0.1699, 0.2843, 0.1385, 0.4722], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0349, 0.0435, 0.0368, 0.0403, 0.0373, 0.0398, 0.0412], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:37:29,236 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 12:37:29,748 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9717.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:37:36,039 INFO [optim.py:369] (2/7) 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 12:37:36,676 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9719.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:37:47,956 INFO [finetune.py:976] (2/7) Epoch 2, batch 4000, loss[loss=0.2631, simple_loss=0.2973, pruned_loss=0.1144, over 4704.00 frames. ], tot_loss[loss=0.2546, simple_loss=0.3026, pruned_loss=0.1033, over 952722.80 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:37:58,340 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9743.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:38:15,575 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9770.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:38:20,829 INFO [finetune.py:976] (2/7) Epoch 2, batch 4050, loss[loss=0.251, simple_loss=0.28, pruned_loss=0.111, over 4470.00 frames. ], tot_loss[loss=0.2561, simple_loss=0.3048, pruned_loss=0.1037, over 952653.46 frames. ], batch size: 19, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:38:38,411 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9804.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:38:41,407 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7012, 1.2844, 1.1950, 1.4667, 1.9851, 1.6503, 1.3798, 1.2003], + device='cuda:2'), covar=tensor([0.1498, 0.1935, 0.2490, 0.1910, 0.1024, 0.1991, 0.2546, 0.2025], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0344, 0.0352, 0.0320, 0.0357, 0.0375, 0.0326, 0.0359], + device='cuda:2'), out_proj_covar=tensor([7.1649e-05, 7.3991e-05, 7.6182e-05, 6.7265e-05, 7.6221e-05, 8.2080e-05, + 7.0899e-05, 7.7491e-05], device='cuda:2') +2023-04-26 12:38:47,823 INFO [optim.py:369] (2/7) 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 12:38:54,759 INFO [finetune.py:976] (2/7) Epoch 2, batch 4100, loss[loss=0.3102, simple_loss=0.3451, pruned_loss=0.1376, over 4241.00 frames. ], tot_loss[loss=0.2595, simple_loss=0.3091, pruned_loss=0.1049, over 953290.62 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:38:57,838 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7205, 4.0106, 0.8823, 2.1644, 2.1636, 2.6225, 2.5086, 0.9575], + device='cuda:2'), covar=tensor([0.1271, 0.0889, 0.2055, 0.1245, 0.0980, 0.1127, 0.1237, 0.2280], + device='cuda:2'), in_proj_covar=tensor([0.0125, 0.0272, 0.0151, 0.0132, 0.0144, 0.0166, 0.0129, 0.0134], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 12:39:20,796 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2034, 1.4864, 1.3546, 1.9894, 2.2050, 1.9215, 1.7458, 1.5078], + device='cuda:2'), covar=tensor([0.2390, 0.2507, 0.2547, 0.2294, 0.1308, 0.2795, 0.3185, 0.2160], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0344, 0.0352, 0.0320, 0.0358, 0.0375, 0.0327, 0.0359], + device='cuda:2'), out_proj_covar=tensor([7.1734e-05, 7.3904e-05, 7.6204e-05, 6.7273e-05, 7.6342e-05, 8.2050e-05, + 7.1106e-05, 7.7658e-05], device='cuda:2') +2023-04-26 12:39:34,299 INFO [finetune.py:976] (2/7) Epoch 2, batch 4150, loss[loss=0.2964, simple_loss=0.3427, pruned_loss=0.125, over 4906.00 frames. ], tot_loss[loss=0.2616, simple_loss=0.311, pruned_loss=0.1061, over 953704.84 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:39:43,556 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 12:40:18,083 INFO [optim.py:369] (2/7) 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 12:40:28,539 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2169, 1.6352, 1.5262, 1.9516, 1.8142, 2.2842, 1.6012, 3.8739], + device='cuda:2'), covar=tensor([0.0689, 0.0800, 0.0825, 0.1257, 0.0671, 0.0490, 0.0753, 0.0135], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 12:40:29,029 INFO [finetune.py:976] (2/7) Epoch 2, batch 4200, loss[loss=0.1971, simple_loss=0.2621, pruned_loss=0.06604, over 4809.00 frames. ], tot_loss[loss=0.2582, simple_loss=0.3081, pruned_loss=0.1041, over 951221.97 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:40:43,341 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 4250, loss[loss=0.2211, simple_loss=0.2712, pruned_loss=0.08554, over 4793.00 frames. ], tot_loss[loss=0.2577, simple_loss=0.3068, pruned_loss=0.1043, over 951520.30 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:42:30,340 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:42:43,921 INFO [finetune.py:976] (2/7) Epoch 2, batch 4300, loss[loss=0.2262, simple_loss=0.2846, pruned_loss=0.08388, over 4863.00 frames. ], tot_loss[loss=0.2546, simple_loss=0.3037, pruned_loss=0.1028, over 954077.34 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:43:02,768 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10039.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:43:18,147 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10062.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:43:23,044 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10070.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:43:27,737 INFO [finetune.py:976] (2/7) Epoch 2, batch 4350, loss[loss=0.2732, simple_loss=0.3037, pruned_loss=0.1214, over 4863.00 frames. ], tot_loss[loss=0.2514, simple_loss=0.3003, pruned_loss=0.1012, over 954918.26 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:43:30,482 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 12:43:39,627 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2266, 2.2215, 2.5114, 2.6807, 2.5336, 1.9685, 1.4988, 2.2121], + device='cuda:2'), covar=tensor([0.1251, 0.1070, 0.0637, 0.0787, 0.0734, 0.1294, 0.1498, 0.0798], + device='cuda:2'), in_proj_covar=tensor([0.0212, 0.0211, 0.0190, 0.0184, 0.0181, 0.0201, 0.0178, 0.0196], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:43:42,009 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10100.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:43:54,962 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:43:58,059 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10123.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:44:00,980 INFO [finetune.py:976] (2/7) Epoch 2, batch 4400, loss[loss=0.338, simple_loss=0.3785, pruned_loss=0.1487, over 4280.00 frames. ], tot_loss[loss=0.2532, simple_loss=0.3017, pruned_loss=0.1024, over 954856.38 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:44:34,791 INFO [finetune.py:976] (2/7) Epoch 2, batch 4450, loss[loss=0.2788, simple_loss=0.3366, pruned_loss=0.1105, over 4824.00 frames. ], tot_loss[loss=0.2564, simple_loss=0.3059, pruned_loss=0.1035, over 954781.84 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:45:03,112 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 2, batch 4500, loss[loss=0.224, simple_loss=0.2782, pruned_loss=0.08489, over 4045.00 frames. ], tot_loss[loss=0.2568, simple_loss=0.3065, pruned_loss=0.1036, over 952669.90 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:45:15,380 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7839, 3.7497, 2.8004, 4.3372, 3.7598, 3.7732, 1.7527, 3.6349], + device='cuda:2'), covar=tensor([0.1550, 0.1225, 0.3375, 0.1301, 0.2609, 0.1775, 0.5216, 0.2020], + device='cuda:2'), in_proj_covar=tensor([0.0256, 0.0230, 0.0273, 0.0322, 0.0318, 0.0267, 0.0281, 0.0284], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:45:15,987 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10240.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:45:38,665 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0843, 1.5475, 1.4217, 1.8329, 1.5867, 1.8504, 1.4375, 3.1122], + device='cuda:2'), covar=tensor([0.0718, 0.0782, 0.0831, 0.1226, 0.0690, 0.0514, 0.0749, 0.0209], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 12:45:42,233 INFO [finetune.py:976] (2/7) Epoch 2, batch 4550, loss[loss=0.237, simple_loss=0.2948, pruned_loss=0.08961, over 4817.00 frames. ], tot_loss[loss=0.2582, simple_loss=0.308, pruned_loss=0.1042, over 951599.82 frames. ], batch size: 30, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:45:48,411 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10288.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:45:57,831 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3888, 1.4069, 3.7936, 3.5240, 3.3758, 3.6366, 3.5832, 3.3597], + device='cuda:2'), covar=tensor([0.6122, 0.5178, 0.1073, 0.1656, 0.1044, 0.1663, 0.1933, 0.1282], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0311, 0.0445, 0.0449, 0.0376, 0.0431, 0.0338, 0.0396], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 12:46:21,445 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10314.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:46:28,877 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-26 12:46:29,829 INFO [optim.py:369] (2/7) 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 12:46:40,809 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6382, 1.7659, 0.6134, 1.3220, 1.9429, 1.5485, 1.3898, 1.5244], + device='cuda:2'), covar=tensor([0.0614, 0.0466, 0.0508, 0.0656, 0.0323, 0.0613, 0.0597, 0.0724], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0051], + device='cuda:2') +2023-04-26 12:46:41,916 INFO [finetune.py:976] (2/7) Epoch 2, batch 4600, loss[loss=0.2317, simple_loss=0.2842, pruned_loss=0.08954, over 4805.00 frames. ], tot_loss[loss=0.2558, simple_loss=0.3066, pruned_loss=0.1025, over 951865.51 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:47:08,791 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8706, 2.3694, 1.9581, 2.2168, 1.6804, 1.8983, 1.9555, 1.5338], + device='cuda:2'), covar=tensor([0.2615, 0.1815, 0.1312, 0.1872, 0.3629, 0.1754, 0.2308, 0.3687], + device='cuda:2'), in_proj_covar=tensor([0.0319, 0.0339, 0.0249, 0.0313, 0.0318, 0.0290, 0.0282, 0.0305], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:47:09,969 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 4650, loss[loss=0.2572, simple_loss=0.2956, pruned_loss=0.1094, over 4897.00 frames. ], tot_loss[loss=0.2531, simple_loss=0.3036, pruned_loss=0.1013, over 953935.94 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 64.0 +2023-04-26 12:47:50,967 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10395.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 12:47:58,941 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10399.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:48:10,176 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8987, 1.7974, 1.9626, 2.2443, 2.2617, 1.7613, 1.3413, 1.9699], + device='cuda:2'), covar=tensor([0.1220, 0.1302, 0.0847, 0.0773, 0.0750, 0.1330, 0.1427, 0.0788], + device='cuda:2'), in_proj_covar=tensor([0.0216, 0.0215, 0.0193, 0.0187, 0.0184, 0.0204, 0.0181, 0.0199], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:48:12,498 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10418.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:48:25,137 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 2, batch 4700, loss[loss=0.2201, simple_loss=0.2685, pruned_loss=0.08589, over 4924.00 frames. ], tot_loss[loss=0.2496, simple_loss=0.2998, pruned_loss=0.09973, over 955398.96 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:48:53,255 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10447.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:48:55,146 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5275, 3.0804, 0.9411, 1.4887, 2.2640, 1.4476, 4.4140, 2.0403], + device='cuda:2'), covar=tensor([0.0658, 0.0784, 0.1055, 0.1405, 0.0616, 0.1090, 0.0253, 0.0732], + device='cuda:2'), in_proj_covar=tensor([0.0056, 0.0072, 0.0054, 0.0050, 0.0055, 0.0055, 0.0085, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 12:48:57,673 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10454.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:49:09,541 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10470.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:49:15,191 INFO [finetune.py:976] (2/7) Epoch 2, batch 4750, loss[loss=0.2128, simple_loss=0.2743, pruned_loss=0.07562, over 4901.00 frames. ], tot_loss[loss=0.2464, simple_loss=0.2966, pruned_loss=0.09815, over 955791.84 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:49:34,661 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3030, 1.5728, 1.4937, 2.0175, 2.3059, 1.9762, 1.8168, 1.7127], + device='cuda:2'), covar=tensor([0.1599, 0.2040, 0.1899, 0.1906, 0.1284, 0.1934, 0.2510, 0.1861], + device='cuda:2'), in_proj_covar=tensor([0.0324, 0.0339, 0.0348, 0.0316, 0.0353, 0.0367, 0.0320, 0.0354], + device='cuda:2'), out_proj_covar=tensor([7.0733e-05, 7.2861e-05, 7.5264e-05, 6.6473e-05, 7.5280e-05, 8.0395e-05, + 6.9708e-05, 7.6552e-05], device='cuda:2') +2023-04-26 12:49:40,417 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10515.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:49:43,424 INFO [optim.py:369] (2/7) 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 12:49:49,341 INFO [finetune.py:976] (2/7) Epoch 2, batch 4800, loss[loss=0.2608, simple_loss=0.3142, pruned_loss=0.1037, over 4903.00 frames. ], tot_loss[loss=0.25, simple_loss=0.3001, pruned_loss=0.1, over 956852.65 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:50:23,189 INFO [finetune.py:976] (2/7) Epoch 2, batch 4850, loss[loss=0.311, simple_loss=0.3495, pruned_loss=0.1362, over 4841.00 frames. ], tot_loss[loss=0.2531, simple_loss=0.3036, pruned_loss=0.1013, over 954370.10 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:50:23,947 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6712, 1.4498, 1.2555, 1.4335, 1.9676, 1.6074, 1.3167, 1.2129], + device='cuda:2'), covar=tensor([0.1649, 0.1962, 0.2558, 0.2184, 0.1315, 0.2224, 0.2687, 0.2006], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0343, 0.0352, 0.0321, 0.0357, 0.0372, 0.0324, 0.0358], + device='cuda:2'), out_proj_covar=tensor([7.1649e-05, 7.3729e-05, 7.6326e-05, 6.7452e-05, 7.6206e-05, 8.1476e-05, + 7.0640e-05, 7.7457e-05], device='cuda:2') +2023-04-26 12:50:50,776 INFO [optim.py:369] (2/7) 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 12:50:56,075 INFO [finetune.py:976] (2/7) Epoch 2, batch 4900, loss[loss=0.2413, simple_loss=0.2811, pruned_loss=0.1007, over 4717.00 frames. ], tot_loss[loss=0.2536, simple_loss=0.3044, pruned_loss=0.1014, over 955759.77 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:51:03,405 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-26 12:51:46,565 INFO [finetune.py:976] (2/7) Epoch 2, batch 4950, loss[loss=0.2653, simple_loss=0.3138, pruned_loss=0.1084, over 4298.00 frames. ], tot_loss[loss=0.2549, simple_loss=0.3064, pruned_loss=0.1017, over 957299.60 frames. ], batch size: 66, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:51:46,697 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0942, 2.7567, 2.1438, 2.4729, 1.9585, 2.3007, 2.3709, 1.9474], + device='cuda:2'), covar=tensor([0.2440, 0.1388, 0.1266, 0.1691, 0.3463, 0.1446, 0.2511, 0.3192], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0337, 0.0248, 0.0312, 0.0316, 0.0290, 0.0280, 0.0303], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:52:10,671 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10695.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:52:11,281 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1768, 2.0667, 2.3234, 2.6827, 2.6271, 2.0611, 1.6650, 2.2619], + device='cuda:2'), covar=tensor([0.1226, 0.1189, 0.0705, 0.0712, 0.0689, 0.1251, 0.1327, 0.0765], + device='cuda:2'), in_proj_covar=tensor([0.0215, 0.0215, 0.0193, 0.0188, 0.0184, 0.0204, 0.0181, 0.0198], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:52:24,092 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 12:52:36,491 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10718.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 12:52:43,790 INFO [optim.py:369] (2/7) 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 12:52:54,984 INFO [finetune.py:976] (2/7) Epoch 2, batch 5000, loss[loss=0.2086, simple_loss=0.2734, pruned_loss=0.07188, over 4788.00 frames. ], tot_loss[loss=0.2506, simple_loss=0.3022, pruned_loss=0.09954, over 957270.91 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:53:16,590 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10743.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:53:17,248 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10744.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:53:26,972 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0678, 1.9207, 4.5934, 4.2946, 4.1059, 4.3056, 4.2803, 4.1075], + device='cuda:2'), covar=tensor([0.5873, 0.4908, 0.0998, 0.1566, 0.0971, 0.1789, 0.1071, 0.1257], + device='cuda:2'), in_proj_covar=tensor([0.0333, 0.0312, 0.0445, 0.0449, 0.0378, 0.0432, 0.0339, 0.0399], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 12:53:29,935 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:53:30,542 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10766.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:53:33,507 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2017, 1.2209, 1.3282, 1.5316, 1.5747, 1.2252, 0.8280, 1.3995], + device='cuda:2'), covar=tensor([0.0926, 0.1485, 0.0950, 0.0652, 0.0649, 0.0951, 0.1232, 0.0665], + device='cuda:2'), in_proj_covar=tensor([0.0213, 0.0213, 0.0191, 0.0186, 0.0182, 0.0202, 0.0179, 0.0196], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:53:42,296 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 5050, loss[loss=0.227, simple_loss=0.2787, pruned_loss=0.08766, over 4885.00 frames. ], tot_loss[loss=0.2469, simple_loss=0.2978, pruned_loss=0.09799, over 954358.22 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:53:53,144 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-04-26 12:54:15,025 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4227, 1.1690, 1.3904, 1.6749, 1.5757, 1.3251, 1.3343, 1.3647], + device='cuda:2'), covar=tensor([2.8360, 3.9985, 4.5497, 5.1717, 3.1184, 4.8575, 5.0077, 3.6778], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0506, 0.0599, 0.0598, 0.0481, 0.0524, 0.0534, 0.0545], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:54:25,692 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10810.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 12:54:40,510 INFO [optim.py:369] (2/7) 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 12:54:51,192 INFO [finetune.py:976] (2/7) Epoch 2, batch 5100, loss[loss=0.2676, simple_loss=0.3044, pruned_loss=0.1154, over 4748.00 frames. ], tot_loss[loss=0.2445, simple_loss=0.2948, pruned_loss=0.09706, over 950579.58 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:55:01,976 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10835.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:55:45,708 INFO [finetune.py:976] (2/7) Epoch 2, batch 5150, loss[loss=0.1798, simple_loss=0.2337, pruned_loss=0.06294, over 4728.00 frames. ], tot_loss[loss=0.247, simple_loss=0.2965, pruned_loss=0.09875, over 950287.09 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:55:50,606 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10885.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:55:59,077 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-26 12:56:04,158 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4620, 1.5123, 1.7365, 1.8960, 2.0204, 1.4481, 1.0771, 1.6464], + device='cuda:2'), covar=tensor([0.1226, 0.1354, 0.0784, 0.0802, 0.0671, 0.1147, 0.1416, 0.0751], + device='cuda:2'), in_proj_covar=tensor([0.0215, 0.0215, 0.0192, 0.0187, 0.0184, 0.0204, 0.0180, 0.0199], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:56:13,137 INFO [optim.py:369] (2/7) 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 12:56:15,757 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 12:56:18,492 INFO [finetune.py:976] (2/7) Epoch 2, batch 5200, loss[loss=0.311, simple_loss=0.3566, pruned_loss=0.1327, over 4821.00 frames. ], tot_loss[loss=0.2496, simple_loss=0.3002, pruned_loss=0.0995, over 950342.07 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:56:20,498 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9501, 1.9026, 1.5995, 1.5244, 2.0351, 1.6113, 2.4606, 1.3664], + device='cuda:2'), covar=tensor([0.4014, 0.1631, 0.4703, 0.3012, 0.1850, 0.2385, 0.1264, 0.4706], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0350, 0.0434, 0.0367, 0.0402, 0.0374, 0.0399, 0.0413], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 12:56:21,669 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4345, 1.3091, 0.8275, 1.1502, 1.5316, 1.3207, 1.2125, 1.3136], + device='cuda:2'), covar=tensor([0.0587, 0.0500, 0.0480, 0.0630, 0.0376, 0.0573, 0.0596, 0.0715], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0030, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0037, 0.0047, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 12:56:31,970 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10946.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 12:56:57,990 INFO [finetune.py:976] (2/7) Epoch 2, batch 5250, loss[loss=0.2461, simple_loss=0.2882, pruned_loss=0.102, over 4928.00 frames. ], tot_loss[loss=0.252, simple_loss=0.3033, pruned_loss=0.1003, over 949265.93 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:57:21,611 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8916, 2.7970, 2.2053, 3.2680, 2.8098, 2.8695, 1.1733, 2.7774], + device='cuda:2'), covar=tensor([0.2049, 0.1609, 0.3338, 0.2932, 0.3039, 0.2313, 0.5854, 0.2778], + device='cuda:2'), in_proj_covar=tensor([0.0255, 0.0230, 0.0273, 0.0324, 0.0316, 0.0268, 0.0281, 0.0284], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 12:57:40,221 INFO [optim.py:369] (2/7) 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 12:57:51,419 INFO [finetune.py:976] (2/7) Epoch 2, batch 5300, loss[loss=0.333, simple_loss=0.3716, pruned_loss=0.1472, over 4805.00 frames. ], tot_loss[loss=0.2546, simple_loss=0.3058, pruned_loss=0.1017, over 950573.82 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:58:45,101 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 5350, loss[loss=0.2387, simple_loss=0.2926, pruned_loss=0.09238, over 4814.00 frames. ], tot_loss[loss=0.2526, simple_loss=0.3043, pruned_loss=0.1005, over 950037.02 frames. ], batch size: 41, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 12:59:29,511 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11100.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 12:59:47,720 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11110.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 12:59:50,021 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 13:00:11,118 INFO [finetune.py:976] (2/7) Epoch 2, batch 5400, loss[loss=0.2972, simple_loss=0.3465, pruned_loss=0.1239, over 4661.00 frames. ], tot_loss[loss=0.2504, simple_loss=0.3017, pruned_loss=0.09958, over 950676.90 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:00:11,998 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-04-26 13:00:12,417 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:00:32,457 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11158.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:00:37,257 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0871, 1.9494, 2.2323, 2.5140, 2.4449, 1.9427, 1.5538, 2.1255], + device='cuda:2'), covar=tensor([0.1183, 0.1216, 0.0603, 0.0793, 0.0714, 0.1238, 0.1425, 0.0683], + device='cuda:2'), in_proj_covar=tensor([0.0214, 0.0214, 0.0192, 0.0188, 0.0183, 0.0203, 0.0180, 0.0199], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:00:42,736 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 2, batch 5450, loss[loss=0.2385, simple_loss=0.2893, pruned_loss=0.09382, over 4934.00 frames. ], tot_loss[loss=0.2449, simple_loss=0.2962, pruned_loss=0.09683, over 951209.94 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:01:00,531 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3750, 1.6736, 1.5450, 2.1037, 1.7939, 2.0220, 1.6360, 3.4382], + device='cuda:2'), covar=tensor([0.0656, 0.0744, 0.0778, 0.1084, 0.0617, 0.0502, 0.0746, 0.0207], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0041, 0.0041, 0.0041, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 13:01:18,381 INFO [optim.py:369] (2/7) 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 13:01:23,718 INFO [finetune.py:976] (2/7) Epoch 2, batch 5500, loss[loss=0.242, simple_loss=0.296, pruned_loss=0.09402, over 4820.00 frames. ], tot_loss[loss=0.2395, simple_loss=0.2915, pruned_loss=0.09376, over 951066.45 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:01:28,092 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11235.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:01:31,702 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11241.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 13:01:57,533 INFO [finetune.py:976] (2/7) Epoch 2, batch 5550, loss[loss=0.2525, simple_loss=0.3143, pruned_loss=0.09536, over 4910.00 frames. ], tot_loss[loss=0.2421, simple_loss=0.2937, pruned_loss=0.09527, over 950484.57 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:02:03,149 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1830, 1.3757, 1.3363, 1.4637, 1.4529, 1.5309, 1.3945, 1.3791], + device='cuda:2'), covar=tensor([1.7436, 2.9612, 2.2476, 1.9345, 2.2713, 3.6521, 3.0315, 2.5246], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0393, 0.0311, 0.0315, 0.0344, 0.0389, 0.0378, 0.0341], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:02:31,826 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5428, 3.1830, 0.9523, 1.7328, 1.6247, 2.3534, 1.9865, 0.9604], + device='cuda:2'), covar=tensor([0.1490, 0.1156, 0.2265, 0.1631, 0.1395, 0.1218, 0.1560, 0.2403], + device='cuda:2'), in_proj_covar=tensor([0.0125, 0.0272, 0.0152, 0.0133, 0.0144, 0.0165, 0.0129, 0.0134], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:02:35,254 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 2, batch 5600, loss[loss=0.2706, simple_loss=0.3272, pruned_loss=0.107, over 4933.00 frames. ], tot_loss[loss=0.2466, simple_loss=0.2991, pruned_loss=0.09702, over 952114.44 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:03:30,853 INFO [scaling.py:679] (2/7) 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] (2/7) Epoch 2, batch 5650, loss[loss=0.2485, simple_loss=0.3181, pruned_loss=0.08944, over 4910.00 frames. ], tot_loss[loss=0.2478, simple_loss=0.3017, pruned_loss=0.09701, over 953189.60 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:04:07,966 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0718, 1.5163, 1.6932, 1.6614, 2.1828, 1.9767, 1.5561, 1.6066], + device='cuda:2'), covar=tensor([0.1761, 0.1734, 0.1958, 0.1663, 0.1017, 0.1377, 0.2138, 0.1992], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0339, 0.0349, 0.0317, 0.0351, 0.0365, 0.0321, 0.0354], + device='cuda:2'), 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:2') +2023-04-26 13:04:13,907 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11400.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:04:31,190 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 2, batch 5700, loss[loss=0.218, simple_loss=0.2616, pruned_loss=0.08717, over 3848.00 frames. ], tot_loss[loss=0.2461, simple_loss=0.2979, pruned_loss=0.09715, over 934828.82 frames. ], batch size: 16, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:04:37,184 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11430.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:04:44,392 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9357, 1.9149, 2.0555, 1.6438, 2.1367, 1.6430, 2.6751, 1.7435], + device='cuda:2'), covar=tensor([0.4115, 0.1835, 0.4278, 0.2873, 0.1603, 0.2411, 0.1514, 0.4026], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0352, 0.0434, 0.0369, 0.0402, 0.0377, 0.0400, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:04:47,961 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11448.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:05:07,854 INFO [finetune.py:976] (2/7) Epoch 3, batch 0, loss[loss=0.3374, simple_loss=0.376, pruned_loss=0.1494, over 4698.00 frames. ], tot_loss[loss=0.3374, simple_loss=0.376, pruned_loss=0.1494, over 4698.00 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:05:07,854 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 13:05:24,857 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6274MB +2023-04-26 13:05:42,225 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 50, loss[loss=0.3052, simple_loss=0.3244, pruned_loss=0.143, over 4231.00 frames. ], tot_loss[loss=0.2576, simple_loss=0.3072, pruned_loss=0.104, over 212834.08 frames. ], batch size: 66, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:06:11,125 INFO [optim.py:369] (2/7) 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 13:06:17,275 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11530.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:06:18,528 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11532.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:06:24,403 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11541.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:06:34,900 INFO [finetune.py:976] (2/7) Epoch 3, batch 100, loss[loss=0.2327, simple_loss=0.2819, pruned_loss=0.09171, over 4864.00 frames. ], tot_loss[loss=0.2494, simple_loss=0.2986, pruned_loss=0.1001, over 376606.39 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:06:55,970 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11589.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:06:58,871 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6677, 2.1111, 1.1763, 1.3500, 2.3489, 1.5508, 1.4948, 1.6096], + device='cuda:2'), covar=tensor([0.0690, 0.0355, 0.0380, 0.0635, 0.0257, 0.0746, 0.0700, 0.0728], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 13:06:58,896 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 150, loss[loss=0.2256, simple_loss=0.2674, pruned_loss=0.09185, over 4710.00 frames. ], tot_loss[loss=0.2436, simple_loss=0.2932, pruned_loss=0.09702, over 504984.06 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:07:18,013 INFO [optim.py:369] (2/7) 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 13:07:27,317 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11635.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:07:42,006 INFO [finetune.py:976] (2/7) Epoch 3, batch 200, loss[loss=0.2992, simple_loss=0.317, pruned_loss=0.1407, over 4061.00 frames. ], tot_loss[loss=0.241, simple_loss=0.2911, pruned_loss=0.09541, over 603266.45 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:07:46,367 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 13:08:29,867 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11696.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:08:31,707 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 250, loss[loss=0.2016, simple_loss=0.2553, pruned_loss=0.07397, over 4680.00 frames. ], tot_loss[loss=0.2436, simple_loss=0.2942, pruned_loss=0.09652, over 678592.39 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:09:02,116 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 3, batch 300, loss[loss=0.2783, simple_loss=0.3204, pruned_loss=0.1181, over 4931.00 frames. ], tot_loss[loss=0.2467, simple_loss=0.2981, pruned_loss=0.09764, over 739491.75 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:09:40,545 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11760.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:09:57,764 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3435, 2.8232, 1.1058, 1.7370, 1.6883, 2.1622, 1.8956, 1.2007], + device='cuda:2'), covar=tensor([0.1228, 0.0855, 0.1630, 0.1224, 0.1046, 0.0910, 0.1213, 0.1924], + device='cuda:2'), in_proj_covar=tensor([0.0125, 0.0271, 0.0152, 0.0131, 0.0144, 0.0165, 0.0129, 0.0133], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:10:09,493 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1133, 2.3378, 0.8831, 1.3831, 1.4959, 1.7489, 1.5809, 0.8403], + device='cuda:2'), covar=tensor([0.1387, 0.1165, 0.1730, 0.1429, 0.1087, 0.0995, 0.1537, 0.1470], + device='cuda:2'), in_proj_covar=tensor([0.0125, 0.0270, 0.0151, 0.0131, 0.0143, 0.0165, 0.0128, 0.0133], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:10:10,601 INFO [finetune.py:976] (2/7) Epoch 3, batch 350, loss[loss=0.2064, simple_loss=0.2652, pruned_loss=0.07378, over 4699.00 frames. ], tot_loss[loss=0.2493, simple_loss=0.3007, pruned_loss=0.09898, over 784570.04 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:10:18,390 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11816.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:10:20,681 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6622, 1.8887, 1.0721, 1.2967, 2.1485, 1.5197, 1.4198, 1.6177], + device='cuda:2'), covar=tensor([0.0589, 0.0410, 0.0405, 0.0617, 0.0289, 0.0567, 0.0518, 0.0668], + device='cuda:2'), in_proj_covar=tensor([0.0032, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 13:10:21,167 INFO [optim.py:369] (2/7) 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 13:10:28,350 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11830.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:10:38,186 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11846.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:10:50,156 INFO [finetune.py:976] (2/7) Epoch 3, batch 400, loss[loss=0.2927, simple_loss=0.3208, pruned_loss=0.1323, over 4756.00 frames. ], tot_loss[loss=0.2517, simple_loss=0.3039, pruned_loss=0.09972, over 822513.32 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:11:20,865 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11877.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:11:21,407 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11878.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:11:33,332 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11888.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:11:44,732 INFO [finetune.py:976] (2/7) Epoch 3, batch 450, loss[loss=0.2201, simple_loss=0.2691, pruned_loss=0.08557, over 4734.00 frames. ], tot_loss[loss=0.2507, simple_loss=0.3028, pruned_loss=0.09924, over 853495.14 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:11:45,494 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 13:11:54,869 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11920.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:13,316 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 500, loss[loss=0.2241, simple_loss=0.2819, pruned_loss=0.0832, over 4769.00 frames. ], tot_loss[loss=0.2464, simple_loss=0.2983, pruned_loss=0.09719, over 874910.87 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:12:36,395 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11981.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:38,123 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11984.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:42,326 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11991.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:12:49,718 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-04-26 13:12:50,390 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-04-26 13:12:52,573 INFO [finetune.py:976] (2/7) Epoch 3, batch 550, loss[loss=0.2211, simple_loss=0.2775, pruned_loss=0.08236, over 4804.00 frames. ], tot_loss[loss=0.2432, simple_loss=0.2948, pruned_loss=0.09582, over 893674.05 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:12:55,036 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12009.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:13:02,613 INFO [optim.py:369] (2/7) 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 13:13:19,331 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12045.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:13:19,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2128, 1.5856, 2.0738, 2.4384, 1.9596, 1.5915, 1.3466, 1.8088], + device='cuda:2'), covar=tensor([0.5658, 0.6424, 0.3023, 0.5175, 0.5952, 0.4738, 0.8494, 0.6097], + device='cuda:2'), in_proj_covar=tensor([0.0273, 0.0278, 0.0226, 0.0348, 0.0235, 0.0238, 0.0269, 0.0215], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:13:30,911 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12055.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:13:31,439 INFO [finetune.py:976] (2/7) Epoch 3, batch 600, loss[loss=0.2648, simple_loss=0.3246, pruned_loss=0.1026, over 4729.00 frames. ], tot_loss[loss=0.2423, simple_loss=0.2942, pruned_loss=0.0952, over 907389.53 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:14:01,461 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12076.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:14:26,161 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-26 13:14:38,092 INFO [finetune.py:976] (2/7) Epoch 3, batch 650, loss[loss=0.2563, simple_loss=0.325, pruned_loss=0.09377, over 4751.00 frames. ], tot_loss[loss=0.2472, simple_loss=0.2993, pruned_loss=0.09754, over 917633.29 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-04-26 13:14:38,247 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3462, 1.1466, 1.4123, 1.5558, 1.4993, 1.3219, 1.3559, 1.3639], + device='cuda:2'), covar=tensor([2.5552, 3.3530, 3.9810, 4.2068, 2.6196, 4.1451, 4.2681, 3.2296], + device='cuda:2'), in_proj_covar=tensor([0.0449, 0.0504, 0.0599, 0.0600, 0.0481, 0.0518, 0.0529, 0.0542], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:14:47,779 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0131, 1.2644, 1.8093, 2.3071, 1.6271, 1.2572, 1.0896, 1.5313], + device='cuda:2'), covar=tensor([0.5291, 0.6553, 0.2876, 0.5006, 0.6225, 0.4841, 0.8373, 0.5762], + device='cuda:2'), in_proj_covar=tensor([0.0271, 0.0277, 0.0225, 0.0346, 0.0234, 0.0237, 0.0268, 0.0214], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:14:55,287 INFO [optim.py:369] (2/7) 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 13:15:07,089 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4442, 2.3384, 1.9348, 2.1007, 2.5437, 2.0312, 3.2608, 1.7131], + device='cuda:2'), covar=tensor([0.4693, 0.2230, 0.4862, 0.3762, 0.2086, 0.2737, 0.1782, 0.4382], + device='cuda:2'), in_proj_covar=tensor([0.0351, 0.0355, 0.0438, 0.0373, 0.0404, 0.0379, 0.0401, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:15:14,501 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-26 13:15:18,809 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12137.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:15:30,235 INFO [finetune.py:976] (2/7) Epoch 3, batch 700, loss[loss=0.2406, simple_loss=0.2999, pruned_loss=0.09067, over 4813.00 frames. ], tot_loss[loss=0.2474, simple_loss=0.3003, pruned_loss=0.09727, over 926129.31 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:15:40,483 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12172.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:15:40,530 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4645, 1.2901, 1.7651, 1.6182, 1.2746, 1.0438, 1.4643, 1.0475], + device='cuda:2'), covar=tensor([0.0970, 0.0863, 0.0603, 0.0842, 0.1191, 0.1844, 0.0837, 0.1079], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0079, 0.0075, 0.0072, 0.0085, 0.0096, 0.0088, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 13:15:52,188 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12188.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:15:58,346 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 13:16:00,827 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12202.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:16:03,168 INFO [finetune.py:976] (2/7) Epoch 3, batch 750, loss[loss=0.2125, simple_loss=0.2776, pruned_loss=0.07369, over 4874.00 frames. ], tot_loss[loss=0.2456, simple_loss=0.299, pruned_loss=0.09608, over 932531.05 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:16:11,580 INFO [optim.py:369] (2/7) 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 13:16:23,274 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:16:42,544 INFO [finetune.py:976] (2/7) Epoch 3, batch 800, loss[loss=0.2634, simple_loss=0.3113, pruned_loss=0.1077, over 4819.00 frames. ], tot_loss[loss=0.2437, simple_loss=0.2975, pruned_loss=0.09493, over 937185.01 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:17:00,570 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12276.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:04,549 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4272, 3.3612, 2.6939, 3.9019, 3.4547, 3.4512, 1.5075, 3.3455], + device='cuda:2'), covar=tensor([0.1844, 0.1436, 0.3084, 0.2343, 0.3817, 0.2156, 0.5721, 0.2708], + device='cuda:2'), in_proj_covar=tensor([0.0255, 0.0229, 0.0270, 0.0323, 0.0315, 0.0266, 0.0281, 0.0283], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:17:11,620 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12291.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:13,331 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4767, 1.2629, 4.1325, 3.8519, 3.6688, 3.9240, 3.9074, 3.6205], + device='cuda:2'), covar=tensor([0.6554, 0.5608, 0.1115, 0.1674, 0.1090, 0.1523, 0.1331, 0.1531], + device='cuda:2'), in_proj_covar=tensor([0.0326, 0.0307, 0.0435, 0.0442, 0.0372, 0.0422, 0.0335, 0.0390], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:17:19,898 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 850, loss[loss=0.2267, simple_loss=0.2861, pruned_loss=0.08359, over 4911.00 frames. ], tot_loss[loss=0.2417, simple_loss=0.2951, pruned_loss=0.09411, over 939653.81 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:17:29,560 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12339.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:43,715 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12340.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:54,285 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12355.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:17:54,806 INFO [finetune.py:976] (2/7) Epoch 3, batch 900, loss[loss=0.199, simple_loss=0.2492, pruned_loss=0.07441, over 4752.00 frames. ], tot_loss[loss=0.2387, simple_loss=0.2917, pruned_loss=0.09288, over 943127.15 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:18:26,965 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 950, loss[loss=0.2014, simple_loss=0.258, pruned_loss=0.07242, over 4832.00 frames. ], tot_loss[loss=0.2376, simple_loss=0.2904, pruned_loss=0.09242, over 946050.40 frames. ], batch size: 30, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:18:37,304 INFO [optim.py:369] (2/7) 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 13:18:50,722 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12432.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:19:13,312 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 1000, loss[loss=0.2549, simple_loss=0.2811, pruned_loss=0.1143, over 3789.00 frames. ], tot_loss[loss=0.2411, simple_loss=0.2937, pruned_loss=0.09431, over 947004.18 frames. ], batch size: 16, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:19:30,631 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12472.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:19:57,894 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12502.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:20:06,939 INFO [finetune.py:976] (2/7) Epoch 3, batch 1050, loss[loss=0.2184, simple_loss=0.2896, pruned_loss=0.07364, over 4821.00 frames. ], tot_loss[loss=0.2433, simple_loss=0.2964, pruned_loss=0.09511, over 949393.01 frames. ], batch size: 30, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:20:07,655 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 13:20:26,048 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12520.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:20:27,311 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0171, 2.4110, 0.8018, 1.3724, 1.4520, 1.6915, 1.6755, 0.8210], + device='cuda:2'), covar=tensor([0.1599, 0.1382, 0.1921, 0.1495, 0.1224, 0.1164, 0.1478, 0.1892], + device='cuda:2'), in_proj_covar=tensor([0.0125, 0.0270, 0.0152, 0.0132, 0.0143, 0.0166, 0.0129, 0.0133], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:21:02,678 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12550.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:03,403 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 13:21:12,628 INFO [finetune.py:976] (2/7) Epoch 3, batch 1100, loss[loss=0.2386, simple_loss=0.2924, pruned_loss=0.09247, over 4867.00 frames. ], tot_loss[loss=0.2431, simple_loss=0.2965, pruned_loss=0.09489, over 949924.07 frames. ], batch size: 34, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:21:25,845 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12576.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:44,439 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12604.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:46,054 INFO [finetune.py:976] (2/7) Epoch 3, batch 1150, loss[loss=0.2016, simple_loss=0.2536, pruned_loss=0.07478, over 4764.00 frames. ], tot_loss[loss=0.2428, simple_loss=0.297, pruned_loss=0.09426, over 950993.45 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:21:55,572 INFO [optim.py:369] (2/7) 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 13:21:58,104 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12624.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:21:58,761 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1946, 2.6139, 1.0602, 1.2862, 2.1687, 1.2961, 3.5099, 1.7294], + device='cuda:2'), covar=tensor([0.0679, 0.0700, 0.0896, 0.1383, 0.0530, 0.1037, 0.0252, 0.0695], + device='cuda:2'), in_proj_covar=tensor([0.0056, 0.0073, 0.0054, 0.0050, 0.0055, 0.0056, 0.0085, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 13:22:11,370 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5978, 1.5868, 3.6481, 3.3788, 3.2834, 3.4044, 3.3306, 3.2556], + device='cuda:2'), covar=tensor([0.6075, 0.4851, 0.1137, 0.1757, 0.1089, 0.1907, 0.3788, 0.1462], + device='cuda:2'), in_proj_covar=tensor([0.0326, 0.0310, 0.0439, 0.0445, 0.0375, 0.0426, 0.0336, 0.0394], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 13:22:19,232 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12640.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:22:34,206 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 1200, loss[loss=0.2226, simple_loss=0.2885, pruned_loss=0.0784, over 4899.00 frames. ], tot_loss[loss=0.2432, simple_loss=0.2969, pruned_loss=0.0947, over 948570.23 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:22:39,713 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6493, 1.3126, 1.6714, 1.8947, 1.7141, 1.5554, 1.6461, 1.6394], + device='cuda:2'), covar=tensor([2.7090, 3.7194, 4.3085, 4.7034, 2.9230, 4.4343, 4.3949, 3.4035], + device='cuda:2'), in_proj_covar=tensor([0.0450, 0.0505, 0.0598, 0.0604, 0.0483, 0.0519, 0.0531, 0.0543], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:22:57,551 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12688.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:23:04,640 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6468, 3.6263, 2.8024, 4.1928, 3.6498, 3.6438, 1.6838, 3.5050], + device='cuda:2'), covar=tensor([0.1637, 0.1140, 0.3488, 0.1585, 0.3049, 0.1783, 0.4988, 0.2321], + device='cuda:2'), in_proj_covar=tensor([0.0255, 0.0227, 0.0269, 0.0320, 0.0314, 0.0265, 0.0279, 0.0282], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:23:09,934 INFO [finetune.py:976] (2/7) Epoch 3, batch 1250, loss[loss=0.2548, simple_loss=0.2888, pruned_loss=0.1104, over 4267.00 frames. ], tot_loss[loss=0.2398, simple_loss=0.2935, pruned_loss=0.093, over 949888.99 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:23:19,406 INFO [optim.py:369] (2/7) 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 13:23:26,773 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 1300, loss[loss=0.2442, simple_loss=0.2917, pruned_loss=0.09834, over 4827.00 frames. ], tot_loss[loss=0.2371, simple_loss=0.2906, pruned_loss=0.09181, over 952304.64 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:23:47,612 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:24:09,814 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12780.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:24:33,586 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-26 13:24:35,997 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12802.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:24:39,383 INFO [finetune.py:976] (2/7) Epoch 3, batch 1350, loss[loss=0.2244, simple_loss=0.289, pruned_loss=0.07991, over 4812.00 frames. ], tot_loss[loss=0.2389, simple_loss=0.2922, pruned_loss=0.09282, over 949180.75 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:24:48,915 INFO [optim.py:369] (2/7) 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 13:24:51,944 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12824.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:25:12,234 INFO [finetune.py:976] (2/7) Epoch 3, batch 1400, loss[loss=0.2644, simple_loss=0.313, pruned_loss=0.1079, over 4911.00 frames. ], tot_loss[loss=0.241, simple_loss=0.2949, pruned_loss=0.09358, over 950722.40 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:25:27,716 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.43 vs. limit=5.0 +2023-04-26 13:25:56,156 INFO [finetune.py:976] (2/7) Epoch 3, batch 1450, loss[loss=0.2617, simple_loss=0.3108, pruned_loss=0.1063, over 4854.00 frames. ], tot_loss[loss=0.2432, simple_loss=0.297, pruned_loss=0.09468, over 949186.88 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:26:17,015 INFO [optim.py:369] (2/7) 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 13:27:02,918 INFO [finetune.py:976] (2/7) Epoch 3, batch 1500, loss[loss=0.1926, simple_loss=0.2593, pruned_loss=0.06294, over 4793.00 frames. ], tot_loss[loss=0.2443, simple_loss=0.2978, pruned_loss=0.09541, over 947795.75 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:27:17,302 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 13:27:26,253 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 13:27:40,485 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6085, 1.6047, 1.8316, 1.9233, 1.9227, 1.5649, 1.1946, 1.6834], + device='cuda:2'), covar=tensor([0.1080, 0.1302, 0.0723, 0.0683, 0.0717, 0.1156, 0.1214, 0.0740], + device='cuda:2'), in_proj_covar=tensor([0.0211, 0.0211, 0.0190, 0.0185, 0.0184, 0.0201, 0.0177, 0.0197], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:27:41,572 INFO [finetune.py:976] (2/7) Epoch 3, batch 1550, loss[loss=0.2517, simple_loss=0.2974, pruned_loss=0.103, over 4892.00 frames. ], tot_loss[loss=0.2429, simple_loss=0.2969, pruned_loss=0.09441, over 950560.88 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:28:02,731 INFO [optim.py:369] (2/7) 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 13:28:33,840 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-26 13:28:47,302 INFO [finetune.py:976] (2/7) Epoch 3, batch 1600, loss[loss=0.2106, simple_loss=0.2783, pruned_loss=0.07144, over 4832.00 frames. ], tot_loss[loss=0.2417, simple_loss=0.2953, pruned_loss=0.09403, over 951479.32 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:28:47,998 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8335, 1.2734, 5.1765, 4.8012, 4.5228, 4.8379, 4.5251, 4.5804], + device='cuda:2'), covar=tensor([0.7128, 0.6229, 0.1001, 0.1903, 0.1087, 0.1213, 0.1192, 0.1443], + device='cuda:2'), in_proj_covar=tensor([0.0326, 0.0310, 0.0439, 0.0443, 0.0374, 0.0425, 0.0337, 0.0391], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:29:10,873 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-04-26 13:29:11,324 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5818, 1.8608, 1.6164, 1.8415, 1.7477, 1.9563, 1.6216, 1.6443], + device='cuda:2'), covar=tensor([1.4206, 2.7095, 2.1403, 1.7654, 2.0230, 3.0927, 2.8552, 2.3897], + device='cuda:2'), in_proj_covar=tensor([0.0299, 0.0399, 0.0315, 0.0322, 0.0349, 0.0401, 0.0384, 0.0345], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:29:22,134 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 13:29:23,819 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13102.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:29:26,215 INFO [finetune.py:976] (2/7) Epoch 3, batch 1650, loss[loss=0.2135, simple_loss=0.2573, pruned_loss=0.08484, over 4236.00 frames. ], tot_loss[loss=0.239, simple_loss=0.2921, pruned_loss=0.09295, over 952401.88 frames. ], batch size: 18, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:29:31,058 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5370, 2.0347, 1.6554, 1.8930, 1.5708, 1.5550, 1.6478, 1.3940], + device='cuda:2'), covar=tensor([0.2442, 0.1673, 0.1158, 0.1593, 0.3629, 0.1825, 0.2026, 0.2940], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0341, 0.0249, 0.0312, 0.0319, 0.0292, 0.0281, 0.0304], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:29:34,729 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=13119.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:29:35,234 INFO [optim.py:369] (2/7) 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 13:29:44,653 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3169, 2.9517, 0.8427, 1.6738, 1.7455, 2.1928, 1.8624, 1.0157], + device='cuda:2'), covar=tensor([0.1340, 0.1083, 0.2032, 0.1356, 0.1036, 0.1024, 0.1470, 0.1853], + device='cuda:2'), in_proj_covar=tensor([0.0125, 0.0271, 0.0152, 0.0132, 0.0143, 0.0166, 0.0129, 0.0134], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:29:53,537 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4527, 1.0964, 3.8103, 3.4026, 3.4805, 3.5534, 3.5937, 3.2396], + device='cuda:2'), covar=tensor([0.8573, 0.7806, 0.1937, 0.3111, 0.1910, 0.2359, 0.2697, 0.3118], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0313, 0.0441, 0.0446, 0.0376, 0.0429, 0.0339, 0.0395], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 13:29:55,914 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13150.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:29:55,966 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6111, 1.5099, 0.5808, 1.2565, 1.6670, 1.4734, 1.3882, 1.4190], + device='cuda:2'), covar=tensor([0.0604, 0.0468, 0.0519, 0.0655, 0.0333, 0.0612, 0.0606, 0.0721], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0048, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 13:29:59,490 INFO [finetune.py:976] (2/7) Epoch 3, batch 1700, loss[loss=0.2559, simple_loss=0.3038, pruned_loss=0.104, over 4908.00 frames. ], tot_loss[loss=0.2387, simple_loss=0.291, pruned_loss=0.0932, over 953628.74 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:30:00,204 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2988, 3.2940, 2.7114, 2.7957, 2.3289, 2.5701, 2.7139, 2.2237], + device='cuda:2'), covar=tensor([0.2941, 0.1672, 0.1041, 0.1659, 0.2992, 0.1736, 0.2301, 0.3424], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0341, 0.0249, 0.0312, 0.0319, 0.0292, 0.0281, 0.0304], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:30:33,505 INFO [finetune.py:976] (2/7) Epoch 3, batch 1750, loss[loss=0.2894, simple_loss=0.3511, pruned_loss=0.1138, over 4809.00 frames. ], tot_loss[loss=0.2391, simple_loss=0.2917, pruned_loss=0.09323, over 952263.99 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 64.0 +2023-04-26 13:30:43,131 INFO [optim.py:369] (2/7) 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 13:31:07,408 INFO [finetune.py:976] (2/7) Epoch 3, batch 1800, loss[loss=0.2422, simple_loss=0.3137, pruned_loss=0.08535, over 4815.00 frames. ], tot_loss[loss=0.24, simple_loss=0.2943, pruned_loss=0.09286, over 952308.03 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 13:31:20,365 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-04-26 13:31:47,121 INFO [finetune.py:976] (2/7) Epoch 3, batch 1850, loss[loss=0.3311, simple_loss=0.3706, pruned_loss=0.1458, over 4926.00 frames. ], tot_loss[loss=0.2416, simple_loss=0.2957, pruned_loss=0.09371, over 952065.98 frames. ], batch size: 42, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:31:56,813 INFO [optim.py:369] (2/7) 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 13:32:30,328 INFO [finetune.py:976] (2/7) Epoch 3, batch 1900, loss[loss=0.3007, simple_loss=0.3488, pruned_loss=0.1263, over 4241.00 frames. ], tot_loss[loss=0.243, simple_loss=0.2979, pruned_loss=0.09404, over 953348.80 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:32:38,717 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8518, 2.6993, 1.8406, 1.7167, 1.3155, 1.3841, 1.9383, 1.3295], + device='cuda:2'), covar=tensor([0.2469, 0.2125, 0.2495, 0.3160, 0.3617, 0.2749, 0.1925, 0.3011], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0225, 0.0191, 0.0216, 0.0229, 0.0195, 0.0186, 0.0206], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:33:26,629 INFO [finetune.py:976] (2/7) Epoch 3, batch 1950, loss[loss=0.2143, simple_loss=0.2645, pruned_loss=0.08204, over 4907.00 frames. ], tot_loss[loss=0.2401, simple_loss=0.295, pruned_loss=0.09262, over 952512.77 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:33:28,622 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1606, 1.5577, 1.4113, 1.8357, 1.6216, 2.1818, 1.4482, 3.6795], + device='cuda:2'), covar=tensor([0.0782, 0.0807, 0.0823, 0.1238, 0.0701, 0.0548, 0.0764, 0.0156], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 13:33:39,969 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13419.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:33:41,755 INFO [optim.py:369] (2/7) 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 13:34:17,441 INFO [finetune.py:976] (2/7) Epoch 3, batch 2000, loss[loss=0.2386, simple_loss=0.2859, pruned_loss=0.09562, over 4904.00 frames. ], tot_loss[loss=0.2373, simple_loss=0.2915, pruned_loss=0.0916, over 954576.15 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:34:24,602 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:34:25,243 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1125, 1.6659, 1.4735, 2.1255, 1.9411, 2.2466, 1.5576, 4.0223], + device='cuda:2'), covar=tensor([0.0696, 0.0687, 0.0772, 0.1087, 0.0590, 0.0547, 0.0695, 0.0127], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 13:34:30,212 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-26 13:34:51,089 INFO [finetune.py:976] (2/7) Epoch 3, batch 2050, loss[loss=0.216, simple_loss=0.2774, pruned_loss=0.07732, over 4820.00 frames. ], tot_loss[loss=0.236, simple_loss=0.2894, pruned_loss=0.0913, over 955342.65 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:34:51,855 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2319, 1.5230, 1.4066, 1.5068, 1.4144, 1.6082, 1.5774, 1.5382], + device='cuda:2'), covar=tensor([1.3845, 2.3364, 1.8502, 1.5994, 1.9139, 3.0152, 2.2614, 1.9910], + device='cuda:2'), in_proj_covar=tensor([0.0297, 0.0397, 0.0313, 0.0318, 0.0346, 0.0398, 0.0381, 0.0341], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:35:01,249 INFO [optim.py:369] (2/7) 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 13:35:17,957 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4886, 1.7771, 1.5478, 1.6850, 1.5652, 1.8057, 1.6673, 1.6316], + device='cuda:2'), covar=tensor([1.2849, 2.2990, 1.9660, 1.6558, 1.9159, 2.9583, 2.4349, 2.2241], + device='cuda:2'), in_proj_covar=tensor([0.0297, 0.0397, 0.0313, 0.0318, 0.0346, 0.0399, 0.0381, 0.0341], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:35:24,229 INFO [finetune.py:976] (2/7) Epoch 3, batch 2100, loss[loss=0.254, simple_loss=0.3186, pruned_loss=0.0947, over 4851.00 frames. ], tot_loss[loss=0.2373, simple_loss=0.2902, pruned_loss=0.09216, over 957577.59 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:35:26,158 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8841, 1.3097, 1.8359, 2.2221, 1.6219, 1.3668, 0.9890, 1.4598], + device='cuda:2'), covar=tensor([0.5333, 0.6098, 0.2811, 0.4584, 0.6026, 0.4619, 0.8514, 0.5696], + device='cuda:2'), in_proj_covar=tensor([0.0270, 0.0273, 0.0223, 0.0343, 0.0231, 0.0235, 0.0263, 0.0209], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:35:57,669 INFO [finetune.py:976] (2/7) Epoch 3, batch 2150, loss[loss=0.2474, simple_loss=0.3045, pruned_loss=0.09513, over 4808.00 frames. ], tot_loss[loss=0.2391, simple_loss=0.2926, pruned_loss=0.09276, over 954984.90 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:35:59,574 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1895, 1.4635, 1.1267, 1.4091, 1.3155, 1.1383, 1.2323, 1.0154], + device='cuda:2'), covar=tensor([0.2287, 0.1721, 0.1434, 0.1728, 0.4057, 0.1894, 0.2074, 0.2925], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0342, 0.0250, 0.0313, 0.0321, 0.0294, 0.0282, 0.0305], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:36:07,902 INFO [optim.py:369] (2/7) 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 13:36:30,853 INFO [finetune.py:976] (2/7) Epoch 3, batch 2200, loss[loss=0.2516, simple_loss=0.2955, pruned_loss=0.1039, over 4798.00 frames. ], tot_loss[loss=0.2408, simple_loss=0.2951, pruned_loss=0.09326, over 955789.32 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:37:21,693 INFO [finetune.py:976] (2/7) Epoch 3, batch 2250, loss[loss=0.2577, simple_loss=0.3175, pruned_loss=0.09893, over 4903.00 frames. ], tot_loss[loss=0.2414, simple_loss=0.2959, pruned_loss=0.09346, over 954767.62 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:37:31,818 INFO [optim.py:369] (2/7) 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 13:37:43,769 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=13737.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:38:01,333 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 13:38:02,150 INFO [finetune.py:976] (2/7) Epoch 3, batch 2300, loss[loss=0.2394, simple_loss=0.2976, pruned_loss=0.09059, over 4898.00 frames. ], tot_loss[loss=0.2422, simple_loss=0.297, pruned_loss=0.09366, over 954492.38 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:38:57,775 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=13798.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:39:06,933 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3781, 2.3590, 2.6338, 2.7742, 2.7981, 2.1130, 1.7630, 2.3374], + device='cuda:2'), covar=tensor([0.1177, 0.0948, 0.0560, 0.0841, 0.0657, 0.1258, 0.1266, 0.0757], + device='cuda:2'), in_proj_covar=tensor([0.0210, 0.0209, 0.0189, 0.0184, 0.0182, 0.0200, 0.0175, 0.0196], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:39:09,162 INFO [finetune.py:976] (2/7) Epoch 3, batch 2350, loss[loss=0.1964, simple_loss=0.2573, pruned_loss=0.0678, over 4911.00 frames. ], tot_loss[loss=0.2389, simple_loss=0.2934, pruned_loss=0.09218, over 955211.83 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:39:37,984 INFO [optim.py:369] (2/7) 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 13:40:20,272 INFO [finetune.py:976] (2/7) Epoch 3, batch 2400, loss[loss=0.2375, simple_loss=0.2884, pruned_loss=0.09325, over 4756.00 frames. ], tot_loss[loss=0.2372, simple_loss=0.2911, pruned_loss=0.09165, over 957235.43 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:40:37,263 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 13:40:39,683 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 13:40:57,680 INFO [finetune.py:976] (2/7) Epoch 3, batch 2450, loss[loss=0.2758, simple_loss=0.3226, pruned_loss=0.1145, over 3998.00 frames. ], tot_loss[loss=0.236, simple_loss=0.2888, pruned_loss=0.09163, over 953265.60 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:41:09,355 INFO [optim.py:369] (2/7) 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 13:41:11,227 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 13:41:17,921 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8988, 2.0548, 1.7499, 1.6401, 2.1571, 1.6354, 2.7064, 1.5875], + device='cuda:2'), covar=tensor([0.4930, 0.1777, 0.5904, 0.3568, 0.1969, 0.3106, 0.1324, 0.4756], + device='cuda:2'), in_proj_covar=tensor([0.0352, 0.0354, 0.0436, 0.0372, 0.0404, 0.0381, 0.0398, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:41:31,085 INFO [finetune.py:976] (2/7) Epoch 3, batch 2500, loss[loss=0.2164, simple_loss=0.2884, pruned_loss=0.07217, over 4903.00 frames. ], tot_loss[loss=0.2373, simple_loss=0.2906, pruned_loss=0.09205, over 954069.90 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:42:05,891 INFO [finetune.py:976] (2/7) Epoch 3, batch 2550, loss[loss=0.247, simple_loss=0.3084, pruned_loss=0.09283, over 4916.00 frames. ], tot_loss[loss=0.2399, simple_loss=0.2943, pruned_loss=0.09275, over 954183.44 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:42:28,116 INFO [optim.py:369] (2/7) 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 13:42:46,177 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 13:43:13,249 INFO [finetune.py:976] (2/7) Epoch 3, batch 2600, loss[loss=0.2262, simple_loss=0.2885, pruned_loss=0.08196, over 4820.00 frames. ], tot_loss[loss=0.2411, simple_loss=0.2959, pruned_loss=0.0931, over 954394.34 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:43:37,588 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14072.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:44:06,886 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14093.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:44:07,495 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4283, 3.2709, 0.8769, 1.6761, 1.7394, 2.3273, 1.9379, 1.0029], + device='cuda:2'), covar=tensor([0.1360, 0.0962, 0.2201, 0.1513, 0.1110, 0.1051, 0.1483, 0.1958], + device='cuda:2'), in_proj_covar=tensor([0.0124, 0.0268, 0.0150, 0.0131, 0.0142, 0.0164, 0.0127, 0.0132], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:44:18,993 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14103.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:44:25,130 INFO [finetune.py:976] (2/7) Epoch 3, batch 2650, loss[loss=0.2424, simple_loss=0.296, pruned_loss=0.09444, over 4805.00 frames. ], tot_loss[loss=0.2406, simple_loss=0.2959, pruned_loss=0.09269, over 952713.91 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:44:39,918 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14120.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 13:44:47,127 INFO [optim.py:369] (2/7) 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 13:45:00,929 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14133.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:45:30,357 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7017, 1.4337, 1.9674, 1.9406, 1.4960, 1.3091, 1.7351, 1.2232], + device='cuda:2'), covar=tensor([0.0779, 0.1092, 0.0658, 0.0904, 0.1167, 0.1483, 0.0820, 0.1243], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0079, 0.0076, 0.0071, 0.0084, 0.0097, 0.0088, 0.0081], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 13:45:30,810 INFO [finetune.py:976] (2/7) Epoch 3, batch 2700, loss[loss=0.2385, simple_loss=0.2894, pruned_loss=0.09383, over 4851.00 frames. ], tot_loss[loss=0.24, simple_loss=0.2949, pruned_loss=0.09258, over 949093.92 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:45:41,204 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14164.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:45:44,929 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2532, 1.5555, 1.4159, 1.5537, 1.4674, 1.7132, 1.5649, 1.5415], + device='cuda:2'), covar=tensor([1.4078, 2.2722, 1.8995, 1.5746, 1.8584, 2.8558, 2.3781, 1.9477], + device='cuda:2'), in_proj_covar=tensor([0.0300, 0.0400, 0.0315, 0.0321, 0.0348, 0.0402, 0.0383, 0.0342], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:46:03,632 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14181.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 13:46:30,712 INFO [finetune.py:976] (2/7) Epoch 3, batch 2750, loss[loss=0.2039, simple_loss=0.2653, pruned_loss=0.07126, over 4761.00 frames. ], tot_loss[loss=0.2379, simple_loss=0.292, pruned_loss=0.09194, over 950900.28 frames. ], batch size: 27, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:46:40,382 INFO [optim.py:369] (2/7) 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 13:46:49,385 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 13:47:03,042 INFO [finetune.py:976] (2/7) Epoch 3, batch 2800, loss[loss=0.2309, simple_loss=0.2756, pruned_loss=0.0931, over 4851.00 frames. ], tot_loss[loss=0.233, simple_loss=0.2873, pruned_loss=0.08936, over 952263.22 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:47:33,884 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-26 13:47:35,420 INFO [finetune.py:976] (2/7) Epoch 3, batch 2850, loss[loss=0.2335, simple_loss=0.2944, pruned_loss=0.08629, over 4830.00 frames. ], tot_loss[loss=0.2322, simple_loss=0.286, pruned_loss=0.08918, over 952953.01 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:47:40,090 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4783, 2.8427, 0.9286, 1.4986, 2.2595, 1.4240, 4.2108, 2.1192], + device='cuda:2'), covar=tensor([0.0666, 0.0893, 0.1001, 0.1397, 0.0610, 0.1096, 0.0224, 0.0647], + device='cuda:2'), in_proj_covar=tensor([0.0055, 0.0072, 0.0053, 0.0050, 0.0055, 0.0056, 0.0084, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 13:47:44,494 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-04-26 13:47:45,429 INFO [optim.py:369] (2/7) 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 13:48:08,416 INFO [finetune.py:976] (2/7) Epoch 3, batch 2900, loss[loss=0.1629, simple_loss=0.2131, pruned_loss=0.05635, over 4695.00 frames. ], tot_loss[loss=0.2319, simple_loss=0.2867, pruned_loss=0.08857, over 951414.71 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:48:33,680 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14393.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:48:41,651 INFO [finetune.py:976] (2/7) Epoch 3, batch 2950, loss[loss=0.2647, simple_loss=0.3175, pruned_loss=0.1059, over 4897.00 frames. ], tot_loss[loss=0.2351, simple_loss=0.2903, pruned_loss=0.08994, over 951593.35 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:48:46,072 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14441.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:49:14,495 INFO [finetune.py:976] (2/7) Epoch 3, batch 3000, loss[loss=0.2094, simple_loss=0.2759, pruned_loss=0.07143, over 4835.00 frames. ], tot_loss[loss=0.2367, simple_loss=0.292, pruned_loss=0.0907, over 954237.16 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:49:14,495 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 13:49:25,021 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6274MB +2023-04-26 13:49:26,932 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14459.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:49:36,024 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14474.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:49:37,147 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14476.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 13:49:44,329 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14487.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:50:13,198 INFO [finetune.py:976] (2/7) Epoch 3, batch 3050, loss[loss=0.2552, simple_loss=0.3187, pruned_loss=0.09587, over 4760.00 frames. ], tot_loss[loss=0.2367, simple_loss=0.2922, pruned_loss=0.09059, over 953809.43 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:50:24,171 INFO [optim.py:369] (2/7) 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 13:50:28,727 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 13:50:52,326 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14548.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:51:03,205 INFO [finetune.py:976] (2/7) Epoch 3, batch 3100, loss[loss=0.2151, simple_loss=0.2691, pruned_loss=0.08057, over 4893.00 frames. ], tot_loss[loss=0.2361, simple_loss=0.2911, pruned_loss=0.09052, over 953302.84 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:51:15,975 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.72 vs. limit=2.0 +2023-04-26 13:52:09,675 INFO [finetune.py:976] (2/7) Epoch 3, batch 3150, loss[loss=0.2391, simple_loss=0.2935, pruned_loss=0.09236, over 4873.00 frames. ], tot_loss[loss=0.2342, simple_loss=0.2889, pruned_loss=0.0898, over 952805.27 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:52:30,356 INFO [optim.py:369] (2/7) 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 13:52:41,597 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-04-26 13:52:51,772 INFO [finetune.py:976] (2/7) Epoch 3, batch 3200, loss[loss=0.2336, simple_loss=0.2975, pruned_loss=0.08482, over 4786.00 frames. ], tot_loss[loss=0.2301, simple_loss=0.2847, pruned_loss=0.08773, over 952804.37 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:53:25,510 INFO [finetune.py:976] (2/7) Epoch 3, batch 3250, loss[loss=0.2857, simple_loss=0.3348, pruned_loss=0.1183, over 4808.00 frames. ], tot_loss[loss=0.2328, simple_loss=0.2874, pruned_loss=0.0891, over 954105.26 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:53:37,682 INFO [optim.py:369] (2/7) 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 13:53:41,982 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14728.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:53:49,332 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14740.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:53:59,444 INFO [finetune.py:976] (2/7) Epoch 3, batch 3300, loss[loss=0.2731, simple_loss=0.3353, pruned_loss=0.1055, over 4847.00 frames. ], tot_loss[loss=0.2368, simple_loss=0.2919, pruned_loss=0.09087, over 955748.55 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:54:01,389 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14776.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:54:14,699 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14776.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 13:54:30,601 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 3350, loss[loss=0.2514, simple_loss=0.2999, pruned_loss=0.1015, over 4817.00 frames. ], tot_loss[loss=0.238, simple_loss=0.2934, pruned_loss=0.09133, over 954409.48 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:54:34,181 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14807.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:54:44,575 INFO [optim.py:369] (2/7) 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 13:54:46,351 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14824.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 13:54:48,666 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6891, 2.4789, 1.6064, 1.6443, 1.2345, 1.3077, 1.7431, 1.1988], + device='cuda:2'), covar=tensor([0.2181, 0.1727, 0.2312, 0.2596, 0.3440, 0.2672, 0.1584, 0.2784], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0223, 0.0188, 0.0213, 0.0226, 0.0193, 0.0182, 0.0203], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:54:49,792 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3478, 3.3093, 2.5916, 3.8730, 3.3943, 3.3828, 1.3676, 3.2646], + device='cuda:2'), covar=tensor([0.2004, 0.1372, 0.3153, 0.2180, 0.2597, 0.1994, 0.6139, 0.2487], + device='cuda:2'), in_proj_covar=tensor([0.0257, 0.0229, 0.0270, 0.0321, 0.0315, 0.0265, 0.0283, 0.0282], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:54:55,347 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-26 13:54:58,941 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14843.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:55:13,040 INFO [finetune.py:976] (2/7) Epoch 3, batch 3400, loss[loss=0.2269, simple_loss=0.2865, pruned_loss=0.08369, over 4832.00 frames. ], tot_loss[loss=0.2375, simple_loss=0.2937, pruned_loss=0.09069, over 954586.02 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:56:02,500 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8744, 1.8616, 1.5978, 1.4622, 1.8737, 1.4628, 2.1911, 1.3842], + device='cuda:2'), covar=tensor([0.3607, 0.1276, 0.4321, 0.2614, 0.1552, 0.2217, 0.1496, 0.4069], + device='cuda:2'), in_proj_covar=tensor([0.0357, 0.0361, 0.0441, 0.0376, 0.0409, 0.0387, 0.0404, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:56:08,499 INFO [finetune.py:976] (2/7) Epoch 3, batch 3450, loss[loss=0.2086, simple_loss=0.2639, pruned_loss=0.07667, over 4820.00 frames. ], tot_loss[loss=0.2362, simple_loss=0.2929, pruned_loss=0.08971, over 956220.98 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:56:18,850 INFO [optim.py:369] (2/7) 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 13:56:42,394 INFO [finetune.py:976] (2/7) Epoch 3, batch 3500, loss[loss=0.2229, simple_loss=0.2807, pruned_loss=0.08255, over 4825.00 frames. ], tot_loss[loss=0.2329, simple_loss=0.2893, pruned_loss=0.08825, over 956084.66 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:57:16,912 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7357, 2.0762, 1.5657, 1.3096, 1.3275, 1.3356, 1.5330, 1.2805], + device='cuda:2'), covar=tensor([0.1876, 0.1792, 0.2045, 0.2473, 0.3181, 0.2345, 0.1601, 0.2539], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0222, 0.0187, 0.0212, 0.0225, 0.0191, 0.0181, 0.0202], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 13:57:38,374 INFO [finetune.py:976] (2/7) Epoch 3, batch 3550, loss[loss=0.2187, simple_loss=0.2772, pruned_loss=0.08007, over 4737.00 frames. ], tot_loss[loss=0.2308, simple_loss=0.2865, pruned_loss=0.08749, over 955408.23 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:57:54,036 INFO [optim.py:369] (2/7) 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 13:58:13,430 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4582, 1.9879, 2.5160, 2.7915, 1.9831, 1.7282, 2.2431, 1.3100], + device='cuda:2'), covar=tensor([0.0716, 0.1339, 0.0656, 0.1111, 0.1246, 0.1565, 0.1089, 0.1379], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0079, 0.0076, 0.0071, 0.0084, 0.0097, 0.0087, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 13:58:25,886 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0177, 1.5806, 1.9540, 2.3681, 1.8657, 1.5062, 1.4154, 1.7935], + device='cuda:2'), covar=tensor([0.3430, 0.4539, 0.1930, 0.3105, 0.4331, 0.3520, 0.6409, 0.4095], + device='cuda:2'), in_proj_covar=tensor([0.0272, 0.0271, 0.0223, 0.0341, 0.0230, 0.0234, 0.0261, 0.0208], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 13:58:28,220 INFO [finetune.py:976] (2/7) Epoch 3, batch 3600, loss[loss=0.2177, simple_loss=0.2798, pruned_loss=0.07776, over 4734.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2841, pruned_loss=0.08631, over 956911.66 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:58:35,225 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-26 13:58:36,855 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15069.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:58:56,112 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 3650, loss[loss=0.2611, simple_loss=0.3265, pruned_loss=0.09787, over 4856.00 frames. ], tot_loss[loss=0.2303, simple_loss=0.2864, pruned_loss=0.08713, over 957777.01 frames. ], batch size: 44, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:59:09,046 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15117.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:59:12,573 INFO [optim.py:369] (2/7) 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 13:59:17,636 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3921, 1.1542, 1.4774, 1.7161, 1.5883, 1.3721, 1.4581, 1.4388], + device='cuda:2'), covar=tensor([2.0380, 2.5802, 2.8716, 3.3409, 2.1659, 3.2211, 3.1654, 2.4884], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0498, 0.0590, 0.0599, 0.0480, 0.0512, 0.0525, 0.0534], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 13:59:28,016 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15143.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 13:59:36,235 INFO [finetune.py:976] (2/7) Epoch 3, batch 3700, loss[loss=0.2228, simple_loss=0.286, pruned_loss=0.07977, over 4919.00 frames. ], tot_loss[loss=0.2329, simple_loss=0.2893, pruned_loss=0.08828, over 955373.70 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 13:59:45,424 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5372, 1.3044, 1.8991, 1.7905, 1.4103, 1.1844, 1.5946, 1.0354], + device='cuda:2'), covar=tensor([0.0977, 0.1103, 0.0583, 0.1176, 0.1305, 0.1608, 0.0963, 0.1325], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0079, 0.0076, 0.0071, 0.0084, 0.0097, 0.0087, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 13:59:59,374 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15191.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:00:02,737 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15195.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:00:10,003 INFO [finetune.py:976] (2/7) Epoch 3, batch 3750, loss[loss=0.2394, simple_loss=0.2984, pruned_loss=0.09021, over 4744.00 frames. ], tot_loss[loss=0.2353, simple_loss=0.2915, pruned_loss=0.08954, over 956090.23 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 14:00:24,837 INFO [optim.py:369] (2/7) 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 14:00:55,181 INFO [finetune.py:976] (2/7) Epoch 3, batch 3800, loss[loss=0.2355, simple_loss=0.2932, pruned_loss=0.0889, over 4885.00 frames. ], tot_loss[loss=0.2364, simple_loss=0.2929, pruned_loss=0.08994, over 955713.16 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 16.0 +2023-04-26 14:00:55,313 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15256.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 14:01:29,009 INFO [finetune.py:976] (2/7) Epoch 3, batch 3850, loss[loss=0.215, simple_loss=0.2743, pruned_loss=0.07785, over 4848.00 frames. ], tot_loss[loss=0.235, simple_loss=0.2914, pruned_loss=0.08928, over 956096.31 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:01:38,746 INFO [optim.py:369] (2/7) 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 14:01:57,731 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7007, 4.0676, 0.7743, 2.0936, 2.3087, 2.4690, 2.3926, 0.9230], + device='cuda:2'), covar=tensor([0.1398, 0.0913, 0.2353, 0.1392, 0.1064, 0.1305, 0.1428, 0.2201], + device='cuda:2'), in_proj_covar=tensor([0.0124, 0.0269, 0.0150, 0.0131, 0.0141, 0.0165, 0.0128, 0.0132], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 14:01:58,365 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4458, 2.3764, 1.9623, 2.1705, 2.4553, 2.0320, 3.1000, 1.7525], + device='cuda:2'), covar=tensor([0.4666, 0.2208, 0.5114, 0.3669, 0.2064, 0.2802, 0.2240, 0.4643], + device='cuda:2'), in_proj_covar=tensor([0.0356, 0.0359, 0.0442, 0.0375, 0.0409, 0.0384, 0.0405, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:02:01,731 INFO [finetune.py:976] (2/7) Epoch 3, batch 3900, loss[loss=0.1816, simple_loss=0.2528, pruned_loss=0.05514, over 4780.00 frames. ], tot_loss[loss=0.2314, simple_loss=0.2878, pruned_loss=0.08747, over 958135.00 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:02:50,596 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 3950, loss[loss=0.2131, simple_loss=0.2763, pruned_loss=0.0749, over 4805.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2854, pruned_loss=0.08714, over 954159.34 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:03:24,419 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15444.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:04:08,305 INFO [finetune.py:976] (2/7) Epoch 3, batch 4000, loss[loss=0.1922, simple_loss=0.2537, pruned_loss=0.06532, over 4755.00 frames. ], tot_loss[loss=0.2286, simple_loss=0.2836, pruned_loss=0.08685, over 953101.23 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:05:04,586 INFO [finetune.py:976] (2/7) Epoch 3, batch 4050, loss[loss=0.2092, simple_loss=0.2527, pruned_loss=0.08289, over 4717.00 frames. ], tot_loss[loss=0.2329, simple_loss=0.2875, pruned_loss=0.08915, over 951664.07 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:05:08,748 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0142, 1.1570, 1.4049, 1.5902, 1.5187, 1.6879, 1.4177, 1.4760], + device='cuda:2'), covar=tensor([1.1298, 1.8481, 1.5540, 1.3326, 1.5006, 2.3599, 1.8480, 1.6671], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0403, 0.0319, 0.0324, 0.0351, 0.0408, 0.0388, 0.0345], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:05:14,143 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-26 14:05:15,729 INFO [optim.py:369] (2/7) 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 14:05:30,517 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 14:05:33,870 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15551.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:05:36,858 INFO [finetune.py:976] (2/7) Epoch 3, batch 4100, loss[loss=0.2718, simple_loss=0.3216, pruned_loss=0.111, over 4897.00 frames. ], tot_loss[loss=0.2333, simple_loss=0.2889, pruned_loss=0.08882, over 952898.63 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:05:49,441 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3071, 1.5148, 1.5191, 1.6390, 1.5694, 1.7128, 1.6021, 1.6161], + device='cuda:2'), covar=tensor([1.2072, 1.9634, 1.7105, 1.3643, 1.6180, 2.5938, 2.0233, 1.7470], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0404, 0.0320, 0.0325, 0.0352, 0.0409, 0.0390, 0.0346], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:06:07,136 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6883, 1.1816, 1.4472, 1.6186, 1.3506, 1.1187, 0.5979, 1.0833], + device='cuda:2'), covar=tensor([0.4719, 0.5618, 0.2571, 0.3737, 0.4821, 0.4276, 0.7085, 0.4455], + device='cuda:2'), in_proj_covar=tensor([0.0273, 0.0272, 0.0224, 0.0342, 0.0230, 0.0235, 0.0260, 0.0208], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:06:10,674 INFO [finetune.py:976] (2/7) Epoch 3, batch 4150, loss[loss=0.2757, simple_loss=0.3289, pruned_loss=0.1113, over 4822.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.2911, pruned_loss=0.08928, over 954747.23 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:06:14,773 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1237, 1.4243, 1.9593, 2.6243, 1.8531, 1.3950, 1.3377, 1.8661], + device='cuda:2'), covar=tensor([0.5160, 0.5987, 0.2676, 0.4823, 0.5750, 0.4524, 0.7187, 0.5110], + device='cuda:2'), in_proj_covar=tensor([0.0273, 0.0272, 0.0224, 0.0342, 0.0230, 0.0235, 0.0260, 0.0208], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:06:16,479 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3729, 1.0902, 1.3780, 1.7247, 1.5448, 1.3626, 1.4313, 1.3840], + device='cuda:2'), covar=tensor([1.7866, 2.5518, 2.7620, 3.0128, 1.9489, 2.9504, 3.0105, 2.3876], + device='cuda:2'), in_proj_covar=tensor([0.0445, 0.0495, 0.0585, 0.0597, 0.0478, 0.0508, 0.0520, 0.0530], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:06:23,087 INFO [optim.py:369] (2/7) 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 14:06:34,253 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-04-26 14:06:44,251 INFO [finetune.py:976] (2/7) Epoch 3, batch 4200, loss[loss=0.2243, simple_loss=0.2818, pruned_loss=0.08342, over 4763.00 frames. ], tot_loss[loss=0.2336, simple_loss=0.2903, pruned_loss=0.08849, over 954585.23 frames. ], batch size: 28, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:07:17,990 INFO [finetune.py:976] (2/7) Epoch 3, batch 4250, loss[loss=0.2529, simple_loss=0.3007, pruned_loss=0.1025, over 4783.00 frames. ], tot_loss[loss=0.2327, simple_loss=0.2889, pruned_loss=0.08826, over 954719.66 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:07:27,379 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8325, 2.2631, 1.8398, 2.1441, 1.6545, 1.7292, 1.9252, 1.5386], + device='cuda:2'), covar=tensor([0.1986, 0.1108, 0.0994, 0.1182, 0.3008, 0.1537, 0.1830, 0.2603], + device='cuda:2'), in_proj_covar=tensor([0.0319, 0.0337, 0.0248, 0.0310, 0.0321, 0.0291, 0.0280, 0.0302], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:07:29,538 INFO [optim.py:369] (2/7) 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 14:07:50,583 INFO [finetune.py:976] (2/7) Epoch 3, batch 4300, loss[loss=0.2242, simple_loss=0.2807, pruned_loss=0.08383, over 4911.00 frames. ], tot_loss[loss=0.2301, simple_loss=0.2858, pruned_loss=0.08713, over 953753.68 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:08:40,062 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6644, 1.3138, 1.6352, 1.9018, 1.8084, 1.6195, 1.6871, 1.6489], + device='cuda:2'), covar=tensor([1.8757, 2.4927, 2.9154, 3.1810, 2.0317, 3.0144, 3.0601, 2.2844], + device='cuda:2'), in_proj_covar=tensor([0.0446, 0.0494, 0.0584, 0.0596, 0.0476, 0.0507, 0.0520, 0.0529], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:08:59,709 INFO [finetune.py:976] (2/7) Epoch 3, batch 4350, loss[loss=0.2025, simple_loss=0.2531, pruned_loss=0.07599, over 4812.00 frames. ], tot_loss[loss=0.2254, simple_loss=0.2813, pruned_loss=0.08472, over 953213.75 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 32.0 +2023-04-26 14:09:21,815 INFO [optim.py:369] (2/7) 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 14:10:01,712 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15851.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:10:04,714 INFO [finetune.py:976] (2/7) Epoch 3, batch 4400, loss[loss=0.2835, simple_loss=0.3339, pruned_loss=0.1165, over 4798.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2838, pruned_loss=0.0864, over 953733.47 frames. ], batch size: 45, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:10:05,624 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 14:10:39,587 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 3, batch 4450, loss[loss=0.2237, simple_loss=0.2934, pruned_loss=0.07699, over 4838.00 frames. ], tot_loss[loss=0.2327, simple_loss=0.2884, pruned_loss=0.08849, over 954440.82 frames. ], batch size: 47, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:10:59,475 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-04-26 14:11:04,060 INFO [optim.py:369] (2/7) 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 14:11:12,875 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4357, 2.3086, 1.9865, 2.1648, 2.4705, 2.0509, 3.3778, 1.8827], + device='cuda:2'), covar=tensor([0.5285, 0.2402, 0.5370, 0.4510, 0.2538, 0.3263, 0.1602, 0.4575], + device='cuda:2'), in_proj_covar=tensor([0.0358, 0.0360, 0.0443, 0.0378, 0.0412, 0.0386, 0.0405, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:11:20,092 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15944.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:11:22,480 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2257, 2.9937, 1.1290, 1.3760, 2.3390, 1.2325, 3.8161, 1.7271], + device='cuda:2'), covar=tensor([0.0645, 0.0859, 0.0960, 0.1276, 0.0492, 0.1003, 0.0203, 0.0631], + device='cuda:2'), in_proj_covar=tensor([0.0055, 0.0072, 0.0053, 0.0050, 0.0055, 0.0055, 0.0084, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 14:11:27,286 INFO [finetune.py:976] (2/7) Epoch 3, batch 4500, loss[loss=0.2124, simple_loss=0.2461, pruned_loss=0.08934, over 4095.00 frames. ], tot_loss[loss=0.2335, simple_loss=0.2893, pruned_loss=0.08886, over 952136.46 frames. ], batch size: 17, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:11:34,158 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0074, 1.4029, 1.3861, 1.8449, 2.1073, 1.8072, 1.7923, 1.4616], + device='cuda:2'), covar=tensor([0.2111, 0.1986, 0.2004, 0.2057, 0.1505, 0.2285, 0.2235, 0.1880], + device='cuda:2'), in_proj_covar=tensor([0.0319, 0.0338, 0.0350, 0.0312, 0.0350, 0.0361, 0.0320, 0.0353], + device='cuda:2'), out_proj_covar=tensor([6.9300e-05, 7.2568e-05, 7.5973e-05, 6.5314e-05, 7.4260e-05, 7.8730e-05, + 6.9619e-05, 7.6228e-05], device='cuda:2') +2023-04-26 14:11:34,748 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7491, 1.6569, 1.9374, 2.1925, 2.1207, 1.6029, 1.2961, 1.7817], + device='cuda:2'), covar=tensor([0.0860, 0.1037, 0.0624, 0.0557, 0.0631, 0.1010, 0.1054, 0.0652], + device='cuda:2'), in_proj_covar=tensor([0.0209, 0.0210, 0.0188, 0.0184, 0.0183, 0.0199, 0.0174, 0.0194], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:11:43,097 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8540, 1.7608, 2.0639, 2.2631, 2.2049, 1.6278, 1.3728, 1.8358], + device='cuda:2'), covar=tensor([0.1016, 0.1184, 0.0681, 0.0659, 0.0723, 0.1178, 0.1229, 0.0757], + device='cuda:2'), in_proj_covar=tensor([0.0210, 0.0210, 0.0188, 0.0184, 0.0183, 0.0200, 0.0174, 0.0194], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:11:44,351 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7666, 1.3485, 1.3374, 1.4370, 2.0247, 1.6491, 1.3686, 1.3077], + device='cuda:2'), covar=tensor([0.1463, 0.1587, 0.2026, 0.1439, 0.0793, 0.1571, 0.2158, 0.1957], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0339, 0.0351, 0.0313, 0.0351, 0.0361, 0.0321, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.9410e-05, 7.2734e-05, 7.6134e-05, 6.5532e-05, 7.4458e-05, 7.8822e-05, + 6.9783e-05, 7.6383e-05], device='cuda:2') +2023-04-26 14:11:53,584 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5800, 1.7274, 1.5498, 1.7232, 1.5587, 1.7938, 1.6431, 1.5965], + device='cuda:2'), covar=tensor([1.3809, 2.3153, 1.8751, 1.4996, 1.8112, 2.7437, 2.5180, 2.0507], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0405, 0.0320, 0.0326, 0.0353, 0.0411, 0.0390, 0.0346], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:11:59,091 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 4550, loss[loss=0.2636, simple_loss=0.3165, pruned_loss=0.1054, over 4896.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.2905, pruned_loss=0.08907, over 952594.52 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:12:05,829 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0760, 1.5155, 1.6906, 1.6911, 2.3053, 1.9641, 1.6049, 1.5498], + device='cuda:2'), covar=tensor([0.1276, 0.1514, 0.1985, 0.1327, 0.0797, 0.1453, 0.2083, 0.1677], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0338, 0.0350, 0.0311, 0.0349, 0.0359, 0.0318, 0.0352], + device='cuda:2'), out_proj_covar=tensor([6.9024e-05, 7.2375e-05, 7.5804e-05, 6.5219e-05, 7.4148e-05, 7.8310e-05, + 6.9219e-05, 7.6037e-05], device='cuda:2') +2023-04-26 14:12:11,702 INFO [optim.py:369] (2/7) 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 14:12:18,132 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1109, 2.6705, 2.2409, 2.4108, 2.0250, 2.1670, 2.4391, 1.7323], + device='cuda:2'), covar=tensor([0.2423, 0.1464, 0.0945, 0.1635, 0.2815, 0.1537, 0.2081, 0.3278], + device='cuda:2'), in_proj_covar=tensor([0.0319, 0.0337, 0.0248, 0.0312, 0.0323, 0.0290, 0.0280, 0.0302], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:12:35,172 INFO [finetune.py:976] (2/7) Epoch 3, batch 4600, loss[loss=0.2161, simple_loss=0.2811, pruned_loss=0.07554, over 4913.00 frames. ], tot_loss[loss=0.2337, simple_loss=0.2902, pruned_loss=0.08862, over 952487.38 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:12:38,932 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16062.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:12:50,365 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8169, 1.3582, 1.4531, 1.5243, 2.0658, 1.7337, 1.3689, 1.3838], + device='cuda:2'), covar=tensor([0.1819, 0.1559, 0.2115, 0.1359, 0.0884, 0.1659, 0.2334, 0.2010], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0336, 0.0349, 0.0310, 0.0348, 0.0357, 0.0317, 0.0352], + device='cuda:2'), out_proj_covar=tensor([6.8734e-05, 7.2072e-05, 7.5568e-05, 6.4903e-05, 7.3870e-05, 7.7861e-05, + 6.9064e-05, 7.6054e-05], device='cuda:2') +2023-04-26 14:12:53,918 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16085.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:13:06,144 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-26 14:13:08,933 INFO [finetune.py:976] (2/7) Epoch 3, batch 4650, loss[loss=0.236, simple_loss=0.2788, pruned_loss=0.09657, over 4753.00 frames. ], tot_loss[loss=0.2328, simple_loss=0.2882, pruned_loss=0.08872, over 953446.11 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:13:18,702 INFO [optim.py:369] (2/7) 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 14:13:47,592 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16146.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:13:49,895 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16149.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:13:53,997 INFO [finetune.py:976] (2/7) Epoch 3, batch 4700, loss[loss=0.2027, simple_loss=0.2609, pruned_loss=0.07223, over 4717.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2837, pruned_loss=0.08645, over 953973.24 frames. ], batch size: 23, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:14:05,681 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5977, 1.4264, 1.9082, 1.8913, 1.4732, 1.1897, 1.6744, 1.0095], + device='cuda:2'), covar=tensor([0.1028, 0.1140, 0.0628, 0.1161, 0.1173, 0.1809, 0.0927, 0.1296], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0079, 0.0078, 0.0071, 0.0084, 0.0099, 0.0088, 0.0081], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 14:14:38,790 INFO [finetune.py:976] (2/7) Epoch 3, batch 4750, loss[loss=0.2548, simple_loss=0.3076, pruned_loss=0.101, over 4925.00 frames. ], tot_loss[loss=0.2261, simple_loss=0.2816, pruned_loss=0.08534, over 954087.98 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:14:47,161 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 14:15:10,810 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16239.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:15:17,777 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2199, 1.4411, 1.3339, 1.6958, 1.4162, 1.9892, 1.3798, 3.4324], + device='cuda:2'), covar=tensor([0.0644, 0.0772, 0.0800, 0.1207, 0.0658, 0.0573, 0.0769, 0.0175], + device='cuda:2'), in_proj_covar=tensor([0.0041, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 14:15:35,035 INFO [finetune.py:976] (2/7) Epoch 3, batch 4800, loss[loss=0.2417, simple_loss=0.3029, pruned_loss=0.09021, over 4760.00 frames. ], tot_loss[loss=0.2288, simple_loss=0.2844, pruned_loss=0.08661, over 954586.76 frames. ], batch size: 59, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:15:47,273 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1122, 1.3487, 1.3121, 1.6608, 1.4009, 1.6408, 1.3425, 2.5033], + device='cuda:2'), covar=tensor([0.0679, 0.0850, 0.0865, 0.1285, 0.0731, 0.0554, 0.0814, 0.0263], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 14:16:20,078 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 14:16:42,026 INFO [finetune.py:976] (2/7) Epoch 3, batch 4850, loss[loss=0.2526, simple_loss=0.285, pruned_loss=0.1102, over 4708.00 frames. ], tot_loss[loss=0.2304, simple_loss=0.2864, pruned_loss=0.08716, over 954627.74 frames. ], batch size: 23, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:16:43,248 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1827, 1.5011, 1.3932, 1.7645, 1.5515, 2.2080, 1.4461, 3.6774], + device='cuda:2'), covar=tensor([0.0652, 0.0754, 0.0796, 0.1219, 0.0672, 0.0536, 0.0759, 0.0124], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0041, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 14:17:03,614 INFO [optim.py:369] (2/7) 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 14:17:31,160 INFO [finetune.py:976] (2/7) Epoch 3, batch 4900, loss[loss=0.2299, simple_loss=0.2838, pruned_loss=0.08805, over 4909.00 frames. ], tot_loss[loss=0.2315, simple_loss=0.2882, pruned_loss=0.08739, over 954236.97 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:17:32,331 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16357.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:17:32,479 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 14:17:42,344 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 4950, loss[loss=0.2165, simple_loss=0.2803, pruned_loss=0.07631, over 4791.00 frames. ], tot_loss[loss=0.2338, simple_loss=0.2907, pruned_loss=0.08847, over 956478.19 frames. ], batch size: 51, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:18:16,359 INFO [optim.py:369] (2/7) 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 14:18:21,887 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16432.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:18:27,288 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16441.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:18:29,702 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8588, 2.8035, 2.3457, 3.2986, 2.8435, 2.8400, 1.2139, 2.7757], + device='cuda:2'), covar=tensor([0.2225, 0.1709, 0.3070, 0.2594, 0.3183, 0.2206, 0.5867, 0.2732], + device='cuda:2'), in_proj_covar=tensor([0.0256, 0.0228, 0.0269, 0.0321, 0.0316, 0.0264, 0.0282, 0.0285], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:18:33,365 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9136, 1.2441, 1.4891, 1.6239, 2.1109, 1.7134, 1.3863, 1.3717], + device='cuda:2'), covar=tensor([0.1598, 0.1743, 0.2360, 0.1302, 0.0847, 0.1768, 0.2588, 0.2045], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0337, 0.0350, 0.0312, 0.0349, 0.0360, 0.0318, 0.0353], + device='cuda:2'), out_proj_covar=tensor([6.8910e-05, 7.2364e-05, 7.5951e-05, 6.5372e-05, 7.4213e-05, 7.8420e-05, + 6.9263e-05, 7.6208e-05], device='cuda:2') +2023-04-26 14:18:37,159 INFO [finetune.py:976] (2/7) Epoch 3, batch 5000, loss[loss=0.1931, simple_loss=0.251, pruned_loss=0.06762, over 4766.00 frames. ], tot_loss[loss=0.2316, simple_loss=0.2884, pruned_loss=0.08738, over 957547.15 frames. ], batch size: 28, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:18:37,822 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0550, 1.1407, 5.1086, 4.7577, 4.4498, 4.7854, 4.4343, 4.4609], + device='cuda:2'), covar=tensor([0.6857, 0.6591, 0.0855, 0.1589, 0.1011, 0.1105, 0.1456, 0.1435], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0315, 0.0440, 0.0442, 0.0374, 0.0427, 0.0335, 0.0396], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002], + device='cuda:2') +2023-04-26 14:19:09,889 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16505.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:19:10,441 INFO [finetune.py:976] (2/7) Epoch 3, batch 5050, loss[loss=0.1983, simple_loss=0.2724, pruned_loss=0.06208, over 4801.00 frames. ], tot_loss[loss=0.2295, simple_loss=0.2864, pruned_loss=0.08624, over 957155.38 frames. ], batch size: 29, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:19:23,710 INFO [optim.py:369] (2/7) 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 14:19:29,923 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7512, 2.2389, 1.6381, 1.4269, 1.3118, 1.3210, 1.6359, 1.2957], + device='cuda:2'), covar=tensor([0.2342, 0.2070, 0.2470, 0.3037, 0.3411, 0.2723, 0.1783, 0.2786], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0222, 0.0185, 0.0211, 0.0223, 0.0191, 0.0180, 0.0201], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 14:19:33,535 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16539.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:19:43,698 INFO [finetune.py:976] (2/7) Epoch 3, batch 5100, loss[loss=0.2167, simple_loss=0.2791, pruned_loss=0.07715, over 4919.00 frames. ], tot_loss[loss=0.2256, simple_loss=0.2821, pruned_loss=0.08454, over 959127.19 frames. ], batch size: 46, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:19:57,650 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4932, 0.9215, 0.3782, 1.1681, 1.3213, 1.3871, 1.2405, 1.2114], + device='cuda:2'), covar=tensor([0.0596, 0.0467, 0.0516, 0.0623, 0.0312, 0.0584, 0.0583, 0.0675], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 14:20:10,602 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16587.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:20:11,421 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-26 14:20:33,135 INFO [finetune.py:976] (2/7) Epoch 3, batch 5150, loss[loss=0.1875, simple_loss=0.2485, pruned_loss=0.06327, over 4775.00 frames. ], tot_loss[loss=0.2251, simple_loss=0.2816, pruned_loss=0.08432, over 958493.90 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:20:45,264 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5156, 1.1958, 4.1802, 3.8885, 3.6960, 3.9327, 3.8410, 3.6664], + device='cuda:2'), covar=tensor([0.7584, 0.6358, 0.1065, 0.1725, 0.1182, 0.1640, 0.1619, 0.1724], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0314, 0.0439, 0.0440, 0.0372, 0.0427, 0.0333, 0.0394], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:20:56,490 INFO [optim.py:369] (2/7) 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 14:21:30,491 INFO [finetune.py:976] (2/7) Epoch 3, batch 5200, loss[loss=0.275, simple_loss=0.3109, pruned_loss=0.1195, over 4131.00 frames. ], tot_loss[loss=0.2281, simple_loss=0.2851, pruned_loss=0.08557, over 956742.37 frames. ], batch size: 18, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:21:31,207 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16657.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:22:02,897 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16705.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:22:03,452 INFO [finetune.py:976] (2/7) Epoch 3, batch 5250, loss[loss=0.2795, simple_loss=0.3263, pruned_loss=0.1163, over 4846.00 frames. ], tot_loss[loss=0.2301, simple_loss=0.2872, pruned_loss=0.08651, over 955956.49 frames. ], batch size: 47, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:22:08,948 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7484, 1.5751, 1.8520, 1.9571, 1.9331, 1.6667, 1.7366, 1.8123], + device='cuda:2'), covar=tensor([1.8279, 2.4101, 2.9446, 2.8410, 1.8974, 3.2092, 3.2529, 2.3583], + device='cuda:2'), in_proj_covar=tensor([0.0443, 0.0490, 0.0581, 0.0594, 0.0474, 0.0504, 0.0517, 0.0525], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:22:14,742 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16727.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:22:33,194 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 3, batch 5300, loss[loss=0.2357, simple_loss=0.2941, pruned_loss=0.08867, over 4837.00 frames. ], tot_loss[loss=0.2312, simple_loss=0.2887, pruned_loss=0.08686, over 955782.94 frames. ], batch size: 49, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:23:08,887 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-26 14:23:10,434 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16789.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:23:20,220 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16805.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:23:20,770 INFO [finetune.py:976] (2/7) Epoch 3, batch 5350, loss[loss=0.2973, simple_loss=0.3289, pruned_loss=0.1329, over 4208.00 frames. ], tot_loss[loss=0.2317, simple_loss=0.2889, pruned_loss=0.08723, over 956482.22 frames. ], batch size: 66, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:23:31,443 INFO [optim.py:369] (2/7) 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 14:23:33,844 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5949, 1.3553, 1.7903, 1.8203, 1.4444, 1.0461, 1.4536, 0.9281], + device='cuda:2'), covar=tensor([0.0729, 0.0931, 0.0570, 0.0874, 0.0981, 0.1954, 0.0964, 0.1330], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0079, 0.0077, 0.0072, 0.0084, 0.0098, 0.0087, 0.0081], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 14:23:52,047 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16853.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:23:53,795 INFO [finetune.py:976] (2/7) Epoch 3, batch 5400, loss[loss=0.2581, simple_loss=0.307, pruned_loss=0.1046, over 4809.00 frames. ], tot_loss[loss=0.2301, simple_loss=0.2867, pruned_loss=0.08674, over 955830.03 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:24:01,172 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16868.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:24:22,066 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-04-26 14:24:27,334 INFO [finetune.py:976] (2/7) Epoch 3, batch 5450, loss[loss=0.1678, simple_loss=0.2334, pruned_loss=0.05105, over 4756.00 frames. ], tot_loss[loss=0.2274, simple_loss=0.2835, pruned_loss=0.08566, over 956493.47 frames. ], batch size: 27, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:24:27,441 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16906.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:24:37,666 INFO [optim.py:369] (2/7) 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 14:24:37,850 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-26 14:24:40,104 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3714, 1.2923, 1.3919, 1.0854, 1.3441, 0.9827, 1.6690, 1.2186], + device='cuda:2'), covar=tensor([0.3445, 0.1725, 0.4529, 0.2408, 0.1513, 0.2520, 0.1611, 0.4442], + device='cuda:2'), in_proj_covar=tensor([0.0354, 0.0355, 0.0438, 0.0369, 0.0407, 0.0384, 0.0401, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:24:41,872 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16929.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:24:56,975 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6035, 1.4858, 4.3714, 4.0595, 3.8353, 4.1409, 4.0701, 3.8362], + device='cuda:2'), covar=tensor([0.7079, 0.5740, 0.1094, 0.1732, 0.1034, 0.1754, 0.1168, 0.1532], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0311, 0.0432, 0.0436, 0.0368, 0.0421, 0.0329, 0.0390], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:25:00,558 INFO [finetune.py:976] (2/7) Epoch 3, batch 5500, loss[loss=0.1932, simple_loss=0.2466, pruned_loss=0.0699, over 4775.00 frames. ], tot_loss[loss=0.2239, simple_loss=0.2799, pruned_loss=0.08392, over 956749.11 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:25:03,673 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3886, 1.2013, 1.6655, 1.5293, 1.2653, 1.0407, 1.2707, 0.9076], + device='cuda:2'), covar=tensor([0.0710, 0.1009, 0.0616, 0.0845, 0.1046, 0.1612, 0.0756, 0.1062], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0078, 0.0076, 0.0071, 0.0083, 0.0097, 0.0086, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 14:25:07,412 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16967.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:25:33,762 INFO [finetune.py:976] (2/7) Epoch 3, batch 5550, loss[loss=0.2627, simple_loss=0.3142, pruned_loss=0.1056, over 4829.00 frames. ], tot_loss[loss=0.2262, simple_loss=0.2821, pruned_loss=0.08512, over 956135.59 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:25:54,159 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3691, 1.1181, 1.4908, 1.7087, 1.5968, 1.3869, 1.4382, 1.4479], + device='cuda:2'), covar=tensor([1.7781, 2.2026, 2.5205, 2.8718, 1.7476, 2.8003, 2.7624, 2.2015], + device='cuda:2'), in_proj_covar=tensor([0.0440, 0.0487, 0.0577, 0.0589, 0.0471, 0.0501, 0.0514, 0.0522], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:25:54,585 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17027.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:26:35,598 INFO [finetune.py:976] (2/7) Epoch 3, batch 5600, loss[loss=0.2396, simple_loss=0.2997, pruned_loss=0.08973, over 4739.00 frames. ], tot_loss[loss=0.2286, simple_loss=0.2857, pruned_loss=0.08579, over 957716.98 frames. ], batch size: 27, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:26:59,009 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17075.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:27:16,334 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0231, 1.9323, 1.6470, 1.6508, 2.0991, 1.5408, 2.5361, 1.4363], + device='cuda:2'), covar=tensor([0.4815, 0.1950, 0.5660, 0.3795, 0.2141, 0.3072, 0.1888, 0.5221], + device='cuda:2'), in_proj_covar=tensor([0.0354, 0.0356, 0.0440, 0.0369, 0.0407, 0.0383, 0.0400, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:27:29,329 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.07 vs. limit=5.0 +2023-04-26 14:27:38,803 INFO [finetune.py:976] (2/7) Epoch 3, batch 5650, loss[loss=0.2125, simple_loss=0.281, pruned_loss=0.07196, over 4890.00 frames. ], tot_loss[loss=0.232, simple_loss=0.2899, pruned_loss=0.08708, over 958200.83 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:27:55,214 INFO [optim.py:369] (2/7) 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 14:28:19,286 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-26 14:28:26,721 INFO [finetune.py:976] (2/7) Epoch 3, batch 5700, loss[loss=0.2025, simple_loss=0.2499, pruned_loss=0.07755, over 4197.00 frames. ], tot_loss[loss=0.2304, simple_loss=0.2862, pruned_loss=0.08726, over 937874.39 frames. ], batch size: 18, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:29:04,043 INFO [finetune.py:976] (2/7) Epoch 4, batch 0, loss[loss=0.2569, simple_loss=0.3161, pruned_loss=0.09882, over 4867.00 frames. ], tot_loss[loss=0.2569, simple_loss=0.3161, pruned_loss=0.09882, over 4867.00 frames. ], batch size: 34, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:29:04,043 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 14:29:10,839 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5142, 1.1044, 1.2736, 1.1399, 1.7178, 1.4059, 1.1313, 1.2794], + device='cuda:2'), covar=tensor([0.1865, 0.1988, 0.2936, 0.1993, 0.1086, 0.1809, 0.2755, 0.2349], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0334, 0.0348, 0.0309, 0.0345, 0.0353, 0.0314, 0.0351], + device='cuda:2'), 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:2') +2023-04-26 14:29:26,719 INFO [finetune.py:1010] (2/7) Epoch 4, validation: loss=0.1686, simple_loss=0.2415, pruned_loss=0.04788, over 2265189.00 frames. +2023-04-26 14:29:26,719 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 14:29:31,451 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17189.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:29:52,777 INFO [optim.py:369] (2/7) 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 14:29:53,449 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17224.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 14:29:54,746 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4197, 0.6650, 1.0559, 1.6995, 1.5699, 1.2923, 1.2805, 1.2872], + device='cuda:2'), covar=tensor([1.4874, 1.9076, 2.1956, 2.5272, 1.5279, 2.3922, 2.1529, 1.7945], + device='cuda:2'), in_proj_covar=tensor([0.0438, 0.0486, 0.0576, 0.0587, 0.0470, 0.0498, 0.0511, 0.0520], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:29:58,839 INFO [finetune.py:976] (2/7) Epoch 4, batch 50, loss[loss=0.2265, simple_loss=0.2892, pruned_loss=0.08193, over 4818.00 frames. ], tot_loss[loss=0.2296, simple_loss=0.288, pruned_loss=0.08561, over 217340.66 frames. ], batch size: 30, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:30:11,291 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17250.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:30:16,196 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1633, 2.7022, 1.0473, 1.3713, 1.9915, 1.2491, 3.7425, 1.6698], + device='cuda:2'), covar=tensor([0.0731, 0.0749, 0.0963, 0.1523, 0.0586, 0.1231, 0.0266, 0.0789], + device='cuda:2'), in_proj_covar=tensor([0.0054, 0.0071, 0.0053, 0.0050, 0.0055, 0.0055, 0.0084, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 14:30:18,608 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17262.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:30:31,853 INFO [finetune.py:976] (2/7) Epoch 4, batch 100, loss[loss=0.1865, simple_loss=0.2451, pruned_loss=0.06395, over 4898.00 frames. ], tot_loss[loss=0.2271, simple_loss=0.2829, pruned_loss=0.08568, over 380701.95 frames. ], batch size: 35, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:30:58,842 INFO [optim.py:369] (2/7) 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 14:31:04,967 INFO [finetune.py:976] (2/7) Epoch 4, batch 150, loss[loss=0.2182, simple_loss=0.2717, pruned_loss=0.0823, over 4816.00 frames. ], tot_loss[loss=0.221, simple_loss=0.2769, pruned_loss=0.0826, over 509695.73 frames. ], batch size: 41, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:31:13,152 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1072, 1.3684, 1.2349, 1.7566, 1.5192, 1.5073, 1.3531, 2.5092], + device='cuda:2'), covar=tensor([0.0678, 0.0864, 0.0930, 0.1214, 0.0684, 0.0539, 0.0765, 0.0269], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0040, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 14:31:24,774 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17362.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:31:33,303 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-26 14:31:36,405 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 14:31:38,035 INFO [finetune.py:976] (2/7) Epoch 4, batch 200, loss[loss=0.2644, simple_loss=0.3257, pruned_loss=0.1015, over 4852.00 frames. ], tot_loss[loss=0.224, simple_loss=0.2788, pruned_loss=0.08463, over 610430.34 frames. ], batch size: 44, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:31:38,167 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17383.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:32:05,046 INFO [optim.py:369] (2/7) 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 14:32:05,202 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 250, loss[loss=0.2131, simple_loss=0.2759, pruned_loss=0.07516, over 4755.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2818, pruned_loss=0.08485, over 687141.97 frames. ], batch size: 27, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:32:25,185 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17444.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 14:33:09,718 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.37 vs. limit=5.0 +2023-04-26 14:33:16,612 INFO [finetune.py:976] (2/7) Epoch 4, batch 300, loss[loss=0.2245, simple_loss=0.2803, pruned_loss=0.08435, over 4782.00 frames. ], tot_loss[loss=0.2277, simple_loss=0.2847, pruned_loss=0.08536, over 748284.78 frames. ], batch size: 28, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:33:19,135 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7112, 1.2088, 1.4988, 1.4712, 1.3562, 1.1817, 0.5319, 1.0881], + device='cuda:2'), covar=tensor([0.4973, 0.5981, 0.2635, 0.4342, 0.5332, 0.4305, 0.7495, 0.4710], + device='cuda:2'), in_proj_covar=tensor([0.0276, 0.0272, 0.0226, 0.0343, 0.0230, 0.0236, 0.0259, 0.0206], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:33:50,245 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-26 14:34:04,407 INFO [optim.py:369] (2/7) 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 14:34:10,655 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17524.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:34:22,587 INFO [finetune.py:976] (2/7) Epoch 4, batch 350, loss[loss=0.2496, simple_loss=0.2989, pruned_loss=0.1001, over 4823.00 frames. ], tot_loss[loss=0.2319, simple_loss=0.2884, pruned_loss=0.08769, over 794724.86 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:34:36,259 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17545.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:34:59,127 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17562.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 14:35:10,345 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17572.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:35:17,455 INFO [finetune.py:976] (2/7) Epoch 4, batch 400, loss[loss=0.2248, simple_loss=0.2927, pruned_loss=0.07844, over 4844.00 frames. ], tot_loss[loss=0.2309, simple_loss=0.2884, pruned_loss=0.08665, over 832358.96 frames. ], batch size: 49, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:35:37,142 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17610.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:35:45,053 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 4, batch 450, loss[loss=0.2373, simple_loss=0.2896, pruned_loss=0.09252, over 4824.00 frames. ], tot_loss[loss=0.23, simple_loss=0.2877, pruned_loss=0.08614, over 860982.46 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:36:25,012 INFO [finetune.py:976] (2/7) Epoch 4, batch 500, loss[loss=0.199, simple_loss=0.249, pruned_loss=0.07451, over 4762.00 frames. ], tot_loss[loss=0.2261, simple_loss=0.2836, pruned_loss=0.08433, over 883571.73 frames. ], batch size: 27, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:36:35,724 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17700.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:36:49,495 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17718.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:36:52,485 INFO [optim.py:369] (2/7) 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 14:36:56,854 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17730.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:36:58,589 INFO [finetune.py:976] (2/7) Epoch 4, batch 550, loss[loss=0.2116, simple_loss=0.2585, pruned_loss=0.08232, over 4828.00 frames. ], tot_loss[loss=0.223, simple_loss=0.2799, pruned_loss=0.08302, over 901586.56 frames. ], batch size: 40, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:37:02,831 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17739.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:37:17,096 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6685, 2.1055, 1.7649, 1.9708, 1.7258, 1.7376, 1.8277, 1.4130], + device='cuda:2'), covar=tensor([0.2293, 0.1766, 0.1260, 0.1663, 0.3326, 0.1542, 0.2158, 0.3042], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0336, 0.0246, 0.0308, 0.0321, 0.0288, 0.0277, 0.0299], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:37:18,654 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17761.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:37:43,916 INFO [finetune.py:976] (2/7) Epoch 4, batch 600, loss[loss=0.2403, simple_loss=0.286, pruned_loss=0.09729, over 4871.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2794, pruned_loss=0.08287, over 913763.02 frames. ], batch size: 31, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:37:49,336 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17791.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:37:51,299 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-04-26 14:38:23,100 INFO [optim.py:369] (2/7) 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 14:38:34,870 INFO [finetune.py:976] (2/7) Epoch 4, batch 650, loss[loss=0.214, simple_loss=0.2879, pruned_loss=0.07007, over 4922.00 frames. ], tot_loss[loss=0.2263, simple_loss=0.2835, pruned_loss=0.08458, over 921358.32 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:38:52,794 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17845.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:39:11,528 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1347, 1.3429, 1.1373, 1.3069, 1.1033, 1.1093, 1.1943, 0.9673], + device='cuda:2'), covar=tensor([0.2028, 0.1662, 0.1373, 0.1588, 0.3034, 0.1652, 0.1901, 0.2585], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0335, 0.0245, 0.0308, 0.0321, 0.0287, 0.0276, 0.0300], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:39:22,791 INFO [finetune.py:976] (2/7) Epoch 4, batch 700, loss[loss=0.2562, simple_loss=0.3093, pruned_loss=0.1015, over 4907.00 frames. ], tot_loss[loss=0.2286, simple_loss=0.2862, pruned_loss=0.08547, over 929921.01 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:39:28,940 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17893.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:39:38,618 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7932, 1.1986, 1.3663, 1.4542, 2.0214, 1.6812, 1.3391, 1.3631], + device='cuda:2'), covar=tensor([0.1515, 0.1925, 0.1888, 0.1639, 0.0900, 0.1722, 0.2310, 0.1810], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0338, 0.0352, 0.0311, 0.0348, 0.0356, 0.0318, 0.0353], + device='cuda:2'), out_proj_covar=tensor([6.8495e-05, 7.2540e-05, 7.6409e-05, 6.5173e-05, 7.3970e-05, 7.7726e-05, + 6.9229e-05, 7.6322e-05], device='cuda:2') +2023-04-26 14:40:07,831 INFO [optim.py:369] (2/7) 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 14:40:19,660 INFO [finetune.py:976] (2/7) Epoch 4, batch 750, loss[loss=0.2346, simple_loss=0.2979, pruned_loss=0.08568, over 4842.00 frames. ], tot_loss[loss=0.2308, simple_loss=0.2885, pruned_loss=0.08652, over 936003.78 frames. ], batch size: 44, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:40:52,668 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6466, 1.5493, 1.8962, 1.9517, 1.8631, 1.5429, 1.6914, 1.7231], + device='cuda:2'), covar=tensor([1.7242, 2.1646, 2.6478, 2.7764, 1.9572, 3.0197, 2.9394, 2.4097], + device='cuda:2'), in_proj_covar=tensor([0.0437, 0.0484, 0.0575, 0.0588, 0.0468, 0.0498, 0.0510, 0.0519], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:40:53,850 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17960.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:41:24,410 INFO [finetune.py:976] (2/7) Epoch 4, batch 800, loss[loss=0.2615, simple_loss=0.3074, pruned_loss=0.1078, over 4807.00 frames. ], tot_loss[loss=0.2292, simple_loss=0.2872, pruned_loss=0.08558, over 938789.18 frames. ], batch size: 45, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:41:48,952 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18018.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:41:51,295 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 4, batch 850, loss[loss=0.1999, simple_loss=0.2656, pruned_loss=0.06709, over 4838.00 frames. ], tot_loss[loss=0.227, simple_loss=0.2845, pruned_loss=0.08478, over 943834.78 frames. ], batch size: 44, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:42:02,668 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18039.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:42:07,102 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.65 vs. limit=2.0 +2023-04-26 14:42:13,531 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18056.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:42:20,541 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18066.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:42:32,277 INFO [finetune.py:976] (2/7) Epoch 4, batch 900, loss[loss=0.1777, simple_loss=0.2391, pruned_loss=0.05817, over 4910.00 frames. ], tot_loss[loss=0.2237, simple_loss=0.2809, pruned_loss=0.08328, over 945420.56 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:42:34,141 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18086.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:42:34,723 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18087.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:42:58,116 INFO [optim.py:369] (2/7) 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 14:43:05,176 INFO [finetune.py:976] (2/7) Epoch 4, batch 950, loss[loss=0.1769, simple_loss=0.2385, pruned_loss=0.05765, over 4899.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2795, pruned_loss=0.08287, over 949599.03 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:43:33,531 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4428, 2.3787, 2.6962, 3.0630, 2.6135, 2.2959, 1.9436, 2.4223], + device='cuda:2'), covar=tensor([0.1127, 0.1106, 0.0620, 0.0640, 0.0807, 0.1143, 0.1225, 0.0729], + device='cuda:2'), in_proj_covar=tensor([0.0211, 0.0212, 0.0189, 0.0184, 0.0184, 0.0200, 0.0174, 0.0194], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:43:44,146 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18174.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:43:44,763 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18175.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:43:53,135 INFO [finetune.py:976] (2/7) Epoch 4, batch 1000, loss[loss=0.2321, simple_loss=0.2949, pruned_loss=0.08466, over 4931.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.2817, pruned_loss=0.08348, over 950720.37 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:44:06,222 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1192, 1.5028, 4.3871, 4.1160, 3.8764, 3.9652, 3.8526, 3.9152], + device='cuda:2'), covar=tensor([0.6279, 0.5486, 0.1061, 0.1663, 0.1150, 0.1670, 0.3414, 0.1324], + device='cuda:2'), in_proj_covar=tensor([0.0322, 0.0311, 0.0431, 0.0432, 0.0370, 0.0421, 0.0331, 0.0388], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:44:29,361 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6418, 1.3063, 1.2248, 1.4037, 1.9338, 1.5514, 1.2831, 1.2622], + device='cuda:2'), covar=tensor([0.1637, 0.1675, 0.2104, 0.1545, 0.0770, 0.1724, 0.2407, 0.1878], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0337, 0.0351, 0.0310, 0.0347, 0.0354, 0.0316, 0.0351], + device='cuda:2'), out_proj_covar=tensor([6.8392e-05, 7.2230e-05, 7.6017e-05, 6.4813e-05, 7.3599e-05, 7.7231e-05, + 6.8635e-05, 7.5812e-05], device='cuda:2') +2023-04-26 14:44:30,497 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 4, batch 1050, loss[loss=0.2282, simple_loss=0.2844, pruned_loss=0.08595, over 4831.00 frames. ], tot_loss[loss=0.2262, simple_loss=0.2845, pruned_loss=0.08389, over 952513.02 frames. ], batch size: 30, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:44:39,866 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18235.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:44:40,477 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:44:43,485 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6775, 1.8236, 1.6095, 1.7747, 1.5804, 1.9423, 1.7604, 1.6992], + device='cuda:2'), covar=tensor([1.1055, 1.9701, 1.6193, 1.3505, 1.5791, 2.3064, 2.0665, 1.7696], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0402, 0.0320, 0.0325, 0.0350, 0.0412, 0.0386, 0.0341], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:45:12,451 INFO [finetune.py:976] (2/7) Epoch 4, batch 1100, loss[loss=0.2038, simple_loss=0.275, pruned_loss=0.06636, over 4924.00 frames. ], tot_loss[loss=0.2275, simple_loss=0.2861, pruned_loss=0.08447, over 953413.91 frames. ], batch size: 29, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:45:50,513 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18316.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:46:00,501 INFO [optim.py:369] (2/7) 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 14:46:08,158 INFO [finetune.py:976] (2/7) Epoch 4, batch 1150, loss[loss=0.2422, simple_loss=0.2852, pruned_loss=0.09961, over 4268.00 frames. ], tot_loss[loss=0.2274, simple_loss=0.2859, pruned_loss=0.08443, over 951847.97 frames. ], batch size: 66, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:46:20,794 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0060, 2.0080, 1.8287, 1.6689, 2.2002, 1.5896, 2.7394, 1.5313], + device='cuda:2'), covar=tensor([0.5049, 0.1886, 0.5066, 0.3695, 0.2164, 0.3263, 0.1690, 0.5243], + device='cuda:2'), in_proj_covar=tensor([0.0354, 0.0358, 0.0442, 0.0370, 0.0408, 0.0384, 0.0401, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:46:23,839 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18356.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:46:52,607 INFO [finetune.py:976] (2/7) Epoch 4, batch 1200, loss[loss=0.2168, simple_loss=0.2735, pruned_loss=0.0801, over 4858.00 frames. ], tot_loss[loss=0.2256, simple_loss=0.2838, pruned_loss=0.08374, over 953930.75 frames. ], batch size: 31, lr: 3.97e-03, grad_scale: 16.0 +2023-04-26 14:46:54,502 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 14:46:55,525 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18386.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:47:18,677 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18404.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:47:18,912 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-26 14:47:26,668 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-26 14:47:36,012 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 4, batch 1250, loss[loss=0.2445, simple_loss=0.296, pruned_loss=0.09649, over 4824.00 frames. ], tot_loss[loss=0.2223, simple_loss=0.28, pruned_loss=0.08234, over 953712.77 frames. ], batch size: 40, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:47:43,323 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18434.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:48:02,168 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3588, 1.2891, 3.8243, 3.5505, 3.3825, 3.6577, 3.6738, 3.3361], + device='cuda:2'), covar=tensor([0.6947, 0.5949, 0.1180, 0.1787, 0.1271, 0.1888, 0.1628, 0.1622], + device='cuda:2'), in_proj_covar=tensor([0.0324, 0.0313, 0.0433, 0.0435, 0.0371, 0.0423, 0.0333, 0.0391], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:48:09,369 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5666, 1.3015, 0.6599, 1.2938, 1.4177, 1.4749, 1.3726, 1.3727], + device='cuda:2'), covar=tensor([0.0547, 0.0417, 0.0474, 0.0569, 0.0328, 0.0543, 0.0520, 0.0590], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 14:48:10,196 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 14:48:15,942 INFO [finetune.py:976] (2/7) Epoch 4, batch 1300, loss[loss=0.2009, simple_loss=0.2627, pruned_loss=0.06961, over 4824.00 frames. ], tot_loss[loss=0.2192, simple_loss=0.2766, pruned_loss=0.08087, over 954586.90 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:48:21,739 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18490.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:48:34,007 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18508.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:48:43,005 INFO [optim.py:369] (2/7) 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 14:48:47,800 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18530.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:48:48,423 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 1350, loss[loss=0.2378, simple_loss=0.2979, pruned_loss=0.08885, over 4915.00 frames. ], tot_loss[loss=0.2216, simple_loss=0.2788, pruned_loss=0.0822, over 954288.02 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:49:08,084 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:49:32,019 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18569.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 14:49:44,334 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 14:49:45,978 INFO [finetune.py:976] (2/7) Epoch 4, batch 1400, loss[loss=0.2518, simple_loss=0.314, pruned_loss=0.09486, over 4803.00 frames. ], tot_loss[loss=0.2265, simple_loss=0.2842, pruned_loss=0.08444, over 953954.35 frames. ], batch size: 45, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:49:46,080 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3840, 0.9692, 0.4011, 1.0520, 1.2106, 1.2964, 1.1312, 1.1551], + device='cuda:2'), covar=tensor([0.0601, 0.0508, 0.0530, 0.0673, 0.0332, 0.0603, 0.0634, 0.0701], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 14:50:06,717 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9438, 1.3918, 1.7924, 2.2477, 1.7173, 1.3431, 1.1307, 1.7014], + device='cuda:2'), covar=tensor([0.4617, 0.5220, 0.2502, 0.3987, 0.4708, 0.3965, 0.7082, 0.4063], + device='cuda:2'), in_proj_covar=tensor([0.0272, 0.0267, 0.0222, 0.0336, 0.0225, 0.0232, 0.0252, 0.0202], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:50:09,110 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18616.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:50:11,031 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-26 14:50:13,226 INFO [optim.py:369] (2/7) 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 14:50:19,721 INFO [finetune.py:976] (2/7) Epoch 4, batch 1450, loss[loss=0.2989, simple_loss=0.3485, pruned_loss=0.1247, over 4805.00 frames. ], tot_loss[loss=0.228, simple_loss=0.2857, pruned_loss=0.08516, over 954262.15 frames. ], batch size: 45, lr: 3.97e-03, grad_scale: 32.0 +2023-04-26 14:50:21,218 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-26 14:50:39,179 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8141, 1.3248, 1.4361, 1.4645, 2.0718, 1.6603, 1.2820, 1.3507], + device='cuda:2'), covar=tensor([0.1696, 0.1675, 0.2176, 0.1582, 0.0822, 0.1773, 0.2606, 0.2140], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0334, 0.0347, 0.0306, 0.0344, 0.0350, 0.0313, 0.0350], + device='cuda:2'), out_proj_covar=tensor([6.7688e-05, 7.1615e-05, 7.5154e-05, 6.4132e-05, 7.2891e-05, 7.6411e-05, + 6.7982e-05, 7.5548e-05], device='cuda:2') +2023-04-26 14:50:40,790 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18664.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:50:52,629 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18674.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:51:03,069 INFO [finetune.py:976] (2/7) Epoch 4, batch 1500, loss[loss=0.2507, simple_loss=0.3105, pruned_loss=0.09545, over 4924.00 frames. ], tot_loss[loss=0.2307, simple_loss=0.2883, pruned_loss=0.08657, over 951755.23 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:51:14,065 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8849, 1.2461, 3.0041, 2.7736, 2.7134, 2.9424, 2.9377, 2.6730], + device='cuda:2'), covar=tensor([0.6515, 0.4635, 0.1404, 0.1945, 0.1297, 0.1637, 0.1398, 0.1655], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0308, 0.0429, 0.0431, 0.0367, 0.0417, 0.0330, 0.0388], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:51:57,996 INFO [optim.py:369] (2/7) 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 14:51:59,254 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9205, 1.3445, 3.3021, 3.0474, 2.9814, 3.2158, 3.1852, 2.8866], + device='cuda:2'), covar=tensor([0.7028, 0.4993, 0.1409, 0.2013, 0.1535, 0.1918, 0.1623, 0.1756], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0309, 0.0428, 0.0432, 0.0367, 0.0418, 0.0330, 0.0388], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:52:09,749 INFO [finetune.py:976] (2/7) Epoch 4, batch 1550, loss[loss=0.2755, simple_loss=0.3245, pruned_loss=0.1132, over 4844.00 frames. ], tot_loss[loss=0.2303, simple_loss=0.2878, pruned_loss=0.08638, over 951728.40 frames. ], batch size: 47, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:52:11,107 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18735.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:53:12,328 INFO [finetune.py:976] (2/7) Epoch 4, batch 1600, loss[loss=0.2117, simple_loss=0.2681, pruned_loss=0.07769, over 4721.00 frames. ], tot_loss[loss=0.2276, simple_loss=0.2844, pruned_loss=0.08538, over 949770.87 frames. ], batch size: 54, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:53:35,553 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 14:53:44,152 INFO [optim.py:369] (2/7) 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 14:53:48,481 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18830.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:53:49,089 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18831.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:53:50,210 INFO [finetune.py:976] (2/7) Epoch 4, batch 1650, loss[loss=0.1878, simple_loss=0.2498, pruned_loss=0.0629, over 4930.00 frames. ], tot_loss[loss=0.2254, simple_loss=0.2819, pruned_loss=0.08449, over 953054.98 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:53:58,587 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18846.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:54:11,014 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18864.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:54:20,472 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18879.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:54:23,387 INFO [finetune.py:976] (2/7) Epoch 4, batch 1700, loss[loss=0.1986, simple_loss=0.2699, pruned_loss=0.06366, over 4783.00 frames. ], tot_loss[loss=0.2228, simple_loss=0.2795, pruned_loss=0.08306, over 952560.11 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:54:29,757 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1450, 1.4994, 2.0808, 2.6830, 2.0147, 1.5564, 1.4834, 1.8558], + device='cuda:2'), covar=tensor([0.4386, 0.4966, 0.2227, 0.3492, 0.4691, 0.3683, 0.6151, 0.4076], + device='cuda:2'), in_proj_covar=tensor([0.0271, 0.0265, 0.0221, 0.0334, 0.0224, 0.0231, 0.0252, 0.0201], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:55:17,077 INFO [optim.py:369] (2/7) 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 14:55:21,956 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 1750, loss[loss=0.2867, simple_loss=0.3508, pruned_loss=0.1113, over 4724.00 frames. ], tot_loss[loss=0.2241, simple_loss=0.2811, pruned_loss=0.08349, over 953748.04 frames. ], batch size: 54, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:55:37,651 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.4703, 4.3277, 3.0716, 5.1210, 4.3753, 4.3850, 2.0736, 4.4684], + device='cuda:2'), covar=tensor([0.1575, 0.0940, 0.3357, 0.0812, 0.2856, 0.1708, 0.4985, 0.1919], + device='cuda:2'), in_proj_covar=tensor([0.0254, 0.0226, 0.0265, 0.0320, 0.0312, 0.0263, 0.0279, 0.0283], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:55:55,036 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5228, 3.5022, 2.5659, 4.0890, 3.5575, 3.4989, 1.5691, 3.4582], + device='cuda:2'), covar=tensor([0.1673, 0.1264, 0.3427, 0.1838, 0.2691, 0.2015, 0.5304, 0.2339], + device='cuda:2'), in_proj_covar=tensor([0.0253, 0.0225, 0.0264, 0.0318, 0.0311, 0.0262, 0.0277, 0.0281], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 14:55:56,819 INFO [finetune.py:976] (2/7) Epoch 4, batch 1800, loss[loss=0.2338, simple_loss=0.2974, pruned_loss=0.08512, over 4816.00 frames. ], tot_loss[loss=0.2266, simple_loss=0.2839, pruned_loss=0.08463, over 953698.21 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:56:08,195 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18991.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:56:11,895 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6309, 1.6318, 1.6406, 1.3597, 1.7611, 1.3413, 2.2360, 1.4382], + device='cuda:2'), covar=tensor([0.4764, 0.1969, 0.6336, 0.3529, 0.1989, 0.2979, 0.1698, 0.5147], + device='cuda:2'), in_proj_covar=tensor([0.0356, 0.0358, 0.0443, 0.0373, 0.0406, 0.0385, 0.0400, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 14:56:44,255 INFO [optim.py:369] (2/7) 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 14:56:44,372 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2663, 1.5513, 1.3935, 1.9986, 1.7381, 2.0331, 1.4262, 4.2954], + device='cuda:2'), covar=tensor([0.0691, 0.0778, 0.0819, 0.1183, 0.0655, 0.0691, 0.0789, 0.0113], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0040, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 14:56:46,720 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4969, 3.5942, 0.8329, 1.9017, 1.9978, 2.4289, 1.9866, 1.0349], + device='cuda:2'), covar=tensor([0.1440, 0.0926, 0.2328, 0.1353, 0.1092, 0.1131, 0.1579, 0.2023], + device='cuda:2'), in_proj_covar=tensor([0.0122, 0.0265, 0.0149, 0.0129, 0.0140, 0.0161, 0.0126, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 14:56:49,160 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 1850, loss[loss=0.2392, simple_loss=0.3015, pruned_loss=0.08843, over 4780.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2867, pruned_loss=0.08517, over 954791.52 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:57:24,083 INFO [finetune.py:976] (2/7) Epoch 4, batch 1900, loss[loss=0.2061, simple_loss=0.2712, pruned_loss=0.07052, over 4770.00 frames. ], tot_loss[loss=0.2286, simple_loss=0.2873, pruned_loss=0.08498, over 955115.68 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:57:50,668 INFO [optim.py:369] (2/7) 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 14:57:54,905 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19129.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 14:57:57,174 INFO [finetune.py:976] (2/7) Epoch 4, batch 1950, loss[loss=0.3332, simple_loss=0.3409, pruned_loss=0.1627, over 4345.00 frames. ], tot_loss[loss=0.2267, simple_loss=0.285, pruned_loss=0.08424, over 954157.03 frames. ], batch size: 66, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:58:05,511 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19146.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:58:17,048 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19164.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:58:27,652 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-04-26 14:58:30,343 INFO [finetune.py:976] (2/7) Epoch 4, batch 2000, loss[loss=0.1739, simple_loss=0.2426, pruned_loss=0.05259, over 4865.00 frames. ], tot_loss[loss=0.224, simple_loss=0.2816, pruned_loss=0.08323, over 953736.38 frames. ], batch size: 31, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:58:40,465 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19190.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 14:58:48,412 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19194.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 14:59:10,842 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19212.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 14:59:24,205 INFO [optim.py:369] (2/7) 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 14:59:36,723 INFO [finetune.py:976] (2/7) Epoch 4, batch 2050, loss[loss=0.2037, simple_loss=0.2559, pruned_loss=0.07576, over 4753.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2776, pruned_loss=0.08169, over 954242.86 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 14:59:54,047 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.7544, 4.6579, 3.4021, 5.4903, 4.8592, 4.6657, 2.7675, 4.6648], + device='cuda:2'), covar=tensor([0.1404, 0.0899, 0.2881, 0.0812, 0.3669, 0.1755, 0.4569, 0.2194], + device='cuda:2'), in_proj_covar=tensor([0.0252, 0.0224, 0.0263, 0.0316, 0.0311, 0.0261, 0.0277, 0.0281], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:00:01,360 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4825, 1.3775, 4.1218, 3.8142, 3.6648, 3.9598, 3.8922, 3.6703], + device='cuda:2'), covar=tensor([0.6839, 0.5879, 0.1211, 0.1969, 0.1203, 0.1823, 0.1376, 0.1578], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0308, 0.0426, 0.0431, 0.0365, 0.0417, 0.0328, 0.0384], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:00:03,650 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5758, 3.8110, 0.7844, 2.0992, 2.0412, 2.5120, 2.3190, 1.1195], + device='cuda:2'), covar=tensor([0.1407, 0.0906, 0.2315, 0.1389, 0.1159, 0.1231, 0.1345, 0.2177], + device='cuda:2'), in_proj_covar=tensor([0.0122, 0.0265, 0.0148, 0.0129, 0.0140, 0.0162, 0.0126, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 15:00:15,254 INFO [finetune.py:976] (2/7) Epoch 4, batch 2100, loss[loss=0.2537, simple_loss=0.3128, pruned_loss=0.09728, over 4852.00 frames. ], tot_loss[loss=0.2203, simple_loss=0.2772, pruned_loss=0.0817, over 953924.94 frames. ], batch size: 47, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:00:17,154 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19286.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:00:46,624 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-26 15:01:01,582 INFO [optim.py:369] (2/7) 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 15:01:07,018 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 2150, loss[loss=0.2694, simple_loss=0.3362, pruned_loss=0.1013, over 4071.00 frames. ], tot_loss[loss=0.2268, simple_loss=0.2842, pruned_loss=0.08473, over 954997.52 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:01:25,229 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5266, 1.3194, 0.6420, 1.2535, 1.5634, 1.4109, 1.3073, 1.3143], + device='cuda:2'), covar=tensor([0.0564, 0.0457, 0.0471, 0.0599, 0.0318, 0.0550, 0.0585, 0.0660], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 15:02:07,704 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19378.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:02:16,423 INFO [finetune.py:976] (2/7) Epoch 4, batch 2200, loss[loss=0.249, simple_loss=0.31, pruned_loss=0.09399, over 4892.00 frames. ], tot_loss[loss=0.2275, simple_loss=0.2854, pruned_loss=0.08483, over 956172.27 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:02:30,493 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19394.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:03:06,439 INFO [optim.py:369] (2/7) 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 15:03:18,525 INFO [finetune.py:976] (2/7) Epoch 4, batch 2250, loss[loss=0.2095, simple_loss=0.2684, pruned_loss=0.07532, over 4837.00 frames. ], tot_loss[loss=0.2289, simple_loss=0.2868, pruned_loss=0.08552, over 957142.12 frames. ], batch size: 44, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:03:51,638 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19455.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:04:06,327 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.13 vs. limit=5.0 +2023-04-26 15:04:21,504 INFO [finetune.py:976] (2/7) Epoch 4, batch 2300, loss[loss=0.2722, simple_loss=0.3125, pruned_loss=0.1159, over 4808.00 frames. ], tot_loss[loss=0.2287, simple_loss=0.287, pruned_loss=0.08515, over 957479.23 frames. ], batch size: 40, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:04:23,293 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19485.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 15:04:58,953 INFO [optim.py:369] (2/7) 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 15:05:02,209 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-26 15:05:05,527 INFO [finetune.py:976] (2/7) Epoch 4, batch 2350, loss[loss=0.2297, simple_loss=0.2875, pruned_loss=0.08598, over 4847.00 frames. ], tot_loss[loss=0.2247, simple_loss=0.2826, pruned_loss=0.08342, over 954978.76 frames. ], batch size: 47, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:05:39,105 INFO [finetune.py:976] (2/7) Epoch 4, batch 2400, loss[loss=0.155, simple_loss=0.2168, pruned_loss=0.04659, over 4688.00 frames. ], tot_loss[loss=0.221, simple_loss=0.2787, pruned_loss=0.08164, over 956961.47 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:05:41,041 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19586.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:05:41,146 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-04-26 15:06:00,690 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.5167, 4.3523, 3.2908, 5.2185, 4.4967, 4.5020, 2.2964, 4.4024], + device='cuda:2'), covar=tensor([0.1640, 0.0910, 0.2867, 0.1004, 0.3254, 0.1770, 0.5221, 0.2295], + device='cuda:2'), in_proj_covar=tensor([0.0254, 0.0226, 0.0264, 0.0318, 0.0312, 0.0262, 0.0279, 0.0283], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:06:03,644 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:06:12,787 INFO [finetune.py:976] (2/7) Epoch 4, batch 2450, loss[loss=0.2306, simple_loss=0.2882, pruned_loss=0.08649, over 4863.00 frames. ], tot_loss[loss=0.2172, simple_loss=0.275, pruned_loss=0.07965, over 957803.20 frames. ], batch size: 31, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:06:13,447 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19634.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:06:37,089 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6870, 2.0425, 1.7686, 1.9395, 1.6957, 1.6363, 1.7437, 1.3518], + device='cuda:2'), covar=tensor([0.1949, 0.1568, 0.1089, 0.1349, 0.3203, 0.1709, 0.1903, 0.2850], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0328, 0.0239, 0.0302, 0.0316, 0.0281, 0.0271, 0.0293], + device='cuda:2'), out_proj_covar=tensor([1.2593e-04, 1.3363e-04, 9.7412e-05, 1.2180e-04, 1.3047e-04, 1.1347e-04, + 1.1209e-04, 1.1829e-04], device='cuda:2') +2023-04-26 15:06:41,751 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0357, 1.5786, 5.5344, 5.0989, 4.7515, 5.1559, 4.6789, 4.8594], + device='cuda:2'), covar=tensor([0.6903, 0.6408, 0.0965, 0.1884, 0.1094, 0.1631, 0.1307, 0.1483], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0309, 0.0428, 0.0434, 0.0367, 0.0420, 0.0329, 0.0386], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:06:44,237 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 2500, loss[loss=0.2219, simple_loss=0.2957, pruned_loss=0.07403, over 4819.00 frames. ], tot_loss[loss=0.2194, simple_loss=0.2772, pruned_loss=0.0808, over 956011.64 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:07:31,980 INFO [optim.py:369] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.5928, 2.0921, 1.6174, 1.4131, 1.2289, 1.2059, 1.6294, 1.1832], + device='cuda:2'), covar=tensor([0.2129, 0.1856, 0.2117, 0.2447, 0.3259, 0.2548, 0.1605, 0.2590], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0222, 0.0183, 0.0211, 0.0222, 0.0189, 0.0177, 0.0200], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 15:07:43,942 INFO [finetune.py:976] (2/7) Epoch 4, batch 2550, loss[loss=0.2611, simple_loss=0.3165, pruned_loss=0.1029, over 4858.00 frames. ], tot_loss[loss=0.2223, simple_loss=0.2812, pruned_loss=0.08174, over 956520.97 frames. ], batch size: 44, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:08:07,139 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19750.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:08:50,446 INFO [finetune.py:976] (2/7) Epoch 4, batch 2600, loss[loss=0.2524, simple_loss=0.3, pruned_loss=0.1024, over 4865.00 frames. ], tot_loss[loss=0.225, simple_loss=0.2839, pruned_loss=0.08303, over 956359.11 frames. ], batch size: 34, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:08:51,170 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6926, 1.4819, 0.7021, 1.3491, 1.5256, 1.5684, 1.4184, 1.4375], + device='cuda:2'), covar=tensor([0.0524, 0.0444, 0.0460, 0.0589, 0.0320, 0.0560, 0.0539, 0.0651], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 15:08:57,300 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19785.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 15:09:06,415 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 15:09:23,390 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 4, batch 2650, loss[loss=0.2455, simple_loss=0.3046, pruned_loss=0.09321, over 4808.00 frames. ], tot_loss[loss=0.2271, simple_loss=0.2859, pruned_loss=0.08417, over 954391.93 frames. ], batch size: 40, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:09:29,968 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19833.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 15:10:35,942 INFO [finetune.py:976] (2/7) Epoch 4, batch 2700, loss[loss=0.1908, simple_loss=0.2542, pruned_loss=0.06367, over 4823.00 frames. ], tot_loss[loss=0.2255, simple_loss=0.284, pruned_loss=0.0835, over 953280.34 frames. ], batch size: 25, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:10:45,476 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4391, 0.7622, 1.1089, 1.0307, 1.5771, 1.2148, 1.0369, 1.0530], + device='cuda:2'), covar=tensor([0.1503, 0.2155, 0.2486, 0.1705, 0.0981, 0.1717, 0.2301, 0.2346], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0338, 0.0353, 0.0311, 0.0352, 0.0355, 0.0317, 0.0355], + device='cuda:2'), 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:2') +2023-04-26 15:11:06,173 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6254, 1.2185, 1.2394, 1.4472, 1.9062, 1.5923, 1.3371, 1.2429], + device='cuda:2'), covar=tensor([0.1579, 0.1812, 0.2101, 0.1477, 0.0810, 0.1590, 0.1984, 0.1865], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0338, 0.0352, 0.0311, 0.0351, 0.0354, 0.0316, 0.0355], + device='cuda:2'), 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:2') +2023-04-26 15:11:14,321 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 15:11:15,313 INFO [optim.py:369] (2/7) 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 15:11:18,641 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 15:11:21,340 INFO [finetune.py:976] (2/7) Epoch 4, batch 2750, loss[loss=0.2719, simple_loss=0.3084, pruned_loss=0.1177, over 4740.00 frames. ], tot_loss[loss=0.2231, simple_loss=0.2809, pruned_loss=0.08265, over 952928.61 frames. ], batch size: 59, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:11:49,852 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 2800, loss[loss=0.2115, simple_loss=0.2666, pruned_loss=0.07824, over 4806.00 frames. ], tot_loss[loss=0.2193, simple_loss=0.2768, pruned_loss=0.08087, over 955022.40 frames. ], batch size: 45, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:12:21,322 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20020.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:12:23,471 INFO [optim.py:369] (2/7) 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 15:12:30,049 INFO [finetune.py:976] (2/7) Epoch 4, batch 2850, loss[loss=0.2205, simple_loss=0.2779, pruned_loss=0.08162, over 4899.00 frames. ], tot_loss[loss=0.2164, simple_loss=0.2739, pruned_loss=0.07943, over 956017.64 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:12:30,314 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.96 vs. limit=5.0 +2023-04-26 15:12:40,866 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20050.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:12:43,373 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8989, 1.3026, 1.4270, 1.5226, 2.1578, 1.6834, 1.4002, 1.3915], + device='cuda:2'), covar=tensor([0.1694, 0.1945, 0.2291, 0.1703, 0.0911, 0.2066, 0.2848, 0.2267], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0337, 0.0352, 0.0310, 0.0351, 0.0353, 0.0315, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.8290e-05, 7.2280e-05, 7.6138e-05, 6.4898e-05, 7.4669e-05, 7.7067e-05, + 6.8480e-05, 7.6453e-05], device='cuda:2') +2023-04-26 15:12:59,675 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6239, 1.7469, 0.7977, 1.2631, 1.9111, 1.5490, 1.3732, 1.4075], + device='cuda:2'), covar=tensor([0.0538, 0.0431, 0.0440, 0.0615, 0.0295, 0.0597, 0.0596, 0.0667], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 15:13:02,644 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20081.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:03,323 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-04-26 15:13:03,754 INFO [finetune.py:976] (2/7) Epoch 4, batch 2900, loss[loss=0.2729, simple_loss=0.3425, pruned_loss=0.1016, over 4844.00 frames. ], tot_loss[loss=0.2202, simple_loss=0.2779, pruned_loss=0.08121, over 956036.06 frames. ], batch size: 49, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:13:07,523 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20089.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:13,375 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20098.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:21,899 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20112.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:13:29,802 INFO [optim.py:369] (2/7) 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 15:13:47,512 INFO [finetune.py:976] (2/7) Epoch 4, batch 2950, loss[loss=0.242, simple_loss=0.2978, pruned_loss=0.09309, over 4751.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.2824, pruned_loss=0.08312, over 954649.48 frames. ], batch size: 59, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:14:09,929 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20150.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:14:35,350 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20173.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:14:54,005 INFO [finetune.py:976] (2/7) Epoch 4, batch 3000, loss[loss=0.2779, simple_loss=0.3281, pruned_loss=0.1138, over 4892.00 frames. ], tot_loss[loss=0.2256, simple_loss=0.2836, pruned_loss=0.08379, over 955020.91 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:14:54,005 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 15:15:10,814 INFO [finetune.py:1010] (2/7) Epoch 4, validation: loss=0.1632, simple_loss=0.2363, pruned_loss=0.04509, over 2265189.00 frames. +2023-04-26 15:15:10,814 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 15:15:17,700 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3742, 4.0844, 0.6188, 1.9445, 1.8762, 2.3235, 2.3236, 0.9345], + device='cuda:2'), covar=tensor([0.2028, 0.1805, 0.3067, 0.2048, 0.1519, 0.1811, 0.1777, 0.2534], + device='cuda:2'), in_proj_covar=tensor([0.0122, 0.0264, 0.0148, 0.0129, 0.0138, 0.0161, 0.0125, 0.0128], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 15:15:35,632 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 4, batch 3050, loss[loss=0.2228, simple_loss=0.283, pruned_loss=0.08125, over 4781.00 frames. ], tot_loss[loss=0.2263, simple_loss=0.2847, pruned_loss=0.08393, over 953716.88 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:16:35,255 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9505, 1.1927, 5.2003, 4.8303, 4.5114, 4.8338, 4.5434, 4.5698], + device='cuda:2'), covar=tensor([0.6652, 0.6564, 0.0831, 0.1744, 0.1007, 0.1555, 0.1186, 0.1659], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0308, 0.0425, 0.0431, 0.0364, 0.0416, 0.0326, 0.0386], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:16:43,918 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:16:49,407 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 3100, loss[loss=0.2127, simple_loss=0.2665, pruned_loss=0.07943, over 4846.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.2827, pruned_loss=0.08297, over 952551.31 frames. ], batch size: 44, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:17:06,145 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 15:17:13,307 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4690, 1.1415, 0.4826, 1.1466, 1.1508, 1.3814, 1.2442, 1.2178], + device='cuda:2'), covar=tensor([0.0587, 0.0487, 0.0508, 0.0646, 0.0342, 0.0604, 0.0617, 0.0672], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0032], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 15:17:21,578 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20322.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:17:22,123 INFO [optim.py:369] (2/7) 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 15:17:28,233 INFO [finetune.py:976] (2/7) Epoch 4, batch 3150, loss[loss=0.1831, simple_loss=0.2464, pruned_loss=0.05991, over 4751.00 frames. ], tot_loss[loss=0.2225, simple_loss=0.2799, pruned_loss=0.08257, over 952559.12 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:17:57,215 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20376.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:01,415 INFO [finetune.py:976] (2/7) Epoch 4, batch 3200, loss[loss=0.2458, simple_loss=0.2959, pruned_loss=0.09788, over 4873.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2752, pruned_loss=0.08025, over 954070.06 frames. ], batch size: 31, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:18:04,532 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20388.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:25,092 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20418.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:28,518 INFO [optim.py:369] (2/7) 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 15:18:33,001 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5218, 1.6476, 1.5663, 1.7635, 1.6280, 1.8536, 1.6387, 1.6430], + device='cuda:2'), covar=tensor([0.9176, 1.6409, 1.4131, 1.1322, 1.3842, 2.0401, 1.7775, 1.4818], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0400, 0.0321, 0.0327, 0.0351, 0.0412, 0.0386, 0.0339], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:18:34,645 INFO [finetune.py:976] (2/7) Epoch 4, batch 3250, loss[loss=0.201, simple_loss=0.2739, pruned_loss=0.06404, over 4146.00 frames. ], tot_loss[loss=0.2195, simple_loss=0.2769, pruned_loss=0.08109, over 952685.61 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 64.0 +2023-04-26 15:18:42,472 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20445.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:18:45,889 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20468.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:19:11,497 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20479.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:19:13,809 INFO [finetune.py:976] (2/7) Epoch 4, batch 3300, loss[loss=0.2248, simple_loss=0.2846, pruned_loss=0.08253, over 4772.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2813, pruned_loss=0.08273, over 952027.29 frames. ], batch size: 28, lr: 3.96e-03, grad_scale: 32.0 +2023-04-26 15:19:33,456 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.93 vs. limit=5.0 +2023-04-26 15:19:47,325 INFO [optim.py:369] (2/7) 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 15:19:58,662 INFO [finetune.py:976] (2/7) Epoch 4, batch 3350, loss[loss=0.2339, simple_loss=0.2918, pruned_loss=0.08803, over 4902.00 frames. ], tot_loss[loss=0.226, simple_loss=0.2845, pruned_loss=0.08374, over 954074.13 frames. ], batch size: 36, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:20:39,486 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 3400, loss[loss=0.2148, simple_loss=0.2863, pruned_loss=0.07161, over 4810.00 frames. ], tot_loss[loss=0.2263, simple_loss=0.2851, pruned_loss=0.08377, over 952835.07 frames. ], batch size: 25, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:21:04,548 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:21:34,360 INFO [finetune.py:976] (2/7) Epoch 4, batch 3450, loss[loss=0.1972, simple_loss=0.2492, pruned_loss=0.07256, over 4700.00 frames. ], tot_loss[loss=0.2242, simple_loss=0.2833, pruned_loss=0.08251, over 954715.19 frames. ], batch size: 54, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:21:44,643 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20649.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:22:19,757 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20676.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:22:23,911 INFO [finetune.py:976] (2/7) Epoch 4, batch 3500, loss[loss=0.2194, simple_loss=0.2732, pruned_loss=0.08283, over 4781.00 frames. ], tot_loss[loss=0.2219, simple_loss=0.2804, pruned_loss=0.08165, over 956942.79 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:22:58,017 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:22:58,703 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9062, 1.3316, 1.6717, 1.8857, 1.5574, 1.2910, 0.7963, 1.3131], + device='cuda:2'), covar=tensor([0.4595, 0.5599, 0.2545, 0.4075, 0.5085, 0.4245, 0.6958, 0.4517], + device='cuda:2'), in_proj_covar=tensor([0.0271, 0.0264, 0.0222, 0.0334, 0.0224, 0.0232, 0.0250, 0.0201], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:23:03,439 INFO [finetune.py:976] (2/7) Epoch 4, batch 3550, loss[loss=0.2467, simple_loss=0.2986, pruned_loss=0.09741, over 4890.00 frames. ], tot_loss[loss=0.219, simple_loss=0.2773, pruned_loss=0.08037, over 959086.64 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:23:15,956 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20744.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:23:16,587 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20745.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:23:33,276 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7381, 1.6135, 1.9814, 2.1279, 2.0033, 1.6838, 1.7698, 1.8825], + device='cuda:2'), covar=tensor([1.4734, 1.9206, 2.2900, 2.1963, 1.5401, 2.6750, 2.6135, 1.9557], + device='cuda:2'), in_proj_covar=tensor([0.0430, 0.0473, 0.0562, 0.0580, 0.0462, 0.0488, 0.0500, 0.0506], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:23:48,040 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20774.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:24:00,554 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4688, 1.6652, 1.6627, 1.8223, 1.6888, 1.8877, 1.7717, 1.7599], + device='cuda:2'), covar=tensor([0.9315, 1.4445, 1.2063, 1.0591, 1.2926, 1.9805, 1.4513, 1.2656], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0400, 0.0321, 0.0328, 0.0350, 0.0412, 0.0386, 0.0341], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:24:04,022 INFO [finetune.py:976] (2/7) Epoch 4, batch 3600, loss[loss=0.2171, simple_loss=0.2574, pruned_loss=0.08835, over 4764.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.2746, pruned_loss=0.0798, over 958960.59 frames. ], batch size: 28, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:24:10,254 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20793.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:24:31,214 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:24:38,732 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1443, 1.3852, 1.2743, 1.6944, 1.5645, 1.6305, 1.3313, 2.4401], + device='cuda:2'), covar=tensor([0.0662, 0.0801, 0.0877, 0.1300, 0.0688, 0.0503, 0.0820, 0.0257], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0041, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 15:24:43,526 INFO [finetune.py:976] (2/7) Epoch 4, batch 3650, loss[loss=0.3288, simple_loss=0.3678, pruned_loss=0.1449, over 4829.00 frames. ], tot_loss[loss=0.2209, simple_loss=0.2783, pruned_loss=0.08178, over 957143.84 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:25:00,499 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20860.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:25:09,794 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-26 15:25:16,800 INFO [finetune.py:976] (2/7) Epoch 4, batch 3700, loss[loss=0.2535, simple_loss=0.3034, pruned_loss=0.1018, over 4758.00 frames. ], tot_loss[loss=0.2223, simple_loss=0.2807, pruned_loss=0.0819, over 955403.03 frames. ], batch size: 28, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:25:24,324 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 15:25:25,359 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4275, 0.8685, 1.3038, 1.7709, 1.5545, 1.3135, 1.3571, 1.4101], + device='cuda:2'), covar=tensor([1.3205, 1.7953, 1.8404, 2.1504, 1.5144, 2.1086, 2.1230, 1.6968], + device='cuda:2'), in_proj_covar=tensor([0.0431, 0.0474, 0.0563, 0.0581, 0.0464, 0.0488, 0.0501, 0.0506], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:25:29,409 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.3084, 4.2349, 3.1530, 4.9610, 4.2441, 4.2568, 2.1913, 4.2006], + device='cuda:2'), covar=tensor([0.1393, 0.0949, 0.3341, 0.0984, 0.2604, 0.1427, 0.5009, 0.2164], + device='cuda:2'), in_proj_covar=tensor([0.0250, 0.0222, 0.0260, 0.0314, 0.0309, 0.0257, 0.0276, 0.0278], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:25:32,430 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20908.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:25:50,199 INFO [optim.py:369] (2/7) 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 15:25:53,800 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 15:26:01,984 INFO [finetune.py:976] (2/7) Epoch 4, batch 3750, loss[loss=0.2216, simple_loss=0.2904, pruned_loss=0.0764, over 4815.00 frames. ], tot_loss[loss=0.223, simple_loss=0.2815, pruned_loss=0.08223, over 954170.78 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:26:10,578 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9334, 2.8962, 2.2149, 3.3246, 2.8803, 2.9318, 1.0955, 2.7927], + device='cuda:2'), covar=tensor([0.2115, 0.1416, 0.2922, 0.2555, 0.3044, 0.1972, 0.5905, 0.2735], + device='cuda:2'), in_proj_covar=tensor([0.0249, 0.0221, 0.0259, 0.0313, 0.0308, 0.0256, 0.0275, 0.0278], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:26:14,219 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20944.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:27:02,460 INFO [finetune.py:976] (2/7) Epoch 4, batch 3800, loss[loss=0.1606, simple_loss=0.2444, pruned_loss=0.03837, over 4917.00 frames. ], tot_loss[loss=0.2236, simple_loss=0.2822, pruned_loss=0.08246, over 951846.89 frames. ], batch size: 42, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:27:09,402 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-26 15:27:46,870 INFO [optim.py:369] (2/7) 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 15:28:05,808 INFO [finetune.py:976] (2/7) Epoch 4, batch 3850, loss[loss=0.2832, simple_loss=0.3276, pruned_loss=0.1194, over 4826.00 frames. ], tot_loss[loss=0.2224, simple_loss=0.2815, pruned_loss=0.08168, over 953528.41 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:28:18,881 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21044.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:28:29,648 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6297, 2.0418, 1.7774, 1.9531, 1.6516, 1.6809, 1.7109, 1.4397], + device='cuda:2'), covar=tensor([0.2388, 0.1543, 0.0962, 0.1498, 0.3462, 0.1501, 0.1916, 0.2686], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0334, 0.0242, 0.0307, 0.0321, 0.0285, 0.0274, 0.0297], + device='cuda:2'), out_proj_covar=tensor([1.2726e-04, 1.3581e-04, 9.8589e-05, 1.2366e-04, 1.3231e-04, 1.1515e-04, + 1.1286e-04, 1.2006e-04], device='cuda:2') +2023-04-26 15:28:29,654 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21052.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:28:54,708 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21074.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:11,950 INFO [finetune.py:976] (2/7) Epoch 4, batch 3900, loss[loss=0.2124, simple_loss=0.2597, pruned_loss=0.08253, over 4819.00 frames. ], tot_loss[loss=0.2209, simple_loss=0.279, pruned_loss=0.08137, over 953562.00 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:29:14,399 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2078, 1.4956, 1.4783, 1.8181, 1.6563, 1.8899, 1.4649, 3.4125], + device='cuda:2'), covar=tensor([0.0761, 0.0806, 0.0866, 0.1324, 0.0681, 0.0552, 0.0800, 0.0154], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0042, 0.0041, 0.0041, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 15:29:16,774 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4856, 1.0379, 1.4987, 1.8456, 1.6812, 1.4478, 1.5138, 1.5376], + device='cuda:2'), covar=tensor([1.2099, 1.6001, 1.6424, 1.8924, 1.3836, 1.9643, 1.9148, 1.5704], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0472, 0.0562, 0.0580, 0.0462, 0.0487, 0.0500, 0.0505], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:29:18,469 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21092.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:20,431 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-26 15:29:31,297 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21113.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:34,777 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21118.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:37,128 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21122.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:29:38,884 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 4, batch 3950, loss[loss=0.1937, simple_loss=0.249, pruned_loss=0.06916, over 4815.00 frames. ], tot_loss[loss=0.2175, simple_loss=0.2754, pruned_loss=0.07983, over 954993.88 frames. ], batch size: 51, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:30:03,953 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3522, 1.5862, 1.5171, 1.8855, 1.6960, 1.9247, 1.5329, 3.4296], + device='cuda:2'), covar=tensor([0.0651, 0.0781, 0.0772, 0.1165, 0.0648, 0.0513, 0.0720, 0.0148], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0040, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 15:30:11,071 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3222, 1.5208, 1.4688, 1.7391, 1.6403, 1.8189, 1.5167, 2.7992], + device='cuda:2'), covar=tensor([0.0606, 0.0637, 0.0668, 0.0963, 0.0514, 0.0594, 0.0640, 0.0249], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0040, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 15:30:14,101 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 4000, loss[loss=0.1948, simple_loss=0.2643, pruned_loss=0.06266, over 4731.00 frames. ], tot_loss[loss=0.2174, simple_loss=0.275, pruned_loss=0.07995, over 954764.42 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:30:37,103 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4260, 1.2454, 1.5465, 1.5683, 1.3506, 1.1605, 1.3368, 0.9012], + device='cuda:2'), covar=tensor([0.0601, 0.1066, 0.0812, 0.0837, 0.0861, 0.1361, 0.0756, 0.1055], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0078, 0.0077, 0.0071, 0.0082, 0.0098, 0.0086, 0.0079], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 15:30:41,965 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5394, 2.2515, 1.4692, 1.4338, 1.1672, 1.2211, 1.4827, 1.1044], + device='cuda:2'), covar=tensor([0.1883, 0.1592, 0.2030, 0.2315, 0.3170, 0.2194, 0.1436, 0.2482], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0220, 0.0181, 0.0210, 0.0219, 0.0187, 0.0175, 0.0198], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 15:30:45,237 INFO [optim.py:369] (2/7) 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 15:30:47,239 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.94 vs. limit=5.0 +2023-04-26 15:30:50,170 INFO [finetune.py:976] (2/7) Epoch 4, batch 4050, loss[loss=0.1999, simple_loss=0.2778, pruned_loss=0.06103, over 4813.00 frames. ], tot_loss[loss=0.22, simple_loss=0.2777, pruned_loss=0.08111, over 952294.38 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:30:58,878 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21244.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:14,643 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7525, 0.9346, 1.1910, 1.3410, 1.3610, 1.5503, 1.2499, 1.2521], + device='cuda:2'), covar=tensor([0.7569, 1.1318, 0.9512, 0.8765, 1.0450, 1.6611, 1.1078, 1.0612], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0398, 0.0319, 0.0326, 0.0349, 0.0411, 0.0384, 0.0339], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:31:17,042 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9599, 2.5799, 0.9788, 1.2411, 1.8710, 1.2168, 3.5869, 1.5841], + device='cuda:2'), covar=tensor([0.0714, 0.0857, 0.0928, 0.1324, 0.0558, 0.0968, 0.0207, 0.0665], + device='cuda:2'), in_proj_covar=tensor([0.0054, 0.0070, 0.0053, 0.0049, 0.0054, 0.0054, 0.0083, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 15:31:20,011 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-26 15:31:23,490 INFO [finetune.py:976] (2/7) Epoch 4, batch 4100, loss[loss=0.2717, simple_loss=0.3219, pruned_loss=0.1108, over 4921.00 frames. ], tot_loss[loss=0.2207, simple_loss=0.2798, pruned_loss=0.08083, over 953142.24 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 16.0 +2023-04-26 15:31:29,564 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21292.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:34,700 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5710, 1.9726, 1.7106, 1.8826, 1.5814, 1.5725, 1.6202, 1.3438], + device='cuda:2'), covar=tensor([0.2116, 0.1641, 0.0937, 0.1439, 0.3747, 0.1555, 0.2054, 0.2707], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0335, 0.0242, 0.0307, 0.0322, 0.0286, 0.0274, 0.0297], + device='cuda:2'), out_proj_covar=tensor([1.2762e-04, 1.3647e-04, 9.8543e-05, 1.2378e-04, 1.3290e-04, 1.1552e-04, + 1.1306e-04, 1.1994e-04], device='cuda:2') +2023-04-26 15:31:43,174 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21312.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:50,918 INFO [optim.py:369] (2/7) 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 15:31:52,893 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21328.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:31:55,798 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6217, 1.5773, 2.0081, 1.9983, 1.5932, 1.2698, 1.8250, 1.1239], + device='cuda:2'), covar=tensor([0.1347, 0.0872, 0.0617, 0.1154, 0.1036, 0.1481, 0.0953, 0.1285], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0078, 0.0076, 0.0070, 0.0082, 0.0097, 0.0086, 0.0079], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 15:31:56,304 INFO [finetune.py:976] (2/7) Epoch 4, batch 4150, loss[loss=0.1925, simple_loss=0.2518, pruned_loss=0.06665, over 4223.00 frames. ], tot_loss[loss=0.222, simple_loss=0.281, pruned_loss=0.08152, over 952366.49 frames. ], batch size: 65, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:32:40,014 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21373.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:32:46,017 INFO [finetune.py:976] (2/7) Epoch 4, batch 4200, loss[loss=0.2296, simple_loss=0.2883, pruned_loss=0.08545, over 4731.00 frames. ], tot_loss[loss=0.2224, simple_loss=0.2817, pruned_loss=0.08152, over 953532.41 frames. ], batch size: 59, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:32:50,260 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21389.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:32:59,255 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0697, 1.2419, 1.4326, 1.6148, 1.5862, 1.7511, 1.4835, 1.5181], + device='cuda:2'), covar=tensor([0.8465, 1.2204, 1.0408, 0.9308, 1.1099, 1.6133, 1.2823, 1.1880], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0399, 0.0320, 0.0327, 0.0349, 0.0412, 0.0385, 0.0340], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:33:21,116 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 15:33:21,602 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:33:52,191 INFO [finetune.py:976] (2/7) Epoch 4, batch 4250, loss[loss=0.1895, simple_loss=0.256, pruned_loss=0.06151, over 4825.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2789, pruned_loss=0.08033, over 955159.16 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:33:52,285 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9432, 1.9335, 4.3942, 4.1296, 3.8755, 4.0817, 4.1335, 3.8888], + device='cuda:2'), covar=tensor([0.6197, 0.4935, 0.1133, 0.1785, 0.1112, 0.1694, 0.1278, 0.1815], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0305, 0.0419, 0.0426, 0.0362, 0.0415, 0.0324, 0.0382], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:34:08,311 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-04-26 15:34:17,774 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 15:34:30,946 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21474.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:34:41,943 INFO [finetune.py:976] (2/7) Epoch 4, batch 4300, loss[loss=0.1745, simple_loss=0.2362, pruned_loss=0.05638, over 4809.00 frames. ], tot_loss[loss=0.2183, simple_loss=0.2767, pruned_loss=0.07997, over 955431.47 frames. ], batch size: 25, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:35:27,095 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 4, batch 4350, loss[loss=0.2209, simple_loss=0.2741, pruned_loss=0.08387, over 4822.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.2732, pruned_loss=0.07853, over 955700.87 frames. ], batch size: 45, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:35:43,249 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6446, 2.1309, 1.9136, 2.1116, 1.6103, 1.7470, 1.9700, 1.4690], + device='cuda:2'), covar=tensor([0.2150, 0.1395, 0.0925, 0.1359, 0.3234, 0.1429, 0.1733, 0.2774], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0332, 0.0241, 0.0304, 0.0321, 0.0285, 0.0273, 0.0296], + device='cuda:2'), out_proj_covar=tensor([1.2680e-04, 1.3540e-04, 9.7948e-05, 1.2264e-04, 1.3234e-04, 1.1513e-04, + 1.1237e-04, 1.1948e-04], device='cuda:2') +2023-04-26 15:36:37,729 INFO [finetune.py:976] (2/7) Epoch 4, batch 4400, loss[loss=0.2166, simple_loss=0.2916, pruned_loss=0.07079, over 4814.00 frames. ], tot_loss[loss=0.2162, simple_loss=0.2745, pruned_loss=0.07895, over 957007.82 frames. ], batch size: 39, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:36:51,074 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21595.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:09,358 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21613.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:12,903 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9902, 2.4330, 1.0607, 1.2926, 1.9670, 1.3039, 3.0680, 1.5562], + device='cuda:2'), covar=tensor([0.0720, 0.0555, 0.0798, 0.1345, 0.0523, 0.1008, 0.0418, 0.0748], + device='cuda:2'), in_proj_covar=tensor([0.0054, 0.0071, 0.0053, 0.0050, 0.0054, 0.0055, 0.0083, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 15:37:17,113 INFO [optim.py:369] (2/7) 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 15:37:22,023 INFO [finetune.py:976] (2/7) Epoch 4, batch 4450, loss[loss=0.269, simple_loss=0.3198, pruned_loss=0.1091, over 4739.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2784, pruned_loss=0.08061, over 956852.13 frames. ], batch size: 59, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:37:27,053 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8648, 2.3123, 2.3201, 2.7030, 2.3298, 2.4605, 2.2448, 4.8843], + device='cuda:2'), covar=tensor([0.0597, 0.0678, 0.0662, 0.1011, 0.0595, 0.0558, 0.0653, 0.0124], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0047, 0.0042, 0.0041, 0.0041, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 15:37:36,748 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21656.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:46,469 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21668.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:48,273 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5160, 1.3026, 1.6723, 1.6276, 1.3623, 1.2104, 1.3836, 0.8870], + device='cuda:2'), covar=tensor([0.0597, 0.0893, 0.0646, 0.0912, 0.0995, 0.1557, 0.0790, 0.1164], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0077, 0.0075, 0.0070, 0.0081, 0.0097, 0.0085, 0.0078], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 15:37:50,675 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21674.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:37:56,116 INFO [finetune.py:976] (2/7) Epoch 4, batch 4500, loss[loss=0.2274, simple_loss=0.29, pruned_loss=0.08238, over 4752.00 frames. ], tot_loss[loss=0.221, simple_loss=0.2801, pruned_loss=0.08091, over 956663.08 frames. ], batch size: 54, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:37:56,778 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21684.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:02,522 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-26 15:38:10,180 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3574, 1.3727, 1.3623, 0.9498, 1.4412, 1.1319, 1.7743, 1.2355], + device='cuda:2'), covar=tensor([0.3845, 0.1671, 0.5013, 0.2851, 0.1468, 0.2109, 0.1529, 0.4906], + device='cuda:2'), in_proj_covar=tensor([0.0355, 0.0356, 0.0439, 0.0369, 0.0401, 0.0382, 0.0395, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:38:11,960 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21708.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:21,108 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 15:38:24,949 INFO [optim.py:369] (2/7) 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 15:38:29,903 INFO [finetune.py:976] (2/7) Epoch 4, batch 4550, loss[loss=0.1808, simple_loss=0.2418, pruned_loss=0.0599, over 4799.00 frames. ], tot_loss[loss=0.2222, simple_loss=0.2818, pruned_loss=0.08128, over 955227.22 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:38:44,652 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21756.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:49,067 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:38:58,612 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21774.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:39:01,091 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7500, 4.2132, 0.8798, 2.0940, 2.2643, 2.5012, 2.4539, 0.9362], + device='cuda:2'), covar=tensor([0.1333, 0.1031, 0.2114, 0.1475, 0.1091, 0.1260, 0.1413, 0.2167], + device='cuda:2'), in_proj_covar=tensor([0.0122, 0.0263, 0.0147, 0.0129, 0.0138, 0.0160, 0.0125, 0.0128], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 15:39:04,081 INFO [finetune.py:976] (2/7) Epoch 4, batch 4600, loss[loss=0.2418, simple_loss=0.3026, pruned_loss=0.09052, over 4847.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.281, pruned_loss=0.08118, over 954383.09 frames. ], batch size: 49, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:39:30,458 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21822.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:39:36,929 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21824.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:39:37,418 INFO [optim.py:369] (2/7) 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 15:39:48,033 INFO [finetune.py:976] (2/7) Epoch 4, batch 4650, loss[loss=0.2369, simple_loss=0.2941, pruned_loss=0.0898, over 4804.00 frames. ], tot_loss[loss=0.2196, simple_loss=0.2783, pruned_loss=0.08047, over 956023.68 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:40:04,095 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21850.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:40:33,084 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2786, 1.5766, 1.4936, 1.7638, 1.6474, 1.9481, 1.4388, 3.4050], + device='cuda:2'), covar=tensor([0.0699, 0.0780, 0.0800, 0.1292, 0.0649, 0.0505, 0.0759, 0.0173], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0047, 0.0042, 0.0041, 0.0041, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 15:40:49,526 INFO [finetune.py:976] (2/7) Epoch 4, batch 4700, loss[loss=0.2103, simple_loss=0.2696, pruned_loss=0.07556, over 4864.00 frames. ], tot_loss[loss=0.2155, simple_loss=0.2741, pruned_loss=0.07848, over 956735.75 frames. ], batch size: 31, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:41:12,782 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21911.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:21,741 INFO [optim.py:369] (2/7) 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 15:41:28,679 INFO [finetune.py:976] (2/7) Epoch 4, batch 4750, loss[loss=0.213, simple_loss=0.2733, pruned_loss=0.07633, over 4767.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2712, pruned_loss=0.07737, over 958636.50 frames. ], batch size: 28, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:41:40,251 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21951.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:42,107 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0270, 2.0651, 1.9018, 1.6319, 2.0449, 1.7466, 2.7677, 1.6198], + device='cuda:2'), covar=tensor([0.4497, 0.1889, 0.5590, 0.3652, 0.2123, 0.2810, 0.1708, 0.5154], + device='cuda:2'), in_proj_covar=tensor([0.0358, 0.0358, 0.0442, 0.0372, 0.0404, 0.0386, 0.0398, 0.0425], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:41:50,661 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21968.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:51,755 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21969.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:41:52,453 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:02,274 INFO [finetune.py:976] (2/7) Epoch 4, batch 4800, loss[loss=0.2702, simple_loss=0.3336, pruned_loss=0.1034, over 4847.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2741, pruned_loss=0.0788, over 956393.29 frames. ], batch size: 49, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:42:02,999 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21984.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:05,398 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9518, 1.7940, 2.0349, 2.3770, 2.2207, 1.7889, 1.4507, 1.9266], + device='cuda:2'), covar=tensor([0.1062, 0.1221, 0.0696, 0.0673, 0.0711, 0.1112, 0.1111, 0.0744], + device='cuda:2'), in_proj_covar=tensor([0.0210, 0.0210, 0.0188, 0.0184, 0.0184, 0.0200, 0.0172, 0.0194], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:42:18,352 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-26 15:42:39,319 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22016.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:50,413 INFO [optim.py:369] (2/7) 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 15:42:54,226 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22031.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:55,317 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22032.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:42:56,368 INFO [finetune.py:976] (2/7) Epoch 4, batch 4850, loss[loss=0.2172, simple_loss=0.2796, pruned_loss=0.07742, over 4821.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.279, pruned_loss=0.08026, over 956093.07 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:43:12,696 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0192, 1.4336, 1.2772, 1.6886, 1.4894, 1.6867, 1.3642, 3.0245], + device='cuda:2'), covar=tensor([0.0741, 0.0770, 0.0862, 0.1251, 0.0706, 0.0541, 0.0758, 0.0181], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0041, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 15:43:45,552 INFO [finetune.py:976] (2/7) Epoch 4, batch 4900, loss[loss=0.2506, simple_loss=0.3239, pruned_loss=0.08868, over 4851.00 frames. ], tot_loss[loss=0.2208, simple_loss=0.2803, pruned_loss=0.08061, over 956004.60 frames. ], batch size: 44, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:44:08,994 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7511, 2.2479, 1.8418, 2.1790, 1.7954, 1.8271, 1.9564, 1.4767], + device='cuda:2'), covar=tensor([0.2768, 0.1921, 0.1212, 0.1615, 0.3609, 0.1701, 0.2418, 0.3410], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0332, 0.0241, 0.0305, 0.0322, 0.0286, 0.0274, 0.0296], + device='cuda:2'), out_proj_covar=tensor([1.2643e-04, 1.3501e-04, 9.8100e-05, 1.2267e-04, 1.3299e-04, 1.1569e-04, + 1.1284e-04, 1.1956e-04], device='cuda:2') +2023-04-26 15:44:15,742 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:44:22,977 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:44:24,752 INFO [finetune.py:976] (2/7) Epoch 4, batch 4950, loss[loss=0.2295, simple_loss=0.2949, pruned_loss=0.08208, over 4723.00 frames. ], tot_loss[loss=0.221, simple_loss=0.2812, pruned_loss=0.08042, over 957868.68 frames. ], batch size: 54, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:44:41,591 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-26 15:44:51,204 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1840, 2.8296, 1.1821, 1.3742, 1.9631, 1.3804, 3.4146, 1.6370], + device='cuda:2'), covar=tensor([0.0779, 0.1272, 0.1151, 0.1547, 0.0705, 0.1251, 0.0414, 0.0939], + device='cuda:2'), in_proj_covar=tensor([0.0054, 0.0071, 0.0053, 0.0049, 0.0054, 0.0055, 0.0083, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0008, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 15:44:53,682 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22176.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:44:58,355 INFO [finetune.py:976] (2/7) Epoch 4, batch 5000, loss[loss=0.1891, simple_loss=0.2562, pruned_loss=0.06102, over 4754.00 frames. ], tot_loss[loss=0.2181, simple_loss=0.2785, pruned_loss=0.07883, over 957822.95 frames. ], batch size: 28, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:45:03,887 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22191.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:14,450 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6002, 1.6917, 1.6704, 1.2296, 1.7640, 1.4649, 2.2697, 1.4637], + device='cuda:2'), covar=tensor([0.4209, 0.1878, 0.5366, 0.3377, 0.1841, 0.2436, 0.1529, 0.4945], + device='cuda:2'), in_proj_covar=tensor([0.0357, 0.0358, 0.0440, 0.0372, 0.0403, 0.0384, 0.0399, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:45:15,020 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22206.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:20,501 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:45:37,942 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-26 15:45:42,549 INFO [finetune.py:976] (2/7) Epoch 4, batch 5050, loss[loss=0.188, simple_loss=0.2529, pruned_loss=0.06148, over 4823.00 frames. ], tot_loss[loss=0.2172, simple_loss=0.2765, pruned_loss=0.07893, over 953900.60 frames. ], batch size: 30, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:45:45,133 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:45:45,142 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:06,551 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22251.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:28,165 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22269.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:32,409 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22276.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:41,824 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.7095, 4.7067, 3.1227, 5.5175, 4.7757, 4.7829, 2.4011, 4.6818], + device='cuda:2'), covar=tensor([0.1431, 0.0834, 0.2969, 0.0776, 0.3022, 0.1565, 0.5035, 0.2096], + device='cuda:2'), in_proj_covar=tensor([0.0252, 0.0224, 0.0262, 0.0317, 0.0312, 0.0260, 0.0278, 0.0281], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 15:46:42,370 INFO [finetune.py:976] (2/7) Epoch 4, batch 5100, loss[loss=0.1619, simple_loss=0.2292, pruned_loss=0.04734, over 4776.00 frames. ], tot_loss[loss=0.2144, simple_loss=0.2728, pruned_loss=0.07803, over 950395.99 frames. ], batch size: 28, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:46:52,065 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22298.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:46:53,599 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22299.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:47:03,023 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8751, 1.3999, 1.4303, 1.5374, 2.1609, 1.7289, 1.3695, 1.3995], + device='cuda:2'), covar=tensor([0.1662, 0.1933, 0.2232, 0.1698, 0.0931, 0.1919, 0.2590, 0.1961], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0336, 0.0354, 0.0311, 0.0350, 0.0351, 0.0316, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.8490e-05, 7.1986e-05, 7.6667e-05, 6.5053e-05, 7.4248e-05, 7.6564e-05, + 6.8654e-05, 7.6463e-05], device='cuda:2') +2023-04-26 15:47:06,007 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22317.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:47:10,807 INFO [optim.py:369] (2/7) 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 15:47:11,502 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22326.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:47:15,714 INFO [finetune.py:976] (2/7) Epoch 4, batch 5150, loss[loss=0.1841, simple_loss=0.2535, pruned_loss=0.05737, over 4813.00 frames. ], tot_loss[loss=0.215, simple_loss=0.273, pruned_loss=0.07856, over 951938.94 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:47:54,331 INFO [finetune.py:976] (2/7) Epoch 4, batch 5200, loss[loss=0.2343, simple_loss=0.3104, pruned_loss=0.07904, over 4790.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2761, pruned_loss=0.07977, over 951705.85 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:48:01,214 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22385.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:48:05,939 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0543, 2.5717, 1.0535, 1.3524, 1.9823, 1.2436, 3.3511, 1.6857], + device='cuda:2'), covar=tensor([0.0669, 0.0792, 0.0912, 0.1275, 0.0497, 0.0957, 0.0224, 0.0622], + device='cuda:2'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0054, 0.0054, 0.0082, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 15:48:18,026 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9609, 1.3925, 1.4027, 1.5657, 2.1613, 1.8000, 1.4239, 1.4444], + device='cuda:2'), covar=tensor([0.1599, 0.1711, 0.2295, 0.1395, 0.0766, 0.1672, 0.2268, 0.1845], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0335, 0.0353, 0.0310, 0.0349, 0.0350, 0.0316, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.8220e-05, 7.1880e-05, 7.6503e-05, 6.4779e-05, 7.3884e-05, 7.6287e-05, + 6.8501e-05, 7.6373e-05], device='cuda:2') +2023-04-26 15:48:23,822 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-26 15:48:30,901 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22419.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:48:32,180 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-26 15:48:34,453 INFO [optim.py:369] (2/7) 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 15:48:39,401 INFO [finetune.py:976] (2/7) Epoch 4, batch 5250, loss[loss=0.2169, simple_loss=0.2789, pruned_loss=0.07742, over 4813.00 frames. ], tot_loss[loss=0.2203, simple_loss=0.2793, pruned_loss=0.08064, over 953934.05 frames. ], batch size: 39, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:48:47,952 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22446.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:49:09,423 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:49:25,136 INFO [finetune.py:976] (2/7) Epoch 4, batch 5300, loss[loss=0.235, simple_loss=0.2917, pruned_loss=0.08915, over 4869.00 frames. ], tot_loss[loss=0.2211, simple_loss=0.2802, pruned_loss=0.08105, over 953180.85 frames. ], batch size: 31, lr: 3.95e-03, grad_scale: 16.0 +2023-04-26 15:49:32,900 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22486.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:49:56,922 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:49:59,225 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.10 vs. limit=5.0 +2023-04-26 15:50:15,635 INFO [optim.py:369] (2/7) 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 15:50:19,983 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 5350, loss[loss=0.2427, simple_loss=0.2927, pruned_loss=0.09629, over 4835.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.2807, pruned_loss=0.0811, over 953447.81 frames. ], batch size: 30, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:50:33,853 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22554.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:50:46,633 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 5400, loss[loss=0.1781, simple_loss=0.2405, pruned_loss=0.05782, over 4822.00 frames. ], tot_loss[loss=0.2184, simple_loss=0.2778, pruned_loss=0.07953, over 954816.85 frames. ], batch size: 39, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:50:59,990 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:51:39,252 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22626.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:51:47,647 INFO [finetune.py:976] (2/7) Epoch 4, batch 5450, loss[loss=0.2007, simple_loss=0.258, pruned_loss=0.07169, over 4908.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.2754, pruned_loss=0.07939, over 954308.18 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:51:51,451 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22639.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:52:00,548 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4964, 1.1337, 0.3792, 1.2134, 1.2899, 1.3756, 1.2406, 1.2850], + device='cuda:2'), covar=tensor([0.0566, 0.0469, 0.0489, 0.0641, 0.0311, 0.0594, 0.0589, 0.0655], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0031, 0.0030, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 15:52:19,096 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6476, 1.5467, 1.7367, 1.9993, 1.9424, 1.5404, 1.2803, 1.6865], + device='cuda:2'), covar=tensor([0.0847, 0.1102, 0.0707, 0.0544, 0.0596, 0.0868, 0.0930, 0.0634], + device='cuda:2'), in_proj_covar=tensor([0.0206, 0.0207, 0.0184, 0.0180, 0.0180, 0.0196, 0.0169, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:52:39,600 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22674.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:52:50,915 INFO [finetune.py:976] (2/7) Epoch 4, batch 5500, loss[loss=0.2013, simple_loss=0.2648, pruned_loss=0.06892, over 4769.00 frames. ], tot_loss[loss=0.214, simple_loss=0.2721, pruned_loss=0.07797, over 953975.75 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:53:12,619 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22700.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 15:53:34,588 INFO [optim.py:369] (2/7) 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 15:53:40,512 INFO [finetune.py:976] (2/7) Epoch 4, batch 5550, loss[loss=0.2839, simple_loss=0.3406, pruned_loss=0.1135, over 4896.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2739, pruned_loss=0.07901, over 952221.76 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:53:45,459 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22741.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:11,071 INFO [finetune.py:976] (2/7) Epoch 4, batch 5600, loss[loss=0.216, simple_loss=0.2881, pruned_loss=0.07189, over 4906.00 frames. ], tot_loss[loss=0.2191, simple_loss=0.2777, pruned_loss=0.08023, over 954099.50 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:54:12,892 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22786.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:18,056 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22794.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:37,300 INFO [optim.py:369] (2/7) 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 15:54:37,404 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22825.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:37,414 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2466, 1.5396, 1.5494, 2.1129, 2.3310, 2.0030, 1.7928, 1.7219], + device='cuda:2'), covar=tensor([0.2638, 0.2316, 0.2430, 0.2307, 0.1546, 0.2337, 0.3268, 0.2317], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0336, 0.0353, 0.0309, 0.0349, 0.0351, 0.0316, 0.0355], + device='cuda:2'), out_proj_covar=tensor([6.8184e-05, 7.1996e-05, 7.6556e-05, 6.4657e-05, 7.3892e-05, 7.6391e-05, + 6.8539e-05, 7.6683e-05], device='cuda:2') +2023-04-26 15:54:41,472 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22832.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:42,026 INFO [finetune.py:976] (2/7) Epoch 4, batch 5650, loss[loss=0.2063, simple_loss=0.2773, pruned_loss=0.06759, over 4890.00 frames. ], tot_loss[loss=0.2211, simple_loss=0.2806, pruned_loss=0.08087, over 953078.21 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:54:42,646 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22834.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:54:42,867 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-04-26 15:54:55,186 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22855.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:55:15,946 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22871.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:55:27,436 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 4, batch 5700, loss[loss=0.1965, simple_loss=0.2375, pruned_loss=0.07774, over 4169.00 frames. ], tot_loss[loss=0.2175, simple_loss=0.2761, pruned_loss=0.07947, over 937623.96 frames. ], batch size: 18, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:55:31,122 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22886.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:55:40,634 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22893.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:56:19,323 INFO [finetune.py:976] (2/7) Epoch 5, batch 0, loss[loss=0.2284, simple_loss=0.2892, pruned_loss=0.08387, over 4760.00 frames. ], tot_loss[loss=0.2284, simple_loss=0.2892, pruned_loss=0.08387, over 4760.00 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:56:19,323 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 15:56:28,482 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5091, 1.3510, 1.6549, 1.6144, 1.4064, 1.2405, 1.4606, 0.8960], + device='cuda:2'), covar=tensor([0.0714, 0.0912, 0.0946, 0.0732, 0.0808, 0.1454, 0.0671, 0.1183], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0077, 0.0075, 0.0069, 0.0080, 0.0096, 0.0083, 0.0078], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 15:56:30,075 INFO [finetune.py:1010] (2/7) Epoch 5, validation: loss=0.1632, simple_loss=0.2369, pruned_loss=0.04473, over 2265189.00 frames. +2023-04-26 15:56:30,075 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 15:56:33,903 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7752, 1.5059, 1.9574, 2.1551, 2.0203, 1.6612, 1.8217, 1.8631], + device='cuda:2'), covar=tensor([1.2053, 1.6608, 1.8351, 1.7142, 1.2768, 2.0574, 2.0257, 1.6425], + device='cuda:2'), in_proj_covar=tensor([0.0426, 0.0466, 0.0554, 0.0573, 0.0458, 0.0483, 0.0495, 0.0498], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 15:56:35,031 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22919.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:56:38,534 INFO [optim.py:369] (2/7) 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 15:56:41,649 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1461, 2.5044, 0.8027, 1.4272, 1.5952, 1.8593, 1.6204, 0.8563], + device='cuda:2'), covar=tensor([0.1372, 0.1071, 0.1774, 0.1357, 0.1049, 0.0932, 0.1530, 0.1660], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0264, 0.0149, 0.0130, 0.0139, 0.0162, 0.0126, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 15:56:48,361 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22941.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:56:49,646 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6015, 1.1762, 1.2490, 1.3090, 1.8460, 1.4417, 1.1139, 1.2559], + device='cuda:2'), covar=tensor([0.1596, 0.1622, 0.2085, 0.1516, 0.0773, 0.1559, 0.2452, 0.1945], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0335, 0.0353, 0.0309, 0.0348, 0.0350, 0.0315, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.8047e-05, 7.1806e-05, 7.6438e-05, 6.4501e-05, 7.3811e-05, 7.6164e-05, + 6.8455e-05, 7.6506e-05], device='cuda:2') +2023-04-26 15:57:02,116 INFO [finetune.py:976] (2/7) Epoch 5, batch 50, loss[loss=0.2183, simple_loss=0.278, pruned_loss=0.07935, over 4862.00 frames. ], tot_loss[loss=0.2236, simple_loss=0.2826, pruned_loss=0.08231, over 216753.44 frames. ], batch size: 31, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:57:20,143 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.47 vs. limit=5.0 +2023-04-26 15:57:30,169 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22995.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 15:57:52,331 INFO [finetune.py:976] (2/7) Epoch 5, batch 100, loss[loss=0.2183, simple_loss=0.2751, pruned_loss=0.08075, over 4826.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2754, pruned_loss=0.08019, over 381352.62 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:58:02,669 INFO [optim.py:369] (2/7) 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 15:58:12,529 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23041.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:58:25,549 INFO [finetune.py:976] (2/7) Epoch 5, batch 150, loss[loss=0.1948, simple_loss=0.2592, pruned_loss=0.06524, over 4870.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2681, pruned_loss=0.0768, over 508545.99 frames. ], batch size: 34, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:58:44,620 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23089.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:58:57,363 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23108.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:58:59,551 INFO [finetune.py:976] (2/7) Epoch 5, batch 200, loss[loss=0.2519, simple_loss=0.2869, pruned_loss=0.1085, over 4930.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2655, pruned_loss=0.07506, over 610682.86 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:59:10,065 INFO [optim.py:369] (2/7) 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 15:59:25,391 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23150.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 15:59:33,030 INFO [finetune.py:976] (2/7) Epoch 5, batch 250, loss[loss=0.2613, simple_loss=0.3015, pruned_loss=0.1106, over 4913.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2659, pruned_loss=0.07392, over 685914.19 frames. ], batch size: 36, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 15:59:38,560 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23181.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:00:00,766 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0345, 1.6631, 2.0953, 2.2389, 2.0570, 1.9418, 2.0859, 2.0364], + device='cuda:2'), covar=tensor([1.4261, 1.8143, 2.2181, 2.6746, 1.7508, 2.3928, 2.2762, 1.8141], + device='cuda:2'), in_proj_covar=tensor([0.0427, 0.0466, 0.0555, 0.0575, 0.0459, 0.0484, 0.0496, 0.0498], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 16:00:03,132 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2116, 1.5560, 1.4735, 1.7458, 1.6336, 1.9613, 1.3857, 3.6326], + device='cuda:2'), covar=tensor([0.0665, 0.0768, 0.0800, 0.1239, 0.0646, 0.0585, 0.0788, 0.0143], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0046, 0.0041, 0.0041, 0.0040, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 16:00:06,066 INFO [finetune.py:976] (2/7) Epoch 5, batch 300, loss[loss=0.2368, simple_loss=0.3008, pruned_loss=0.08636, over 4914.00 frames. ], tot_loss[loss=0.2101, simple_loss=0.2705, pruned_loss=0.07483, over 744958.60 frames. ], batch size: 36, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:00:10,742 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23217.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:00:14,411 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9611, 2.2113, 0.7476, 1.2824, 1.4254, 1.6000, 1.4993, 0.8453], + device='cuda:2'), covar=tensor([0.1344, 0.1210, 0.1639, 0.1333, 0.0989, 0.0960, 0.1355, 0.1400], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0265, 0.0149, 0.0130, 0.0140, 0.0162, 0.0126, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:00:15,563 INFO [optim.py:369] (2/7) 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 16:00:18,579 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23229.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:00:34,811 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-26 16:00:55,701 INFO [finetune.py:976] (2/7) Epoch 5, batch 350, loss[loss=0.233, simple_loss=0.2961, pruned_loss=0.08492, over 4892.00 frames. ], tot_loss[loss=0.2163, simple_loss=0.2761, pruned_loss=0.07824, over 791109.58 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:01:18,170 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23278.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:01:31,714 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23290.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:01:40,418 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4263, 1.4199, 1.5520, 2.1362, 2.3616, 2.0285, 1.8419, 1.8226], + device='cuda:2'), covar=tensor([0.1952, 0.2319, 0.2115, 0.1991, 0.1499, 0.2171, 0.2540, 0.1872], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0331, 0.0348, 0.0305, 0.0344, 0.0345, 0.0310, 0.0351], + device='cuda:2'), out_proj_covar=tensor([6.7006e-05, 7.0894e-05, 7.5373e-05, 6.3716e-05, 7.2842e-05, 7.5057e-05, + 6.7231e-05, 7.5850e-05], device='cuda:2') +2023-04-26 16:01:40,978 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23295.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:01:55,886 INFO [finetune.py:976] (2/7) Epoch 5, batch 400, loss[loss=0.2338, simple_loss=0.3027, pruned_loss=0.08244, over 4814.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2783, pruned_loss=0.07865, over 828656.10 frames. ], batch size: 39, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:02:01,015 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-26 16:02:05,328 INFO [optim.py:369] (2/7) 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 16:02:05,464 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9963, 2.4898, 2.1737, 2.2809, 1.8258, 1.9275, 2.1045, 1.7160], + device='cuda:2'), covar=tensor([0.2170, 0.1305, 0.0925, 0.1384, 0.3332, 0.1380, 0.2056, 0.2971], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0332, 0.0241, 0.0305, 0.0324, 0.0287, 0.0274, 0.0296], + device='cuda:2'), out_proj_covar=tensor([1.2727e-04, 1.3474e-04, 9.8069e-05, 1.2288e-04, 1.3356e-04, 1.1576e-04, + 1.1296e-04, 1.1966e-04], device='cuda:2') +2023-04-26 16:02:05,494 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1415, 1.5166, 1.9698, 2.2735, 1.8753, 1.4914, 1.1266, 1.7201], + device='cuda:2'), covar=tensor([0.4609, 0.5069, 0.2358, 0.3626, 0.4332, 0.3853, 0.6418, 0.3783], + device='cuda:2'), in_proj_covar=tensor([0.0272, 0.0262, 0.0220, 0.0331, 0.0221, 0.0230, 0.0248, 0.0197], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:02:13,057 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4230, 3.5497, 1.0267, 1.8222, 1.9235, 2.4388, 2.0301, 1.0287], + device='cuda:2'), covar=tensor([0.1467, 0.1020, 0.2000, 0.1376, 0.1174, 0.1114, 0.1588, 0.2015], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0264, 0.0148, 0.0130, 0.0140, 0.0162, 0.0127, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:02:18,849 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23343.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:02:29,820 INFO [finetune.py:976] (2/7) Epoch 5, batch 450, loss[loss=0.2741, simple_loss=0.3185, pruned_loss=0.1148, over 4371.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2768, pruned_loss=0.07766, over 857912.79 frames. ], batch size: 65, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:02:49,268 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7889, 2.0079, 1.7888, 2.0734, 1.8531, 2.1620, 1.9440, 1.9445], + device='cuda:2'), covar=tensor([0.8166, 1.4893, 1.2687, 1.1104, 1.2979, 1.8754, 1.5812, 1.3119], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0399, 0.0320, 0.0327, 0.0350, 0.0414, 0.0384, 0.0338], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:03:12,381 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5388, 3.7498, 1.0781, 2.0501, 2.1742, 2.5003, 2.2752, 1.0940], + device='cuda:2'), covar=tensor([0.1348, 0.0837, 0.2010, 0.1290, 0.1014, 0.1114, 0.1403, 0.2167], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0264, 0.0148, 0.0130, 0.0139, 0.0162, 0.0126, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:03:14,687 INFO [finetune.py:976] (2/7) Epoch 5, batch 500, loss[loss=0.2912, simple_loss=0.327, pruned_loss=0.1277, over 4906.00 frames. ], tot_loss[loss=0.2147, simple_loss=0.275, pruned_loss=0.07721, over 879468.87 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:03:29,843 INFO [optim.py:369] (2/7) 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 16:03:47,510 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23450.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:03:54,136 INFO [finetune.py:976] (2/7) Epoch 5, batch 550, loss[loss=0.1985, simple_loss=0.2606, pruned_loss=0.06819, over 4941.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.271, pruned_loss=0.07567, over 894788.53 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 32.0 +2023-04-26 16:03:56,064 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23464.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:07,413 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23481.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:09,295 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6986, 1.8404, 1.0750, 1.4627, 2.1355, 1.6782, 1.5675, 1.5095], + device='cuda:2'), covar=tensor([0.0543, 0.0391, 0.0384, 0.0573, 0.0285, 0.0533, 0.0548, 0.0618], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0025, 0.0031, 0.0022, 0.0030, 0.0030, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 16:04:12,220 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23488.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:19,697 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 600, loss[loss=0.1842, simple_loss=0.2524, pruned_loss=0.05801, over 4907.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2704, pruned_loss=0.07583, over 909595.40 frames. ], batch size: 32, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:04:32,588 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5836, 1.6809, 1.6684, 2.3386, 2.4486, 2.2346, 1.9713, 1.9021], + device='cuda:2'), covar=tensor([0.1910, 0.2156, 0.2565, 0.1634, 0.1523, 0.1946, 0.2809, 0.2190], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0332, 0.0348, 0.0305, 0.0343, 0.0344, 0.0312, 0.0351], + device='cuda:2'), 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:2') +2023-04-26 16:04:36,131 INFO [optim.py:369] (2/7) 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 16:04:39,606 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23529.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:04:53,901 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23549.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 16:05:01,024 INFO [finetune.py:976] (2/7) Epoch 5, batch 650, loss[loss=0.1952, simple_loss=0.2633, pruned_loss=0.06352, over 4801.00 frames. ], tot_loss[loss=0.2143, simple_loss=0.2741, pruned_loss=0.07723, over 919994.44 frames. ], batch size: 45, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:05:01,785 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8283, 2.5658, 1.7233, 1.7365, 1.3322, 1.4169, 1.8881, 1.2655], + device='cuda:2'), covar=tensor([0.2123, 0.1755, 0.2096, 0.2293, 0.3179, 0.2402, 0.1500, 0.2642], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0221, 0.0181, 0.0210, 0.0219, 0.0189, 0.0174, 0.0197], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:05:08,391 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23573.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:05:16,710 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23585.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:05:34,424 INFO [finetune.py:976] (2/7) Epoch 5, batch 700, loss[loss=0.2533, simple_loss=0.3233, pruned_loss=0.09168, over 4804.00 frames. ], tot_loss[loss=0.2177, simple_loss=0.2777, pruned_loss=0.07883, over 927807.99 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:05:42,878 INFO [optim.py:369] (2/7) 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 16:06:19,708 INFO [finetune.py:976] (2/7) Epoch 5, batch 750, loss[loss=0.1983, simple_loss=0.2705, pruned_loss=0.06298, over 4792.00 frames. ], tot_loss[loss=0.2199, simple_loss=0.2799, pruned_loss=0.07997, over 935301.48 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:06:31,011 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6980, 1.4133, 1.8751, 1.8561, 1.4798, 1.2749, 1.5714, 0.9809], + device='cuda:2'), covar=tensor([0.0722, 0.0955, 0.0550, 0.1016, 0.0981, 0.1514, 0.0833, 0.1071], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0078, 0.0076, 0.0071, 0.0082, 0.0098, 0.0085, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 16:07:12,108 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5400, 1.2830, 1.7827, 1.7238, 1.4093, 1.1730, 1.4013, 0.8844], + device='cuda:2'), covar=tensor([0.0840, 0.1127, 0.0595, 0.0963, 0.1054, 0.1694, 0.0917, 0.1245], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0079, 0.0077, 0.0071, 0.0082, 0.0098, 0.0085, 0.0080], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 16:07:26,345 INFO [finetune.py:976] (2/7) Epoch 5, batch 800, loss[loss=0.2067, simple_loss=0.2734, pruned_loss=0.07001, over 4895.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2786, pruned_loss=0.07874, over 941444.08 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:07:34,820 INFO [optim.py:369] (2/7) 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 16:07:51,608 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4066, 1.4294, 1.4258, 1.0534, 1.4411, 1.2030, 1.7742, 1.3857], + device='cuda:2'), covar=tensor([0.3878, 0.1648, 0.5247, 0.2750, 0.1459, 0.2188, 0.1758, 0.5031], + device='cuda:2'), in_proj_covar=tensor([0.0352, 0.0356, 0.0438, 0.0372, 0.0400, 0.0385, 0.0397, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 16:08:00,117 INFO [finetune.py:976] (2/7) Epoch 5, batch 850, loss[loss=0.1743, simple_loss=0.2373, pruned_loss=0.05569, over 4491.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.277, pruned_loss=0.07819, over 944482.40 frames. ], batch size: 20, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:08:02,019 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23764.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:08:39,275 INFO [finetune.py:976] (2/7) Epoch 5, batch 900, loss[loss=0.1917, simple_loss=0.2588, pruned_loss=0.06231, over 4773.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2743, pruned_loss=0.07746, over 947596.76 frames. ], batch size: 28, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:08:40,384 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23812.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:08:54,018 INFO [optim.py:369] (2/7) 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 16:09:15,053 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8103, 3.6670, 2.8125, 4.3672, 3.8155, 3.8069, 1.4654, 3.6932], + device='cuda:2'), covar=tensor([0.1705, 0.1295, 0.3666, 0.1486, 0.3244, 0.2056, 0.6114, 0.2380], + device='cuda:2'), in_proj_covar=tensor([0.0249, 0.0223, 0.0262, 0.0316, 0.0310, 0.0260, 0.0278, 0.0279], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:09:23,844 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23844.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:09:46,929 INFO [finetune.py:976] (2/7) Epoch 5, batch 950, loss[loss=0.2056, simple_loss=0.2567, pruned_loss=0.0772, over 4907.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2721, pruned_loss=0.07694, over 947469.26 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:09:55,189 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 16:09:56,932 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23873.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:10:14,847 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23885.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:10:44,370 INFO [finetune.py:976] (2/7) Epoch 5, batch 1000, loss[loss=0.2098, simple_loss=0.2646, pruned_loss=0.07745, over 4864.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2731, pruned_loss=0.07734, over 950077.93 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:10:52,028 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23921.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:10:54,329 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) Epoch 5, batch 1050, loss[loss=0.2173, simple_loss=0.2726, pruned_loss=0.08101, over 4730.00 frames. ], tot_loss[loss=0.2162, simple_loss=0.276, pruned_loss=0.07823, over 951921.93 frames. ], batch size: 59, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:11:25,503 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5645, 2.2104, 1.5036, 1.3910, 1.1802, 1.2233, 1.6271, 1.1190], + device='cuda:2'), covar=tensor([0.2035, 0.1668, 0.2104, 0.2418, 0.3094, 0.2441, 0.1516, 0.2621], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0221, 0.0180, 0.0210, 0.0218, 0.0189, 0.0173, 0.0197], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:11:46,155 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6119, 3.9084, 0.8795, 1.9357, 2.0578, 2.4711, 2.2486, 1.0264], + device='cuda:2'), covar=tensor([0.1404, 0.0970, 0.2138, 0.1365, 0.1131, 0.1236, 0.1464, 0.2198], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0263, 0.0146, 0.0128, 0.0138, 0.0160, 0.0126, 0.0128], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:12:02,474 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4831, 1.7814, 2.1831, 2.8342, 2.0793, 1.5914, 1.6050, 2.1690], + device='cuda:2'), covar=tensor([0.4148, 0.4879, 0.2174, 0.3830, 0.4588, 0.3808, 0.5669, 0.3798], + device='cuda:2'), in_proj_covar=tensor([0.0273, 0.0262, 0.0221, 0.0331, 0.0222, 0.0230, 0.0248, 0.0198], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:12:21,587 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4748, 1.1896, 1.5375, 1.8125, 1.6938, 1.4357, 1.5030, 1.5175], + device='cuda:2'), covar=tensor([1.0150, 1.4472, 1.4946, 1.6660, 1.1375, 1.6641, 1.6845, 1.4213], + device='cuda:2'), in_proj_covar=tensor([0.0426, 0.0465, 0.0551, 0.0572, 0.0457, 0.0481, 0.0493, 0.0495], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 16:12:31,172 INFO [finetune.py:976] (2/7) Epoch 5, batch 1100, loss[loss=0.2085, simple_loss=0.279, pruned_loss=0.06896, over 4805.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.278, pruned_loss=0.07917, over 952059.79 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:12:45,515 INFO [optim.py:369] (2/7) 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 16:12:53,375 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24037.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:13:04,883 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 16:13:08,245 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 16:13:08,678 INFO [finetune.py:976] (2/7) Epoch 5, batch 1150, loss[loss=0.2233, simple_loss=0.2929, pruned_loss=0.07689, over 4896.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2804, pruned_loss=0.08031, over 953417.42 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:13:28,091 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.5564, 1.5630, 1.6741, 1.1433, 1.7597, 1.2680, 2.2721, 1.4715], + device='cuda:2'), covar=tensor([0.4176, 0.1995, 0.5484, 0.3638, 0.1986, 0.2787, 0.1446, 0.4777], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0353, 0.0435, 0.0368, 0.0397, 0.0382, 0.0393, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 16:13:33,482 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24098.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:13:37,672 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-04-26 16:13:42,162 INFO [finetune.py:976] (2/7) Epoch 5, batch 1200, loss[loss=0.1767, simple_loss=0.2411, pruned_loss=0.05617, over 4769.00 frames. ], tot_loss[loss=0.219, simple_loss=0.2788, pruned_loss=0.0796, over 952859.19 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:13:52,170 INFO [optim.py:369] (2/7) 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 16:14:01,108 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5848, 1.3136, 1.2431, 1.4493, 1.8221, 1.5534, 1.3454, 1.2106], + device='cuda:2'), covar=tensor([0.1367, 0.1293, 0.1713, 0.1174, 0.0655, 0.1200, 0.1901, 0.1681], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0336, 0.0351, 0.0309, 0.0346, 0.0347, 0.0312, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.7693e-05, 7.2048e-05, 7.6122e-05, 6.4527e-05, 7.3433e-05, 7.5509e-05, + 6.7630e-05, 7.6477e-05], device='cuda:2') +2023-04-26 16:14:04,701 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24144.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:14:15,796 INFO [finetune.py:976] (2/7) Epoch 5, batch 1250, loss[loss=0.1906, simple_loss=0.2407, pruned_loss=0.07018, over 4817.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.2762, pruned_loss=0.07862, over 952247.09 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:14:27,770 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-26 16:14:34,639 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9117, 4.1184, 0.7878, 2.2366, 2.2026, 2.6154, 2.4341, 0.9825], + device='cuda:2'), covar=tensor([0.1247, 0.0982, 0.2235, 0.1201, 0.0989, 0.1183, 0.1313, 0.2087], + device='cuda:2'), in_proj_covar=tensor([0.0122, 0.0263, 0.0145, 0.0128, 0.0138, 0.0160, 0.0125, 0.0128], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:14:42,784 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24192.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:14:51,495 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-04-26 16:15:00,605 INFO [finetune.py:976] (2/7) Epoch 5, batch 1300, loss[loss=0.173, simple_loss=0.2293, pruned_loss=0.05832, over 4219.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2717, pruned_loss=0.07662, over 952766.33 frames. ], batch size: 65, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:15:09,762 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 16:15:10,679 INFO [optim.py:369] (2/7) 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 16:15:41,946 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5473, 1.0994, 1.2544, 1.1935, 1.7076, 1.3721, 1.0698, 1.2178], + device='cuda:2'), covar=tensor([0.1896, 0.1516, 0.2370, 0.1445, 0.1002, 0.1717, 0.2535, 0.2332], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0336, 0.0352, 0.0310, 0.0348, 0.0348, 0.0313, 0.0354], + device='cuda:2'), 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:2') +2023-04-26 16:15:49,984 INFO [finetune.py:976] (2/7) Epoch 5, batch 1350, loss[loss=0.1576, simple_loss=0.2154, pruned_loss=0.04987, over 4126.00 frames. ], tot_loss[loss=0.2116, simple_loss=0.2711, pruned_loss=0.07607, over 953592.93 frames. ], batch size: 18, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:16:00,988 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.14 vs. limit=5.0 +2023-04-26 16:16:29,536 INFO [finetune.py:976] (2/7) Epoch 5, batch 1400, loss[loss=0.2466, simple_loss=0.3189, pruned_loss=0.0872, over 4898.00 frames. ], tot_loss[loss=0.2164, simple_loss=0.2762, pruned_loss=0.07828, over 955414.70 frames. ], batch size: 43, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:16:38,503 INFO [optim.py:369] (2/7) 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 16:17:19,152 INFO [finetune.py:976] (2/7) Epoch 5, batch 1450, loss[loss=0.2211, simple_loss=0.2948, pruned_loss=0.07373, over 4789.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2786, pruned_loss=0.07871, over 956077.03 frames. ], batch size: 51, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:18:00,685 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24393.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:18:16,237 INFO [finetune.py:976] (2/7) Epoch 5, batch 1500, loss[loss=0.206, simple_loss=0.2732, pruned_loss=0.06939, over 4874.00 frames. ], tot_loss[loss=0.2201, simple_loss=0.2806, pruned_loss=0.07982, over 955878.12 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:18:25,730 INFO [optim.py:369] (2/7) 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 16:18:46,011 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9683, 2.6858, 1.9049, 1.8055, 1.3794, 1.4328, 2.1077, 1.4056], + device='cuda:2'), covar=tensor([0.1861, 0.1772, 0.1893, 0.2299, 0.2839, 0.2273, 0.1280, 0.2276], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0222, 0.0180, 0.0210, 0.0218, 0.0190, 0.0173, 0.0197], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:18:49,533 INFO [finetune.py:976] (2/7) Epoch 5, batch 1550, loss[loss=0.2524, simple_loss=0.3018, pruned_loss=0.1015, over 4810.00 frames. ], tot_loss[loss=0.2187, simple_loss=0.2794, pruned_loss=0.07895, over 955376.94 frames. ], batch size: 38, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:18:50,950 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-04-26 16:18:57,925 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3780, 1.2021, 4.4099, 4.1002, 3.8696, 4.1832, 4.0718, 3.8061], + device='cuda:2'), covar=tensor([0.7367, 0.6184, 0.1071, 0.1737, 0.1148, 0.2177, 0.1482, 0.1786], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0308, 0.0425, 0.0431, 0.0366, 0.0418, 0.0326, 0.0386], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 16:19:22,930 INFO [finetune.py:976] (2/7) Epoch 5, batch 1600, loss[loss=0.2404, simple_loss=0.2835, pruned_loss=0.09866, over 4295.00 frames. ], tot_loss[loss=0.2164, simple_loss=0.2767, pruned_loss=0.07803, over 954611.33 frames. ], batch size: 66, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:19:32,014 INFO [optim.py:369] (2/7) 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 16:19:48,172 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24547.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:19:56,585 INFO [finetune.py:976] (2/7) Epoch 5, batch 1650, loss[loss=0.2003, simple_loss=0.2653, pruned_loss=0.06762, over 4946.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2737, pruned_loss=0.0769, over 955946.23 frames. ], batch size: 38, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:20:01,611 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24569.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:20:15,599 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 16:20:28,443 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 1700, loss[loss=0.1753, simple_loss=0.2448, pruned_loss=0.05294, over 4824.00 frames. ], tot_loss[loss=0.2118, simple_loss=0.2714, pruned_loss=0.07607, over 956380.14 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:20:38,586 INFO [optim.py:369] (2/7) 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 16:20:42,265 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24630.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:21:03,215 INFO [finetune.py:976] (2/7) Epoch 5, batch 1750, loss[loss=0.2555, simple_loss=0.301, pruned_loss=0.1051, over 4889.00 frames. ], tot_loss[loss=0.2144, simple_loss=0.2737, pruned_loss=0.07749, over 953849.50 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:21:29,303 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1386, 1.4983, 1.2832, 1.6421, 1.5423, 1.8920, 1.3338, 3.3310], + device='cuda:2'), covar=tensor([0.0722, 0.0748, 0.0802, 0.1195, 0.0647, 0.0633, 0.0759, 0.0164], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0041, 0.0046, 0.0041, 0.0041, 0.0040, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 16:21:35,436 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24693.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:21:48,677 INFO [finetune.py:976] (2/7) Epoch 5, batch 1800, loss[loss=0.2058, simple_loss=0.2725, pruned_loss=0.06957, over 4819.00 frames. ], tot_loss[loss=0.2174, simple_loss=0.2774, pruned_loss=0.07871, over 952461.42 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:21:57,282 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24741.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:22:12,875 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7459, 1.5390, 1.8560, 1.9955, 1.5925, 1.2580, 1.7617, 1.1993], + device='cuda:2'), covar=tensor([0.0864, 0.0903, 0.0630, 0.0920, 0.1084, 0.1396, 0.0854, 0.1173], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0078, 0.0076, 0.0070, 0.0081, 0.0097, 0.0084, 0.0079], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 16:22:27,731 INFO [finetune.py:976] (2/7) Epoch 5, batch 1850, loss[loss=0.2449, simple_loss=0.2867, pruned_loss=0.1016, over 4781.00 frames. ], tot_loss[loss=0.2193, simple_loss=0.2796, pruned_loss=0.0795, over 953491.32 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:22:30,270 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:23:29,633 INFO [finetune.py:976] (2/7) Epoch 5, batch 1900, loss[loss=0.2268, simple_loss=0.2899, pruned_loss=0.08179, over 4872.00 frames. ], tot_loss[loss=0.2199, simple_loss=0.2806, pruned_loss=0.07961, over 954754.29 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:23:44,192 INFO [optim.py:369] (2/7) 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 16:23:50,631 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24826.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:24:15,477 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7029, 2.3170, 1.6057, 1.4278, 1.2476, 1.3195, 1.6645, 1.2206], + device='cuda:2'), covar=tensor([0.1871, 0.1518, 0.1921, 0.2284, 0.2804, 0.2151, 0.1384, 0.2350], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0224, 0.0181, 0.0211, 0.0219, 0.0191, 0.0174, 0.0198], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:24:16,072 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:24:29,571 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-26 16:24:31,094 INFO [finetune.py:976] (2/7) Epoch 5, batch 1950, loss[loss=0.1862, simple_loss=0.2665, pruned_loss=0.05295, over 4909.00 frames. ], tot_loss[loss=0.2183, simple_loss=0.2789, pruned_loss=0.07885, over 955877.33 frames. ], batch size: 43, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:24:58,637 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24903.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:25:03,340 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24909.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:25:04,922 INFO [finetune.py:976] (2/7) Epoch 5, batch 2000, loss[loss=0.2515, simple_loss=0.3049, pruned_loss=0.09906, over 4834.00 frames. ], tot_loss[loss=0.2166, simple_loss=0.2766, pruned_loss=0.07835, over 956249.92 frames. ], batch size: 30, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:25:13,836 INFO [optim.py:369] (2/7) 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 16:25:13,936 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24925.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:25:18,846 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3038, 1.5819, 1.3903, 1.5205, 1.4271, 1.2775, 1.3851, 1.1447], + device='cuda:2'), covar=tensor([0.1950, 0.1317, 0.1120, 0.1495, 0.3656, 0.1541, 0.1925, 0.2670], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0330, 0.0241, 0.0303, 0.0322, 0.0284, 0.0273, 0.0297], + device='cuda:2'), out_proj_covar=tensor([1.2687e-04, 1.3419e-04, 9.8014e-05, 1.2200e-04, 1.3281e-04, 1.1474e-04, + 1.1226e-04, 1.1999e-04], device='cuda:2') +2023-04-26 16:25:21,324 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8166, 1.0616, 1.2811, 1.4579, 1.4487, 1.6151, 1.3431, 1.3608], + device='cuda:2'), covar=tensor([0.6594, 0.9354, 0.8133, 0.7634, 0.9441, 1.3672, 0.9899, 0.9030], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0398, 0.0317, 0.0327, 0.0349, 0.0413, 0.0381, 0.0336], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:25:34,925 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4883, 2.4263, 1.9453, 2.2240, 2.5062, 2.0535, 3.0889, 1.7582], + device='cuda:2'), covar=tensor([0.3989, 0.1855, 0.4441, 0.3265, 0.1706, 0.2495, 0.2170, 0.4198], + device='cuda:2'), in_proj_covar=tensor([0.0349, 0.0355, 0.0436, 0.0367, 0.0396, 0.0383, 0.0395, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 16:25:37,718 INFO [finetune.py:976] (2/7) Epoch 5, batch 2050, loss[loss=0.1987, simple_loss=0.2629, pruned_loss=0.0672, over 4235.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2724, pruned_loss=0.07647, over 954368.26 frames. ], batch size: 65, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:25:59,900 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24995.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:26:01,121 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6536, 2.2092, 1.6240, 1.5076, 1.3059, 1.3667, 1.7253, 1.2464], + device='cuda:2'), covar=tensor([0.1761, 0.1547, 0.1743, 0.2087, 0.2794, 0.2105, 0.1259, 0.2220], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0223, 0.0180, 0.0210, 0.0219, 0.0190, 0.0173, 0.0197], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:26:10,615 INFO [finetune.py:976] (2/7) Epoch 5, batch 2100, loss[loss=0.1975, simple_loss=0.2639, pruned_loss=0.0656, over 4916.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2714, pruned_loss=0.07637, over 953665.56 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:26:10,764 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8124, 1.3619, 1.6516, 1.4649, 1.5799, 1.3307, 0.6095, 1.2482], + device='cuda:2'), covar=tensor([0.4484, 0.4943, 0.2227, 0.3668, 0.4132, 0.3759, 0.5983, 0.3568], + device='cuda:2'), in_proj_covar=tensor([0.0275, 0.0261, 0.0221, 0.0331, 0.0221, 0.0230, 0.0248, 0.0197], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:26:21,075 INFO [optim.py:369] (2/7) 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 16:26:29,779 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.33 vs. limit=5.0 +2023-04-26 16:26:40,927 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25056.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 16:26:43,815 INFO [finetune.py:976] (2/7) Epoch 5, batch 2150, loss[loss=0.1666, simple_loss=0.2253, pruned_loss=0.05394, over 4313.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2731, pruned_loss=0.07641, over 954082.72 frames. ], batch size: 18, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:27:17,133 INFO [finetune.py:976] (2/7) Epoch 5, batch 2200, loss[loss=0.2014, simple_loss=0.2518, pruned_loss=0.07556, over 4683.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2761, pruned_loss=0.07726, over 954952.68 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:27:24,270 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25121.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:27:27,080 INFO [optim.py:369] (2/7) 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 16:27:33,265 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-26 16:27:42,240 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3280, 2.9895, 2.4504, 2.5371, 2.1255, 2.3923, 2.6371, 2.0922], + device='cuda:2'), covar=tensor([0.2555, 0.1591, 0.0916, 0.1742, 0.3456, 0.1452, 0.2429, 0.2996], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0329, 0.0239, 0.0302, 0.0321, 0.0283, 0.0272, 0.0296], + device='cuda:2'), out_proj_covar=tensor([1.2660e-04, 1.3344e-04, 9.7449e-05, 1.2164e-04, 1.3223e-04, 1.1439e-04, + 1.1189e-04, 1.1964e-04], device='cuda:2') +2023-04-26 16:27:50,267 INFO [finetune.py:976] (2/7) Epoch 5, batch 2250, loss[loss=0.245, simple_loss=0.2964, pruned_loss=0.0968, over 4901.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.2776, pruned_loss=0.07795, over 954522.91 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:27:51,771 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.12 vs. limit=5.0 +2023-04-26 16:28:26,329 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25203.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:28:26,903 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25204.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:28:32,144 INFO [finetune.py:976] (2/7) Epoch 5, batch 2300, loss[loss=0.2332, simple_loss=0.286, pruned_loss=0.09024, over 4863.00 frames. ], tot_loss[loss=0.2175, simple_loss=0.2786, pruned_loss=0.07826, over 954392.33 frames. ], batch size: 34, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:28:52,192 INFO [optim.py:369] (2/7) 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 16:28:52,295 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25225.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:29:26,417 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25251.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:29:38,265 INFO [finetune.py:976] (2/7) Epoch 5, batch 2350, loss[loss=0.2388, simple_loss=0.2805, pruned_loss=0.09861, over 4866.00 frames. ], tot_loss[loss=0.2156, simple_loss=0.2759, pruned_loss=0.07762, over 951627.01 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:29:57,348 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25273.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:30:39,280 INFO [finetune.py:976] (2/7) Epoch 5, batch 2400, loss[loss=0.1849, simple_loss=0.2526, pruned_loss=0.05862, over 4827.00 frames. ], tot_loss[loss=0.2136, simple_loss=0.2733, pruned_loss=0.07698, over 952708.59 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:30:48,373 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25351.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:31:18,440 INFO [finetune.py:976] (2/7) Epoch 5, batch 2450, loss[loss=0.1921, simple_loss=0.2551, pruned_loss=0.06456, over 4919.00 frames. ], tot_loss[loss=0.2131, simple_loss=0.272, pruned_loss=0.07708, over 953538.57 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:31:22,249 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 16:32:08,174 INFO [finetune.py:976] (2/7) Epoch 5, batch 2500, loss[loss=0.1963, simple_loss=0.2492, pruned_loss=0.07176, over 4398.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2727, pruned_loss=0.07753, over 953394.39 frames. ], batch size: 19, lr: 3.94e-03, grad_scale: 64.0 +2023-04-26 16:32:15,258 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25421.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:32:17,550 INFO [optim.py:369] (2/7) 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 16:32:19,523 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 2550, loss[loss=0.2178, simple_loss=0.2823, pruned_loss=0.0766, over 4874.00 frames. ], tot_loss[loss=0.2183, simple_loss=0.2777, pruned_loss=0.0794, over 955260.72 frames. ], batch size: 34, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:32:47,380 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25469.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:33:01,463 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25489.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 16:33:03,803 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4651, 1.5184, 1.6985, 1.8190, 1.6897, 1.8900, 1.8179, 1.8276], + device='cuda:2'), covar=tensor([0.6724, 1.1233, 0.9528, 0.8450, 1.0053, 1.5124, 1.0875, 1.0033], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0398, 0.0318, 0.0327, 0.0349, 0.0414, 0.0381, 0.0337], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:33:11,419 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25504.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:33:15,646 INFO [finetune.py:976] (2/7) Epoch 5, batch 2600, loss[loss=0.2227, simple_loss=0.2823, pruned_loss=0.08155, over 4820.00 frames. ], tot_loss[loss=0.2208, simple_loss=0.2805, pruned_loss=0.08048, over 954242.86 frames. ], batch size: 47, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:33:24,262 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0319, 1.4145, 1.8911, 2.3349, 1.7527, 1.4044, 1.0685, 1.6201], + device='cuda:2'), covar=tensor([0.4365, 0.4904, 0.2247, 0.3467, 0.4136, 0.3779, 0.6110, 0.3670], + device='cuda:2'), in_proj_covar=tensor([0.0274, 0.0260, 0.0220, 0.0329, 0.0220, 0.0229, 0.0245, 0.0196], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:33:25,300 INFO [optim.py:369] (2/7) 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 16:33:43,927 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25552.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:33:49,361 INFO [finetune.py:976] (2/7) Epoch 5, batch 2650, loss[loss=0.2012, simple_loss=0.2709, pruned_loss=0.06576, over 4818.00 frames. ], tot_loss[loss=0.2206, simple_loss=0.2805, pruned_loss=0.08035, over 953449.40 frames. ], batch size: 38, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:34:28,770 INFO [finetune.py:976] (2/7) Epoch 5, batch 2700, loss[loss=0.1792, simple_loss=0.2479, pruned_loss=0.05525, over 4913.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2798, pruned_loss=0.07991, over 955122.22 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:34:48,614 INFO [optim.py:369] (2/7) 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 16:35:15,640 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3749, 1.6067, 1.4902, 1.6351, 1.5564, 1.7638, 1.6404, 1.5896], + device='cuda:2'), covar=tensor([0.7125, 1.1632, 0.9963, 0.8699, 1.0304, 1.6018, 1.2597, 1.1599], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0398, 0.0318, 0.0328, 0.0350, 0.0415, 0.0381, 0.0337], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:35:20,267 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25651.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:35:37,125 INFO [finetune.py:976] (2/7) Epoch 5, batch 2750, loss[loss=0.1914, simple_loss=0.2655, pruned_loss=0.05871, over 4795.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2765, pruned_loss=0.07828, over 953992.18 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 32.0 +2023-04-26 16:35:47,648 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3675, 3.2522, 0.9989, 1.8830, 1.8421, 2.3487, 1.8391, 1.1178], + device='cuda:2'), covar=tensor([0.1391, 0.0854, 0.1960, 0.1246, 0.1069, 0.1049, 0.1493, 0.2002], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0265, 0.0147, 0.0129, 0.0139, 0.0162, 0.0126, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:35:49,080 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-26 16:36:24,083 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25699.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:36:24,207 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 16:36:29,075 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-26 16:36:32,332 INFO [finetune.py:976] (2/7) Epoch 5, batch 2800, loss[loss=0.188, simple_loss=0.2645, pruned_loss=0.05577, over 4824.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2722, pruned_loss=0.0767, over 954428.47 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:36:47,384 INFO [optim.py:369] (2/7) 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 16:37:23,536 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0735, 0.5576, 0.8895, 0.6991, 1.2112, 0.9304, 0.7543, 0.9353], + device='cuda:2'), covar=tensor([0.2212, 0.1971, 0.1958, 0.1710, 0.0955, 0.1567, 0.2177, 0.1934], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0332, 0.0350, 0.0307, 0.0345, 0.0342, 0.0308, 0.0351], + device='cuda:2'), out_proj_covar=tensor([6.7397e-05, 7.1130e-05, 7.5751e-05, 6.4200e-05, 7.3292e-05, 7.4438e-05, + 6.6814e-05, 7.5623e-05], device='cuda:2') +2023-04-26 16:37:32,910 INFO [finetune.py:976] (2/7) Epoch 5, batch 2850, loss[loss=0.2758, simple_loss=0.3034, pruned_loss=0.124, over 4204.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2703, pruned_loss=0.07598, over 954985.91 frames. ], batch size: 18, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:37:46,998 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25784.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:37:52,802 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3066, 1.4898, 1.3036, 1.4798, 1.3366, 1.1855, 1.3207, 1.1437], + device='cuda:2'), covar=tensor([0.1972, 0.1534, 0.1211, 0.1430, 0.3843, 0.1619, 0.2043, 0.2556], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0326, 0.0237, 0.0299, 0.0319, 0.0283, 0.0270, 0.0294], + device='cuda:2'), out_proj_covar=tensor([1.2558e-04, 1.3256e-04, 9.6480e-05, 1.2029e-04, 1.3142e-04, 1.1435e-04, + 1.1100e-04, 1.1877e-04], device='cuda:2') +2023-04-26 16:37:59,987 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9138, 1.3321, 1.7578, 1.9393, 1.6573, 1.3286, 0.9097, 1.4550], + device='cuda:2'), covar=tensor([0.4107, 0.4667, 0.2219, 0.3315, 0.4060, 0.3586, 0.5811, 0.3510], + device='cuda:2'), in_proj_covar=tensor([0.0273, 0.0259, 0.0219, 0.0329, 0.0219, 0.0228, 0.0244, 0.0195], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:38:06,695 INFO [finetune.py:976] (2/7) Epoch 5, batch 2900, loss[loss=0.1988, simple_loss=0.2739, pruned_loss=0.06184, over 4841.00 frames. ], tot_loss[loss=0.2156, simple_loss=0.275, pruned_loss=0.07809, over 956509.17 frames. ], batch size: 33, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:38:09,842 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5225, 1.7634, 0.7967, 1.2049, 1.7555, 1.3743, 1.3215, 1.3442], + device='cuda:2'), covar=tensor([0.0565, 0.0393, 0.0447, 0.0617, 0.0331, 0.0581, 0.0557, 0.0634], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 16:38:15,790 INFO [optim.py:369] (2/7) 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 16:38:37,779 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25858.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:38:39,949 INFO [finetune.py:976] (2/7) Epoch 5, batch 2950, loss[loss=0.258, simple_loss=0.3098, pruned_loss=0.1031, over 4827.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2783, pruned_loss=0.07866, over 956856.91 frames. ], batch size: 47, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:39:12,517 INFO [finetune.py:976] (2/7) Epoch 5, batch 3000, loss[loss=0.2351, simple_loss=0.2862, pruned_loss=0.09195, over 4742.00 frames. ], tot_loss[loss=0.2194, simple_loss=0.2798, pruned_loss=0.07952, over 957920.66 frames. ], batch size: 54, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:39:12,517 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 16:39:15,468 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6145, 1.6214, 3.7010, 3.4523, 3.3928, 3.4610, 3.6180, 3.3200], + device='cuda:2'), covar=tensor([0.6132, 0.4439, 0.1186, 0.1755, 0.1036, 0.1441, 0.0848, 0.1389], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0307, 0.0417, 0.0425, 0.0359, 0.0411, 0.0324, 0.0379], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 16:39:29,055 INFO [finetune.py:1010] (2/7) Epoch 5, validation: loss=0.1595, simple_loss=0.233, pruned_loss=0.04303, over 2265189.00 frames. +2023-04-26 16:39:29,056 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 16:39:41,242 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25919.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 16:39:51,734 INFO [optim.py:369] (2/7) 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 16:39:58,997 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4678, 1.4021, 0.4650, 1.1619, 1.4935, 1.3289, 1.2735, 1.2114], + device='cuda:2'), covar=tensor([0.0596, 0.0422, 0.0506, 0.0647, 0.0337, 0.0607, 0.0582, 0.0680], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0024, 0.0031, 0.0021, 0.0030, 0.0030, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 16:40:35,904 INFO [finetune.py:976] (2/7) Epoch 5, batch 3050, loss[loss=0.1807, simple_loss=0.247, pruned_loss=0.05724, over 4883.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2791, pruned_loss=0.07822, over 958352.47 frames. ], batch size: 43, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:40:45,142 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3813, 1.5808, 1.3108, 0.9113, 1.1301, 1.1094, 1.2640, 1.0629], + device='cuda:2'), covar=tensor([0.2100, 0.1640, 0.1906, 0.2421, 0.2884, 0.2359, 0.1416, 0.2446], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0224, 0.0180, 0.0211, 0.0217, 0.0190, 0.0173, 0.0198], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:41:21,650 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-26 16:41:32,849 INFO [finetune.py:976] (2/7) Epoch 5, batch 3100, loss[loss=0.1847, simple_loss=0.2503, pruned_loss=0.05956, over 4157.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2758, pruned_loss=0.07661, over 955780.82 frames. ], batch size: 17, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:41:43,947 INFO [optim.py:369] (2/7) 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 16:42:06,249 INFO [finetune.py:976] (2/7) Epoch 5, batch 3150, loss[loss=0.2358, simple_loss=0.2873, pruned_loss=0.09214, over 4258.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2728, pruned_loss=0.07621, over 955927.23 frames. ], batch size: 65, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:42:27,117 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26084.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:42:46,467 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26102.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:42:54,362 INFO [finetune.py:976] (2/7) Epoch 5, batch 3200, loss[loss=0.2197, simple_loss=0.2847, pruned_loss=0.07737, over 4814.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2687, pruned_loss=0.07427, over 955662.06 frames. ], batch size: 41, lr: 3.93e-03, grad_scale: 32.0 +2023-04-26 16:43:09,264 INFO [optim.py:369] (2/7) 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 16:43:14,417 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26132.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:43:15,722 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-26 16:43:24,488 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-04-26 16:43:31,948 INFO [finetune.py:976] (2/7) Epoch 5, batch 3250, loss[loss=0.1995, simple_loss=0.2543, pruned_loss=0.07232, over 4776.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2695, pruned_loss=0.0751, over 955722.79 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:43:33,285 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26163.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:43:37,346 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26168.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:44:03,112 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2740, 4.1497, 3.0131, 4.9243, 4.2590, 4.3156, 1.4944, 4.2522], + device='cuda:2'), covar=tensor([0.1363, 0.1027, 0.3353, 0.0932, 0.2916, 0.1412, 0.5905, 0.1899], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0223, 0.0256, 0.0312, 0.0307, 0.0257, 0.0276, 0.0279], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:44:05,488 INFO [finetune.py:976] (2/7) Epoch 5, batch 3300, loss[loss=0.1498, simple_loss=0.2153, pruned_loss=0.04213, over 4388.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2732, pruned_loss=0.07637, over 955960.47 frames. ], batch size: 19, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:44:07,397 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26214.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:44:16,103 INFO [optim.py:369] (2/7) 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 16:44:17,429 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26229.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:44:17,568 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 16:44:27,880 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 16:44:32,034 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 3350, loss[loss=0.2207, simple_loss=0.2812, pruned_loss=0.08012, over 4764.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2757, pruned_loss=0.07693, over 953966.84 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:45:28,586 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-26 16:45:50,831 INFO [finetune.py:976] (2/7) Epoch 5, batch 3400, loss[loss=0.1845, simple_loss=0.2485, pruned_loss=0.06027, over 4768.00 frames. ], tot_loss[loss=0.2162, simple_loss=0.277, pruned_loss=0.07769, over 955424.60 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:45:50,979 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 16:46:17,083 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4883, 1.3113, 1.7495, 1.6773, 1.3404, 1.1498, 1.5191, 1.1325], + device='cuda:2'), covar=tensor([0.0768, 0.0809, 0.0544, 0.0882, 0.0978, 0.1380, 0.0763, 0.0889], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0078, 0.0076, 0.0070, 0.0081, 0.0097, 0.0083, 0.0079], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:46:56,944 INFO [finetune.py:976] (2/7) Epoch 5, batch 3450, loss[loss=0.1664, simple_loss=0.2356, pruned_loss=0.04863, over 4751.00 frames. ], tot_loss[loss=0.215, simple_loss=0.2765, pruned_loss=0.07673, over 955616.74 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:48:03,178 INFO [finetune.py:976] (2/7) Epoch 5, batch 3500, loss[loss=0.1462, simple_loss=0.2114, pruned_loss=0.04057, over 4761.00 frames. ], tot_loss[loss=0.212, simple_loss=0.2732, pruned_loss=0.0754, over 956325.83 frames. ], batch size: 26, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:48:05,103 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 16:49:08,584 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:49:10,352 INFO [finetune.py:976] (2/7) Epoch 5, batch 3550, loss[loss=0.2192, simple_loss=0.2606, pruned_loss=0.0889, over 4826.00 frames. ], tot_loss[loss=0.2105, simple_loss=0.2708, pruned_loss=0.07505, over 956776.67 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:49:24,630 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26475.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:49:25,792 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.5044, 4.3980, 3.0573, 5.2184, 4.4856, 4.5211, 2.1348, 4.4592], + device='cuda:2'), covar=tensor([0.1448, 0.1002, 0.3164, 0.0896, 0.3749, 0.1649, 0.5274, 0.2159], + device='cuda:2'), in_proj_covar=tensor([0.0249, 0.0223, 0.0257, 0.0313, 0.0308, 0.0258, 0.0277, 0.0280], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:49:32,979 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-26 16:49:49,843 INFO [finetune.py:976] (2/7) Epoch 5, batch 3600, loss[loss=0.2165, simple_loss=0.2766, pruned_loss=0.07819, over 4825.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2676, pruned_loss=0.07395, over 957169.23 frames. ], batch size: 39, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:49:51,760 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26514.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:49:57,837 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26524.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:49:59,591 INFO [optim.py:369] (2/7) 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 16:50:11,701 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-04-26 16:50:23,342 INFO [finetune.py:976] (2/7) Epoch 5, batch 3650, loss[loss=0.2363, simple_loss=0.3053, pruned_loss=0.0836, over 4910.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2707, pruned_loss=0.07586, over 955059.41 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:50:24,024 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26562.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 16:50:53,160 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26606.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:50:57,092 INFO [finetune.py:976] (2/7) Epoch 5, batch 3700, loss[loss=0.197, simple_loss=0.2666, pruned_loss=0.06367, over 4753.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2753, pruned_loss=0.07717, over 955914.11 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:51:06,780 INFO [optim.py:369] (2/7) 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 16:51:08,251 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 16:51:15,331 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5441, 1.3616, 1.8180, 1.7893, 1.3908, 1.1888, 1.5953, 1.0970], + device='cuda:2'), covar=tensor([0.0812, 0.1119, 0.0614, 0.1019, 0.1202, 0.1741, 0.0880, 0.1159], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0078, 0.0076, 0.0070, 0.0081, 0.0098, 0.0084, 0.0079], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 16:51:16,972 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6333, 1.3624, 1.8824, 1.7961, 1.4554, 1.2133, 1.6194, 0.9902], + device='cuda:2'), covar=tensor([0.0789, 0.1136, 0.0560, 0.1017, 0.1160, 0.1630, 0.0882, 0.1184], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0078, 0.0076, 0.0070, 0.0081, 0.0098, 0.0084, 0.0079], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:2') +2023-04-26 16:51:20,755 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-26 16:51:29,840 INFO [finetune.py:976] (2/7) Epoch 5, batch 3750, loss[loss=0.1932, simple_loss=0.2581, pruned_loss=0.06412, over 4864.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.278, pruned_loss=0.07813, over 955033.52 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:52:20,490 INFO [finetune.py:976] (2/7) Epoch 5, batch 3800, loss[loss=0.2212, simple_loss=0.2859, pruned_loss=0.07822, over 4891.00 frames. ], tot_loss[loss=0.2185, simple_loss=0.2794, pruned_loss=0.07878, over 955776.42 frames. ], batch size: 37, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:52:31,675 INFO [optim.py:369] (2/7) 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 16:52:36,577 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3740, 1.6095, 1.5543, 2.0647, 1.7286, 2.2321, 1.4718, 4.2731], + device='cuda:2'), covar=tensor([0.0625, 0.0775, 0.0819, 0.1166, 0.0681, 0.0553, 0.0759, 0.0114], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0041, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 16:52:46,803 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 16:52:52,132 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 3850, loss[loss=0.1933, simple_loss=0.2598, pruned_loss=0.06341, over 4812.00 frames. ], tot_loss[loss=0.2162, simple_loss=0.2773, pruned_loss=0.07761, over 955465.33 frames. ], batch size: 39, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:52:56,823 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 16:53:00,826 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26770.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:53:32,238 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5621, 1.6500, 1.7896, 2.3904, 2.5015, 2.1674, 2.0119, 1.8848], + device='cuda:2'), covar=tensor([0.1931, 0.2252, 0.2435, 0.1929, 0.1546, 0.2283, 0.2815, 0.2204], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0334, 0.0352, 0.0309, 0.0346, 0.0343, 0.0310, 0.0352], + device='cuda:2'), out_proj_covar=tensor([6.7440e-05, 7.1566e-05, 7.6128e-05, 6.4667e-05, 7.3371e-05, 7.4405e-05, + 6.7308e-05, 7.5890e-05], device='cuda:2') +2023-04-26 16:53:40,974 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 3900, loss[loss=0.2112, simple_loss=0.2597, pruned_loss=0.0814, over 4830.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.275, pruned_loss=0.07709, over 956919.38 frames. ], batch size: 33, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:54:03,720 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26824.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:54:05,468 INFO [optim.py:369] (2/7) 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 16:54:06,198 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2877, 3.0543, 1.0749, 1.6402, 1.8496, 2.2460, 1.8221, 1.0192], + device='cuda:2'), covar=tensor([0.1472, 0.0976, 0.1851, 0.1429, 0.1127, 0.1052, 0.1459, 0.1843], + device='cuda:2'), in_proj_covar=tensor([0.0120, 0.0257, 0.0144, 0.0127, 0.0136, 0.0158, 0.0123, 0.0126], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 16:54:47,896 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-26 16:54:50,052 INFO [finetune.py:976] (2/7) Epoch 5, batch 3950, loss[loss=0.1955, simple_loss=0.2622, pruned_loss=0.06442, over 4774.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.271, pruned_loss=0.07504, over 957125.07 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:55:08,484 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26872.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:55:47,039 INFO [zipformer.py:1188] (2/7) 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:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26906.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:55:51,749 INFO [finetune.py:976] (2/7) Epoch 5, batch 4000, loss[loss=0.1875, simple_loss=0.2528, pruned_loss=0.06114, over 4868.00 frames. ], tot_loss[loss=0.2089, simple_loss=0.2689, pruned_loss=0.07444, over 955969.06 frames. ], batch size: 34, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:56:02,922 INFO [optim.py:369] (2/7) 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 16:56:03,702 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-26 16:56:17,493 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26954.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:56:25,068 INFO [finetune.py:976] (2/7) Epoch 5, batch 4050, loss[loss=0.2276, simple_loss=0.2979, pruned_loss=0.07868, over 4747.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.274, pruned_loss=0.07721, over 954589.97 frames. ], batch size: 54, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:56:28,126 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26965.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 16:56:58,980 INFO [finetune.py:976] (2/7) Epoch 5, batch 4100, loss[loss=0.1955, simple_loss=0.2381, pruned_loss=0.07643, over 4109.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.277, pruned_loss=0.078, over 955134.30 frames. ], batch size: 17, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:56:59,104 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27011.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:57:20,556 INFO [optim.py:369] (2/7) 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 16:58:04,498 INFO [finetune.py:976] (2/7) Epoch 5, batch 4150, loss[loss=0.2414, simple_loss=0.3042, pruned_loss=0.0893, over 4847.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.278, pruned_loss=0.07805, over 954884.03 frames. ], batch size: 44, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:58:16,079 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27070.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:58:23,883 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0579, 1.2383, 1.4382, 1.6105, 1.5612, 1.6805, 1.5349, 1.5546], + device='cuda:2'), covar=tensor([0.6243, 0.9190, 0.8158, 0.7340, 0.8686, 1.2909, 0.9207, 0.8257], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0396, 0.0318, 0.0328, 0.0348, 0.0413, 0.0379, 0.0336], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 16:58:46,014 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1177, 2.3588, 1.0812, 1.3337, 1.8431, 1.2274, 2.9154, 1.6104], + device='cuda:2'), covar=tensor([0.0599, 0.0629, 0.0697, 0.1199, 0.0448, 0.0931, 0.0369, 0.0619], + device='cuda:2'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0054, 0.0055, 0.0083, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 16:58:47,720 INFO [finetune.py:976] (2/7) Epoch 5, batch 4200, loss[loss=0.2204, simple_loss=0.2878, pruned_loss=0.07653, over 4865.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2772, pruned_loss=0.07738, over 953153.86 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:58:53,014 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27118.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:58:58,409 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3171, 1.7032, 1.6300, 2.0566, 1.8351, 2.2082, 1.4491, 4.2868], + device='cuda:2'), covar=tensor([0.0685, 0.0805, 0.0778, 0.1192, 0.0665, 0.0573, 0.0805, 0.0120], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0039, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 16:58:58,900 INFO [optim.py:369] (2/7) 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 16:59:13,615 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27148.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:59:17,403 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-26 16:59:18,012 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-26 16:59:21,417 INFO [finetune.py:976] (2/7) Epoch 5, batch 4250, loss[loss=0.2322, simple_loss=0.2896, pruned_loss=0.0874, over 4901.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.275, pruned_loss=0.07709, over 954534.55 frames. ], batch size: 35, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 16:59:54,272 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27209.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 16:59:55,346 INFO [finetune.py:976] (2/7) Epoch 5, batch 4300, loss[loss=0.1954, simple_loss=0.2575, pruned_loss=0.06668, over 4791.00 frames. ], tot_loss[loss=0.2113, simple_loss=0.2713, pruned_loss=0.07561, over 953158.26 frames. ], batch size: 29, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:00:09,461 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3617, 1.5029, 1.3207, 1.5710, 1.3966, 1.3013, 1.4007, 1.1656], + device='cuda:2'), covar=tensor([0.1735, 0.1568, 0.1259, 0.1358, 0.3478, 0.1606, 0.1841, 0.2466], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0328, 0.0237, 0.0302, 0.0321, 0.0282, 0.0271, 0.0294], + device='cuda:2'), out_proj_covar=tensor([1.2610e-04, 1.3325e-04, 9.6561e-05, 1.2140e-04, 1.3217e-04, 1.1413e-04, + 1.1131e-04, 1.1887e-04], device='cuda:2') +2023-04-26 17:00:17,881 INFO [optim.py:369] (2/7) 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 17:01:04,151 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27260.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 17:01:04,678 INFO [finetune.py:976] (2/7) Epoch 5, batch 4350, loss[loss=0.2174, simple_loss=0.2735, pruned_loss=0.08072, over 4906.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2668, pruned_loss=0.07409, over 951275.72 frames. ], batch size: 32, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:01:35,652 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 17:01:46,636 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27306.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:01:49,555 INFO [finetune.py:976] (2/7) Epoch 5, batch 4400, loss[loss=0.2239, simple_loss=0.2741, pruned_loss=0.08685, over 4781.00 frames. ], tot_loss[loss=0.2101, simple_loss=0.2694, pruned_loss=0.07538, over 951129.53 frames. ], batch size: 26, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:01:50,938 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5432, 1.8135, 2.3878, 2.9618, 2.3434, 1.8015, 1.6034, 2.1726], + device='cuda:2'), covar=tensor([0.4092, 0.4475, 0.1981, 0.3319, 0.3916, 0.3298, 0.5236, 0.3606], + device='cuda:2'), in_proj_covar=tensor([0.0277, 0.0260, 0.0221, 0.0330, 0.0220, 0.0230, 0.0245, 0.0196], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:02:00,135 INFO [optim.py:369] (2/7) 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 17:02:11,589 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-26 17:02:13,766 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7250, 2.1671, 1.8161, 2.1501, 1.7148, 1.8640, 1.8859, 1.5632], + device='cuda:2'), covar=tensor([0.2305, 0.1686, 0.1094, 0.1465, 0.3583, 0.1483, 0.2190, 0.3019], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0325, 0.0235, 0.0300, 0.0318, 0.0280, 0.0269, 0.0292], + device='cuda:2'), out_proj_covar=tensor([1.2473e-04, 1.3205e-04, 9.5849e-05, 1.2047e-04, 1.3078e-04, 1.1329e-04, + 1.1050e-04, 1.1786e-04], device='cuda:2') +2023-04-26 17:02:23,093 INFO [finetune.py:976] (2/7) Epoch 5, batch 4450, loss[loss=0.2483, simple_loss=0.3131, pruned_loss=0.09169, over 4800.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2729, pruned_loss=0.0763, over 951720.26 frames. ], batch size: 45, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:02:30,760 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 17:02:49,463 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3329, 0.6951, 1.1546, 1.6629, 1.4476, 1.2390, 1.2388, 1.2827], + device='cuda:2'), covar=tensor([0.9873, 1.3089, 1.3642, 1.8538, 1.1580, 1.6208, 1.6810, 1.3813], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0452, 0.0537, 0.0556, 0.0448, 0.0470, 0.0483, 0.0484], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:02:56,797 INFO [finetune.py:976] (2/7) Epoch 5, batch 4500, loss[loss=0.2065, simple_loss=0.2604, pruned_loss=0.0763, over 4786.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2745, pruned_loss=0.07666, over 950938.90 frames. ], batch size: 51, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:03:17,441 INFO [optim.py:369] (2/7) 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 17:03:46,019 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0303, 2.5665, 1.0946, 1.3219, 2.0214, 1.2047, 3.3329, 1.5818], + device='cuda:2'), covar=tensor([0.0712, 0.0811, 0.0843, 0.1313, 0.0515, 0.1049, 0.0232, 0.0685], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0052, 0.0049, 0.0053, 0.0054, 0.0082, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 17:04:02,848 INFO [finetune.py:976] (2/7) Epoch 5, batch 4550, loss[loss=0.1818, simple_loss=0.2507, pruned_loss=0.05643, over 4801.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.276, pruned_loss=0.07724, over 953440.50 frames. ], batch size: 25, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:04:19,866 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-26 17:04:26,683 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4439, 1.0736, 0.3815, 1.1608, 1.1162, 1.3333, 1.1912, 1.2057], + device='cuda:2'), covar=tensor([0.0567, 0.0458, 0.0493, 0.0633, 0.0318, 0.0602, 0.0573, 0.0639], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0024, 0.0031, 0.0022, 0.0030, 0.0030, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 17:04:31,477 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27504.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:04:36,158 INFO [finetune.py:976] (2/7) Epoch 5, batch 4600, loss[loss=0.1924, simple_loss=0.2589, pruned_loss=0.06299, over 4825.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2747, pruned_loss=0.07641, over 952956.42 frames. ], batch size: 30, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:04:46,279 INFO [optim.py:369] (2/7) 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 17:05:08,928 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27560.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:05:09,447 INFO [finetune.py:976] (2/7) Epoch 5, batch 4650, loss[loss=0.1864, simple_loss=0.2549, pruned_loss=0.05893, over 4816.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2717, pruned_loss=0.07538, over 954782.19 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:05:40,055 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27608.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:05:43,535 INFO [finetune.py:976] (2/7) Epoch 5, batch 4700, loss[loss=0.1994, simple_loss=0.2662, pruned_loss=0.06634, over 4833.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2676, pruned_loss=0.07377, over 955593.88 frames. ], batch size: 30, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:05:54,152 INFO [optim.py:369] (2/7) 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 17:05:54,459 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-26 17:06:23,331 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:06:34,204 INFO [finetune.py:976] (2/7) Epoch 5, batch 4750, loss[loss=0.2071, simple_loss=0.2607, pruned_loss=0.07677, over 4925.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.265, pruned_loss=0.07272, over 956979.68 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:07:40,899 INFO [finetune.py:976] (2/7) Epoch 5, batch 4800, loss[loss=0.2005, simple_loss=0.2728, pruned_loss=0.06407, over 4848.00 frames. ], tot_loss[loss=0.208, simple_loss=0.2678, pruned_loss=0.07412, over 955553.41 frames. ], batch size: 49, lr: 3.93e-03, grad_scale: 16.0 +2023-04-26 17:07:48,560 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27714.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:08:01,165 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 17:08:02,771 INFO [optim.py:369] (2/7) 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 17:08:41,287 INFO [finetune.py:976] (2/7) Epoch 5, batch 4850, loss[loss=0.1915, simple_loss=0.2585, pruned_loss=0.06222, over 4766.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2717, pruned_loss=0.07515, over 954629.62 frames. ], batch size: 28, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:08:43,254 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8770, 2.4840, 2.0685, 2.4259, 1.8807, 2.2119, 2.1635, 1.7236], + device='cuda:2'), covar=tensor([0.2351, 0.1300, 0.0946, 0.1412, 0.2899, 0.1288, 0.2207, 0.2859], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0325, 0.0235, 0.0300, 0.0319, 0.0280, 0.0269, 0.0291], + device='cuda:2'), out_proj_covar=tensor([1.2491e-04, 1.3191e-04, 9.5879e-05, 1.2049e-04, 1.3111e-04, 1.1311e-04, + 1.1045e-04, 1.1760e-04], device='cuda:2') +2023-04-26 17:08:50,943 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27775.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:09:14,396 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 4900, loss[loss=0.2421, simple_loss=0.3002, pruned_loss=0.09202, over 4825.00 frames. ], tot_loss[loss=0.2118, simple_loss=0.2727, pruned_loss=0.07541, over 953228.71 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:09:30,227 INFO [optim.py:369] (2/7) 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 17:09:46,108 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27852.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:09:52,539 INFO [finetune.py:976] (2/7) Epoch 5, batch 4950, loss[loss=0.2168, simple_loss=0.2726, pruned_loss=0.0805, over 4832.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2743, pruned_loss=0.07565, over 955054.07 frames. ], batch size: 30, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:10:31,679 INFO [finetune.py:976] (2/7) Epoch 5, batch 5000, loss[loss=0.1496, simple_loss=0.2149, pruned_loss=0.04215, over 4792.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2726, pruned_loss=0.07489, over 957023.94 frames. ], batch size: 29, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:10:33,626 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9009, 1.5689, 2.1513, 2.0982, 1.7267, 1.3551, 1.9095, 0.8303], + device='cuda:2'), covar=tensor([0.1045, 0.1199, 0.0794, 0.1189, 0.1183, 0.1648, 0.1084, 0.1485], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0076, 0.0076, 0.0069, 0.0080, 0.0097, 0.0082, 0.0078], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:10:38,332 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5545, 1.7577, 1.6458, 2.2515, 1.8915, 2.2764, 1.5295, 4.5371], + device='cuda:2'), covar=tensor([0.0655, 0.0803, 0.0805, 0.1159, 0.0657, 0.0541, 0.0783, 0.0123], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0041, 0.0040, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:10:42,359 INFO [optim.py:369] (2/7) 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 17:10:43,188 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.21 vs. limit=5.0 +2023-04-26 17:10:47,736 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7568, 3.9679, 0.9425, 2.2236, 2.3657, 2.7699, 2.3916, 1.0064], + device='cuda:2'), covar=tensor([0.1301, 0.0883, 0.2113, 0.1307, 0.0991, 0.1094, 0.1421, 0.2273], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0264, 0.0147, 0.0130, 0.0139, 0.0162, 0.0125, 0.0129], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:11:04,343 INFO [finetune.py:976] (2/7) Epoch 5, batch 5050, loss[loss=0.2193, simple_loss=0.2737, pruned_loss=0.08249, over 4889.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.27, pruned_loss=0.07443, over 958419.57 frames. ], batch size: 35, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:11:06,753 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8505, 1.6364, 2.0660, 2.2208, 2.0213, 1.7632, 1.9347, 1.9811], + device='cuda:2'), covar=tensor([1.0149, 1.3215, 1.4764, 1.3657, 1.0933, 1.6798, 1.6823, 1.3937], + device='cuda:2'), in_proj_covar=tensor([0.0420, 0.0454, 0.0538, 0.0559, 0.0449, 0.0472, 0.0484, 0.0484], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:11:10,329 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.5784, 1.6108, 1.7002, 1.2559, 1.7244, 1.4420, 2.2273, 1.4218], + device='cuda:2'), covar=tensor([0.4061, 0.1859, 0.4804, 0.3012, 0.1791, 0.2430, 0.1490, 0.4674], + device='cuda:2'), in_proj_covar=tensor([0.0349, 0.0353, 0.0434, 0.0366, 0.0394, 0.0384, 0.0391, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:11:32,603 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4601, 1.2163, 1.6528, 1.5390, 1.3629, 1.1672, 1.3077, 0.7159], + device='cuda:2'), covar=tensor([0.0626, 0.1034, 0.0590, 0.0836, 0.1001, 0.1467, 0.0790, 0.1118], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0077, 0.0076, 0.0070, 0.0080, 0.0097, 0.0083, 0.0078], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:12:05,263 INFO [finetune.py:976] (2/7) Epoch 5, batch 5100, loss[loss=0.1861, simple_loss=0.24, pruned_loss=0.06604, over 4895.00 frames. ], tot_loss[loss=0.207, simple_loss=0.267, pruned_loss=0.07352, over 958236.31 frames. ], batch size: 32, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:12:12,307 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1145, 1.6236, 1.3863, 1.8849, 1.6389, 2.2226, 1.3834, 3.7677], + device='cuda:2'), covar=tensor([0.0694, 0.0777, 0.0802, 0.1161, 0.0646, 0.0549, 0.0747, 0.0123], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:12:27,655 INFO [optim.py:369] (2/7) 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 17:13:12,922 INFO [finetune.py:976] (2/7) Epoch 5, batch 5150, loss[loss=0.2599, simple_loss=0.3189, pruned_loss=0.1004, over 4932.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2677, pruned_loss=0.07338, over 956870.46 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:13:31,479 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28070.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 17:13:54,614 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-04-26 17:14:08,529 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0434, 2.5925, 1.1349, 1.2967, 1.9805, 1.1284, 3.4534, 1.6920], + device='cuda:2'), covar=tensor([0.0746, 0.0761, 0.0865, 0.1411, 0.0546, 0.1092, 0.0251, 0.0686], + device='cuda:2'), in_proj_covar=tensor([0.0054, 0.0070, 0.0052, 0.0049, 0.0054, 0.0054, 0.0082, 0.0053], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 17:14:09,043 INFO [finetune.py:976] (2/7) Epoch 5, batch 5200, loss[loss=0.2458, simple_loss=0.3141, pruned_loss=0.08877, over 4903.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.272, pruned_loss=0.07521, over 953151.63 frames. ], batch size: 43, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:14:11,126 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-26 17:14:19,887 INFO [optim.py:369] (2/7) 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 17:14:42,336 INFO [finetune.py:976] (2/7) Epoch 5, batch 5250, loss[loss=0.1616, simple_loss=0.2312, pruned_loss=0.04602, over 4692.00 frames. ], tot_loss[loss=0.2132, simple_loss=0.2744, pruned_loss=0.07594, over 954785.99 frames. ], batch size: 23, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:14:49,550 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9043, 1.7324, 1.9988, 2.2620, 2.3088, 1.8408, 1.5518, 1.9418], + device='cuda:2'), covar=tensor([0.1042, 0.1180, 0.0696, 0.0666, 0.0650, 0.1042, 0.0942, 0.0649], + device='cuda:2'), in_proj_covar=tensor([0.0203, 0.0206, 0.0182, 0.0178, 0.0178, 0.0194, 0.0165, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:15:02,566 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2581, 1.5461, 1.5407, 1.7330, 1.5744, 1.7361, 1.6525, 1.6122], + device='cuda:2'), covar=tensor([0.8146, 1.0565, 0.8900, 0.7784, 0.9895, 1.3806, 1.0084, 0.9786], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0395, 0.0317, 0.0326, 0.0347, 0.0413, 0.0377, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:15:16,323 INFO [finetune.py:976] (2/7) Epoch 5, batch 5300, loss[loss=0.2088, simple_loss=0.276, pruned_loss=0.07084, over 4817.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2757, pruned_loss=0.0763, over 955159.59 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:15:27,005 INFO [optim.py:369] (2/7) 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 17:15:27,093 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5407, 3.6017, 2.6647, 4.1665, 3.7075, 3.5677, 1.4264, 3.5117], + device='cuda:2'), covar=tensor([0.1894, 0.1433, 0.3129, 0.1926, 0.2899, 0.1907, 0.6306, 0.2531], + device='cuda:2'), in_proj_covar=tensor([0.0248, 0.0221, 0.0256, 0.0315, 0.0305, 0.0257, 0.0277, 0.0278], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:15:28,324 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28229.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:15:46,199 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28255.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:15:49,729 INFO [finetune.py:976] (2/7) Epoch 5, batch 5350, loss[loss=0.1897, simple_loss=0.2585, pruned_loss=0.06043, over 4815.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2748, pruned_loss=0.0748, over 955996.61 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:16:07,239 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28287.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:16:10,011 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28290.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:16:23,679 INFO [finetune.py:976] (2/7) Epoch 5, batch 5400, loss[loss=0.1892, simple_loss=0.2563, pruned_loss=0.0611, over 4823.00 frames. ], tot_loss[loss=0.2089, simple_loss=0.2711, pruned_loss=0.07341, over 954547.37 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:16:27,308 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28316.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:16:34,348 INFO [optim.py:369] (2/7) 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 17:16:38,815 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-04-26 17:16:49,214 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28348.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:16:57,452 INFO [finetune.py:976] (2/7) Epoch 5, batch 5450, loss[loss=0.2044, simple_loss=0.2608, pruned_loss=0.07393, over 4827.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2674, pruned_loss=0.0721, over 954073.51 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:17:03,560 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28370.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:17:37,439 INFO [finetune.py:976] (2/7) Epoch 5, batch 5500, loss[loss=0.1866, simple_loss=0.2571, pruned_loss=0.05806, over 4773.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2647, pruned_loss=0.0711, over 956998.06 frames. ], batch size: 26, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:17:47,038 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28418.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:17:47,652 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8131, 3.7911, 1.2139, 2.0275, 2.3927, 2.5696, 2.2166, 1.3712], + device='cuda:2'), covar=tensor([0.1241, 0.0869, 0.1891, 0.1274, 0.0962, 0.1074, 0.1429, 0.1745], + device='cuda:2'), in_proj_covar=tensor([0.0124, 0.0267, 0.0148, 0.0130, 0.0141, 0.0164, 0.0126, 0.0130], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:17:59,138 INFO [optim.py:369] (2/7) 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 17:18:49,580 INFO [finetune.py:976] (2/7) Epoch 5, batch 5550, loss[loss=0.2303, simple_loss=0.2737, pruned_loss=0.09349, over 4893.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2664, pruned_loss=0.07263, over 954314.39 frames. ], batch size: 32, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:19:09,812 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28486.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:19:14,457 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28492.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:19:25,620 INFO [finetune.py:976] (2/7) Epoch 5, batch 5600, loss[loss=0.2208, simple_loss=0.2862, pruned_loss=0.07768, over 4806.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2708, pruned_loss=0.07399, over 954231.82 frames. ], batch size: 51, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:19:25,722 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6134, 2.1641, 1.6171, 1.5406, 1.2206, 1.2317, 1.6963, 1.1579], + device='cuda:2'), covar=tensor([0.2002, 0.1720, 0.1895, 0.2234, 0.2925, 0.2340, 0.1362, 0.2385], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0219, 0.0176, 0.0207, 0.0212, 0.0186, 0.0169, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:19:34,905 INFO [optim.py:369] (2/7) 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 17:19:47,073 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28547.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:19:51,031 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 5, batch 5650, loss[loss=0.2168, simple_loss=0.2774, pruned_loss=0.07815, over 4759.00 frames. ], tot_loss[loss=0.2114, simple_loss=0.274, pruned_loss=0.07439, over 953811.56 frames. ], batch size: 26, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:20:04,126 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-26 17:20:09,691 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28585.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:20:25,136 INFO [finetune.py:976] (2/7) Epoch 5, batch 5700, loss[loss=0.1763, simple_loss=0.2281, pruned_loss=0.06227, over 4340.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2715, pruned_loss=0.07458, over 938349.69 frames. ], batch size: 19, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:20:25,175 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28611.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 17:20:31,774 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3817, 2.6422, 1.2108, 1.6043, 2.0259, 1.4324, 3.2965, 1.8787], + device='cuda:2'), covar=tensor([0.0503, 0.0629, 0.0689, 0.1072, 0.0445, 0.0837, 0.0247, 0.0536], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0052, 0.0049, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 17:20:34,598 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 0, loss[loss=0.1703, simple_loss=0.2531, pruned_loss=0.04381, over 4783.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2531, pruned_loss=0.04381, over 4783.00 frames. ], batch size: 29, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:20:58,934 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 17:21:02,907 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1649, 1.6110, 1.3788, 1.8239, 1.5726, 1.9091, 1.3922, 3.0643], + device='cuda:2'), covar=tensor([0.0696, 0.0766, 0.0778, 0.1181, 0.0667, 0.0454, 0.0745, 0.0215], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:21:15,590 INFO [finetune.py:1010] (2/7) Epoch 6, validation: loss=0.1605, simple_loss=0.2337, pruned_loss=0.04366, over 2265189.00 frames. +2023-04-26 17:21:15,591 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 17:21:19,203 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28643.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:21:19,253 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2269, 1.3453, 1.4200, 1.5097, 1.5753, 1.1923, 0.7927, 1.3256], + device='cuda:2'), covar=tensor([0.1103, 0.1381, 0.0931, 0.0831, 0.0774, 0.1138, 0.1261, 0.0781], + device='cuda:2'), in_proj_covar=tensor([0.0202, 0.0204, 0.0181, 0.0177, 0.0177, 0.0192, 0.0165, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:21:21,106 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2695, 2.1800, 2.4921, 2.7255, 2.7341, 2.0549, 1.7565, 2.2431], + device='cuda:2'), covar=tensor([0.0993, 0.1016, 0.0538, 0.0652, 0.0580, 0.1090, 0.1006, 0.0662], + device='cuda:2'), in_proj_covar=tensor([0.0202, 0.0204, 0.0181, 0.0177, 0.0177, 0.0192, 0.0165, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:21:29,883 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6269, 1.1656, 1.2836, 1.3323, 1.8767, 1.4746, 1.1816, 1.2470], + device='cuda:2'), covar=tensor([0.1804, 0.1595, 0.2107, 0.1423, 0.0970, 0.1530, 0.1934, 0.1911], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0331, 0.0349, 0.0305, 0.0341, 0.0336, 0.0305, 0.0351], + device='cuda:2'), 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:2') +2023-04-26 17:21:59,786 INFO [finetune.py:976] (2/7) Epoch 6, batch 50, loss[loss=0.197, simple_loss=0.2701, pruned_loss=0.06192, over 4779.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.278, pruned_loss=0.07808, over 213532.34 frames. ], batch size: 29, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:22:24,839 INFO [optim.py:369] (2/7) 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 17:22:33,640 INFO [finetune.py:976] (2/7) Epoch 6, batch 100, loss[loss=0.182, simple_loss=0.2518, pruned_loss=0.05607, over 4767.00 frames. ], tot_loss[loss=0.2084, simple_loss=0.2696, pruned_loss=0.07356, over 379771.07 frames. ], batch size: 26, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:22:43,757 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0988, 1.4813, 1.3465, 1.6885, 1.4716, 1.8647, 1.3063, 3.0156], + device='cuda:2'), covar=tensor([0.0741, 0.0780, 0.0789, 0.1174, 0.0674, 0.0504, 0.0760, 0.0177], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:23:06,834 INFO [finetune.py:976] (2/7) Epoch 6, batch 150, loss[loss=0.2292, simple_loss=0.2814, pruned_loss=0.0885, over 4895.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2653, pruned_loss=0.07185, over 509366.99 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:23:36,887 INFO [optim.py:369] (2/7) 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 17:23:45,693 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5388, 0.9457, 1.4101, 1.9477, 1.6952, 1.4891, 1.4905, 1.5416], + device='cuda:2'), covar=tensor([0.9476, 1.2635, 1.2673, 1.3365, 1.0598, 1.4763, 1.4042, 1.2014], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0452, 0.0536, 0.0556, 0.0447, 0.0471, 0.0483, 0.0483], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:23:57,269 INFO [finetune.py:976] (2/7) Epoch 6, batch 200, loss[loss=0.2301, simple_loss=0.2798, pruned_loss=0.09017, over 4850.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2637, pruned_loss=0.07276, over 609683.51 frames. ], batch size: 44, lr: 3.92e-03, grad_scale: 32.0 +2023-04-26 17:23:59,203 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28842.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:24:07,618 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4226, 0.9067, 1.4111, 1.7921, 1.5732, 1.3607, 1.4091, 1.4456], + device='cuda:2'), covar=tensor([0.9092, 1.2108, 1.1930, 1.3894, 1.0430, 1.4121, 1.3570, 1.1523], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0452, 0.0536, 0.0558, 0.0448, 0.0471, 0.0483, 0.0484], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:24:08,147 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:24:22,280 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28861.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:24:33,498 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7152, 2.3824, 1.7572, 1.6080, 1.2647, 1.3266, 1.9217, 1.2617], + device='cuda:2'), covar=tensor([0.1860, 0.1672, 0.1691, 0.2141, 0.2832, 0.2179, 0.1261, 0.2285], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0220, 0.0177, 0.0207, 0.0212, 0.0186, 0.0170, 0.0193], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:24:55,342 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28885.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:25:02,844 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-04-26 17:25:03,860 INFO [finetune.py:976] (2/7) Epoch 6, batch 250, loss[loss=0.1971, simple_loss=0.258, pruned_loss=0.06809, over 4770.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2689, pruned_loss=0.07425, over 686682.23 frames. ], batch size: 28, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:25:15,797 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1095, 1.4568, 1.2346, 1.6579, 1.4195, 1.8084, 1.3101, 3.3343], + device='cuda:2'), covar=tensor([0.0715, 0.0808, 0.0830, 0.1205, 0.0690, 0.0569, 0.0765, 0.0150], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:25:29,050 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28911.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:25:40,799 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 17:25:54,903 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28933.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:25:59,501 INFO [finetune.py:976] (2/7) Epoch 6, batch 300, loss[loss=0.2142, simple_loss=0.2758, pruned_loss=0.0763, over 4757.00 frames. ], tot_loss[loss=0.2116, simple_loss=0.273, pruned_loss=0.0751, over 745738.20 frames. ], batch size: 27, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:26:08,041 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28943.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:26:28,686 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 350, loss[loss=0.1879, simple_loss=0.2611, pruned_loss=0.05742, over 4828.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2746, pruned_loss=0.07576, over 792289.79 frames. ], batch size: 30, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:27:11,878 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28991.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:27:24,969 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29002.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:27:50,842 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3365, 1.3202, 1.4194, 1.6688, 1.6403, 1.3836, 0.9584, 1.4425], + device='cuda:2'), covar=tensor([0.1034, 0.1364, 0.0874, 0.0627, 0.0683, 0.1045, 0.1040, 0.0672], + device='cuda:2'), in_proj_covar=tensor([0.0202, 0.0205, 0.0180, 0.0177, 0.0177, 0.0193, 0.0165, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:27:50,863 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0958, 0.6441, 0.9404, 0.7524, 1.2777, 0.9249, 0.7974, 1.0015], + device='cuda:2'), covar=tensor([0.1819, 0.2031, 0.2226, 0.1886, 0.1262, 0.1792, 0.2274, 0.2457], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0329, 0.0348, 0.0305, 0.0340, 0.0335, 0.0306, 0.0348], + device='cuda:2'), out_proj_covar=tensor([6.6176e-05, 7.0315e-05, 7.5186e-05, 6.3630e-05, 7.2078e-05, 7.2571e-05, + 6.6281e-05, 7.4957e-05], device='cuda:2') +2023-04-26 17:27:53,123 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 400, loss[loss=0.1829, simple_loss=0.2573, pruned_loss=0.05431, over 4838.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2748, pruned_loss=0.07555, over 828238.36 frames. ], batch size: 47, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:28:07,448 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9988, 1.4463, 1.8147, 1.9689, 1.6999, 1.3654, 0.9186, 1.4383], + device='cuda:2'), covar=tensor([0.4002, 0.4301, 0.1942, 0.3115, 0.3801, 0.3367, 0.5714, 0.3348], + device='cuda:2'), in_proj_covar=tensor([0.0277, 0.0259, 0.0222, 0.0331, 0.0220, 0.0230, 0.0244, 0.0194], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:28:16,912 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29063.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:28:33,168 INFO [finetune.py:976] (2/7) Epoch 6, batch 450, loss[loss=0.1592, simple_loss=0.2227, pruned_loss=0.04789, over 4893.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.2731, pruned_loss=0.07409, over 857743.00 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:28:36,871 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0608, 2.5087, 2.2328, 2.4033, 2.2058, 2.5187, 2.3738, 2.3197], + device='cuda:2'), covar=tensor([0.5971, 1.0474, 0.9609, 0.8161, 0.9458, 1.2873, 1.1339, 1.0133], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0398, 0.0318, 0.0328, 0.0348, 0.0414, 0.0378, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:28:59,673 INFO [optim.py:369] (2/7) 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 17:29:06,386 INFO [finetune.py:976] (2/7) Epoch 6, batch 500, loss[loss=0.1876, simple_loss=0.2448, pruned_loss=0.06521, over 4246.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2702, pruned_loss=0.07351, over 879372.39 frames. ], batch size: 65, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:29:08,297 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29142.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:13,402 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29148.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:31,341 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9171, 2.0201, 1.9082, 1.5381, 2.1434, 1.7089, 2.7407, 1.5899], + device='cuda:2'), covar=tensor([0.4090, 0.1798, 0.5226, 0.3262, 0.1875, 0.2796, 0.1156, 0.4691], + device='cuda:2'), in_proj_covar=tensor([0.0353, 0.0360, 0.0443, 0.0373, 0.0401, 0.0390, 0.0395, 0.0427], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:29:39,685 INFO [finetune.py:976] (2/7) Epoch 6, batch 550, loss[loss=0.1931, simple_loss=0.2673, pruned_loss=0.05943, over 4812.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2682, pruned_loss=0.07325, over 896501.41 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:29:40,346 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29190.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:44,457 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29196.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:29:59,119 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29217.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:30:06,233 INFO [optim.py:369] (2/7) 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 17:30:11,158 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:30:12,894 INFO [finetune.py:976] (2/7) Epoch 6, batch 600, loss[loss=0.2328, simple_loss=0.2942, pruned_loss=0.08568, over 4868.00 frames. ], tot_loss[loss=0.2079, simple_loss=0.2691, pruned_loss=0.0734, over 909088.13 frames. ], batch size: 34, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:30:13,617 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6677, 1.7208, 1.6647, 1.1622, 1.7419, 1.5257, 2.2982, 1.3715], + device='cuda:2'), covar=tensor([0.3408, 0.1516, 0.4850, 0.2946, 0.1679, 0.2111, 0.1206, 0.4264], + device='cuda:2'), in_proj_covar=tensor([0.0351, 0.0359, 0.0441, 0.0371, 0.0399, 0.0388, 0.0394, 0.0425], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:30:47,030 INFO [finetune.py:976] (2/7) Epoch 6, batch 650, loss[loss=0.2193, simple_loss=0.2949, pruned_loss=0.07189, over 4929.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2725, pruned_loss=0.07453, over 921390.12 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:30:57,626 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29297.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:31:07,014 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8713, 2.1197, 1.9191, 2.1350, 1.8646, 2.1437, 1.9996, 1.9651], + device='cuda:2'), covar=tensor([0.6744, 1.1441, 0.9087, 0.8041, 0.9682, 1.3654, 1.2644, 1.0737], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0394, 0.0315, 0.0325, 0.0345, 0.0411, 0.0374, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:31:42,431 INFO [optim.py:369] (2/7) 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 17:32:01,093 INFO [finetune.py:976] (2/7) Epoch 6, batch 700, loss[loss=0.1708, simple_loss=0.2395, pruned_loss=0.05109, over 4812.00 frames. ], tot_loss[loss=0.212, simple_loss=0.2744, pruned_loss=0.07478, over 929659.89 frames. ], batch size: 25, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:32:25,508 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29358.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:33:07,832 INFO [finetune.py:976] (2/7) Epoch 6, batch 750, loss[loss=0.2272, simple_loss=0.2814, pruned_loss=0.08646, over 4913.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2756, pruned_loss=0.07501, over 934326.51 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:33:11,528 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3913, 0.9837, 0.5051, 1.1110, 1.0486, 1.2941, 1.1572, 1.1618], + device='cuda:2'), covar=tensor([0.0595, 0.0472, 0.0472, 0.0634, 0.0342, 0.0601, 0.0566, 0.0681], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 17:33:12,179 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8972, 1.4011, 1.4330, 1.5948, 2.1851, 1.7507, 1.4348, 1.3858], + device='cuda:2'), covar=tensor([0.1584, 0.1706, 0.2065, 0.1377, 0.0861, 0.1697, 0.2120, 0.1842], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0331, 0.0351, 0.0307, 0.0341, 0.0336, 0.0306, 0.0351], + device='cuda:2'), out_proj_covar=tensor([6.6674e-05, 7.0771e-05, 7.5908e-05, 6.4122e-05, 7.2096e-05, 7.2793e-05, + 6.6338e-05, 7.5628e-05], device='cuda:2') +2023-04-26 17:33:42,205 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1533, 1.9685, 1.6980, 1.6228, 1.9897, 1.6825, 2.3803, 1.4835], + device='cuda:2'), covar=tensor([0.3242, 0.1299, 0.3988, 0.2769, 0.1572, 0.1974, 0.1346, 0.4187], + device='cuda:2'), in_proj_covar=tensor([0.0352, 0.0359, 0.0440, 0.0371, 0.0399, 0.0388, 0.0393, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:34:03,539 INFO [optim.py:369] (2/7) 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 17:34:14,519 INFO [finetune.py:976] (2/7) Epoch 6, batch 800, loss[loss=0.2168, simple_loss=0.2723, pruned_loss=0.08061, over 4830.00 frames. ], tot_loss[loss=0.214, simple_loss=0.2763, pruned_loss=0.07589, over 937630.95 frames. ], batch size: 30, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:34:44,625 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-26 17:34:48,041 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6080, 1.5078, 0.4967, 1.2639, 1.4997, 1.4734, 1.3619, 1.4036], + device='cuda:2'), covar=tensor([0.0530, 0.0391, 0.0451, 0.0585, 0.0280, 0.0582, 0.0528, 0.0612], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 17:34:48,534 INFO [finetune.py:976] (2/7) Epoch 6, batch 850, loss[loss=0.2198, simple_loss=0.2742, pruned_loss=0.08272, over 4816.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2734, pruned_loss=0.07502, over 940691.63 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 16.0 +2023-04-26 17:35:00,866 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-04-26 17:35:07,986 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29517.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:35:14,514 INFO [optim.py:369] (2/7) 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 17:35:18,178 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-04-26 17:35:22,137 INFO [finetune.py:976] (2/7) Epoch 6, batch 900, loss[loss=0.1888, simple_loss=0.2545, pruned_loss=0.06153, over 4900.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.27, pruned_loss=0.07372, over 945496.04 frames. ], batch size: 35, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:35:30,012 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:35:38,988 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:35:44,795 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3281, 1.7575, 1.4209, 2.0551, 1.7641, 2.1435, 1.5560, 4.0404], + device='cuda:2'), covar=tensor([0.0626, 0.0708, 0.0810, 0.1128, 0.0622, 0.0654, 0.0700, 0.0139], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:35:47,246 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3058, 1.6508, 1.4194, 1.8634, 1.6364, 1.9756, 1.4424, 3.5860], + device='cuda:2'), covar=tensor([0.0695, 0.0776, 0.0855, 0.1207, 0.0689, 0.0555, 0.0781, 0.0150], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:35:55,008 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-04-26 17:35:56,021 INFO [finetune.py:976] (2/7) Epoch 6, batch 950, loss[loss=0.1554, simple_loss=0.2266, pruned_loss=0.04204, over 4920.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2672, pruned_loss=0.07241, over 946673.12 frames. ], batch size: 37, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:35:57,878 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29592.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:36:10,616 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29612.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:36:27,225 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 1000, loss[loss=0.1799, simple_loss=0.2404, pruned_loss=0.05966, over 4769.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2701, pruned_loss=0.07363, over 950465.99 frames. ], batch size: 26, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:36:57,627 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8094, 1.6935, 1.9806, 2.2262, 2.2383, 1.8722, 1.4781, 1.9327], + device='cuda:2'), covar=tensor([0.1069, 0.1143, 0.0688, 0.0651, 0.0630, 0.0935, 0.1094, 0.0644], + device='cuda:2'), in_proj_covar=tensor([0.0203, 0.0206, 0.0181, 0.0179, 0.0179, 0.0194, 0.0166, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:36:58,942 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 17:37:09,284 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29658.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:37:11,706 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 1050, loss[loss=0.2056, simple_loss=0.2753, pruned_loss=0.0679, over 4856.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2719, pruned_loss=0.07346, over 953144.24 frames. ], batch size: 44, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:37:52,287 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8059, 1.7137, 2.0507, 2.1641, 1.9920, 1.6629, 1.8036, 1.8768], + device='cuda:2'), covar=tensor([0.9974, 1.2047, 1.4612, 1.3367, 1.0226, 1.7213, 1.6612, 1.3614], + device='cuda:2'), in_proj_covar=tensor([0.0414, 0.0443, 0.0528, 0.0549, 0.0443, 0.0465, 0.0478, 0.0476], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:38:08,202 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2355, 1.5630, 1.2864, 1.8258, 1.5773, 1.8376, 1.3902, 3.2998], + device='cuda:2'), covar=tensor([0.0694, 0.0842, 0.0876, 0.1159, 0.0671, 0.0565, 0.0807, 0.0156], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0041, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 17:38:13,521 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29706.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:38:34,972 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29723.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 17:38:37,903 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 17:38:39,284 INFO [optim.py:369] (2/7) 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 17:38:57,680 INFO [finetune.py:976] (2/7) Epoch 6, batch 1100, loss[loss=0.1872, simple_loss=0.2555, pruned_loss=0.05947, over 4799.00 frames. ], tot_loss[loss=0.2104, simple_loss=0.2733, pruned_loss=0.0738, over 953500.29 frames. ], batch size: 40, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:39:11,284 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-26 17:39:58,701 INFO [finetune.py:976] (2/7) Epoch 6, batch 1150, loss[loss=0.1927, simple_loss=0.2539, pruned_loss=0.06572, over 4775.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2736, pruned_loss=0.07404, over 951762.77 frames. ], batch size: 51, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:40:07,753 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29802.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:40:24,366 INFO [optim.py:369] (2/7) 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 17:40:31,970 INFO [finetune.py:976] (2/7) Epoch 6, batch 1200, loss[loss=0.1713, simple_loss=0.2473, pruned_loss=0.0476, over 4816.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2718, pruned_loss=0.07322, over 953827.64 frames. ], batch size: 38, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:40:48,167 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29863.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:41:05,310 INFO [finetune.py:976] (2/7) Epoch 6, batch 1250, loss[loss=0.2443, simple_loss=0.2873, pruned_loss=0.1007, over 4894.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2694, pruned_loss=0.07284, over 955063.22 frames. ], batch size: 35, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:41:07,201 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29892.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:41:17,262 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29907.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:41:30,521 INFO [optim.py:369] (2/7) 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 17:41:30,646 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6107, 1.4421, 0.6151, 1.3317, 1.4903, 1.5172, 1.4068, 1.3931], + device='cuda:2'), covar=tensor([0.0522, 0.0457, 0.0446, 0.0642, 0.0314, 0.0601, 0.0583, 0.0632], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0047, 0.0050], + device='cuda:2') +2023-04-26 17:41:38,730 INFO [finetune.py:976] (2/7) Epoch 6, batch 1300, loss[loss=0.1532, simple_loss=0.2229, pruned_loss=0.04173, over 4798.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2663, pruned_loss=0.07142, over 957973.71 frames. ], batch size: 29, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:41:39,345 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:41:41,888 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1760, 1.5687, 2.0351, 2.5539, 1.9514, 1.5983, 1.3626, 1.7778], + device='cuda:2'), covar=tensor([0.4005, 0.4532, 0.2030, 0.3094, 0.3562, 0.3380, 0.5554, 0.3321], + device='cuda:2'), in_proj_covar=tensor([0.0275, 0.0257, 0.0219, 0.0327, 0.0217, 0.0229, 0.0241, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:42:12,107 INFO [finetune.py:976] (2/7) Epoch 6, batch 1350, loss[loss=0.1921, simple_loss=0.2566, pruned_loss=0.06376, over 4764.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2666, pruned_loss=0.07152, over 957496.87 frames. ], batch size: 54, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:42:32,672 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 17:42:52,096 INFO [finetune.py:976] (2/7) Epoch 6, batch 1400, loss[loss=0.1724, simple_loss=0.255, pruned_loss=0.04489, over 4928.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2695, pruned_loss=0.07258, over 957370.98 frames. ], batch size: 33, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:43:32,887 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 17:43:56,757 INFO [finetune.py:976] (2/7) Epoch 6, batch 1450, loss[loss=0.2302, simple_loss=0.3029, pruned_loss=0.07875, over 4824.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2709, pruned_loss=0.07326, over 955941.78 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:44:42,766 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30125.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:44:44,031 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-04-26 17:44:44,462 INFO [optim.py:369] (2/7) 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 17:44:57,143 INFO [finetune.py:976] (2/7) Epoch 6, batch 1500, loss[loss=0.2244, simple_loss=0.2912, pruned_loss=0.07883, over 4928.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2719, pruned_loss=0.07373, over 955570.18 frames. ], batch size: 33, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:45:25,798 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30158.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:45:59,256 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30186.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:46:00,950 INFO [finetune.py:976] (2/7) Epoch 6, batch 1550, loss[loss=0.1949, simple_loss=0.2531, pruned_loss=0.06836, over 4819.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2724, pruned_loss=0.07384, over 954767.73 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:46:30,715 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30207.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:46:31,933 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30209.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:46:56,027 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 17:46:56,313 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 1600, loss[loss=0.191, simple_loss=0.2553, pruned_loss=0.06333, over 4846.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2706, pruned_loss=0.07366, over 955369.14 frames. ], batch size: 44, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:47:09,620 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5422, 1.5734, 1.7941, 2.0667, 2.0390, 1.6155, 1.2217, 1.7491], + device='cuda:2'), covar=tensor([0.0952, 0.1194, 0.0700, 0.0556, 0.0621, 0.0922, 0.0937, 0.0582], + device='cuda:2'), in_proj_covar=tensor([0.0203, 0.0207, 0.0182, 0.0179, 0.0180, 0.0194, 0.0166, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:47:19,488 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 17:47:31,233 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30255.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:47:41,431 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30270.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:47:53,334 INFO [finetune.py:976] (2/7) Epoch 6, batch 1650, loss[loss=0.1418, simple_loss=0.2086, pruned_loss=0.03754, over 4816.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.268, pruned_loss=0.0728, over 955902.01 frames. ], batch size: 25, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:48:13,368 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30318.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 17:48:19,321 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 1700, loss[loss=0.2344, simple_loss=0.2864, pruned_loss=0.09124, over 4821.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2658, pruned_loss=0.07202, over 957317.66 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:48:50,854 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-04-26 17:48:51,278 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 6, batch 1750, loss[loss=0.244, simple_loss=0.3163, pruned_loss=0.08582, over 4113.00 frames. ], tot_loss[loss=0.2064, simple_loss=0.2679, pruned_loss=0.07241, over 956269.69 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:49:33,071 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-04-26 17:50:02,937 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-26 17:50:03,365 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30421.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:50:08,042 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 1800, loss[loss=0.1909, simple_loss=0.2673, pruned_loss=0.05721, over 4808.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2716, pruned_loss=0.07352, over 954983.01 frames. ], batch size: 51, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:50:43,031 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:51:02,125 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6365, 1.7898, 1.7752, 1.9956, 1.6998, 1.9687, 1.8559, 1.8571], + device='cuda:2'), covar=tensor([0.7870, 1.1125, 0.9235, 0.7692, 1.0111, 1.4131, 1.1559, 0.9620], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0394, 0.0317, 0.0327, 0.0345, 0.0411, 0.0374, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:51:09,061 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 1850, loss[loss=0.2046, simple_loss=0.279, pruned_loss=0.06517, over 4818.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2732, pruned_loss=0.07457, over 954720.81 frames. ], batch size: 38, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:51:23,172 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-26 17:51:24,865 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:51:40,049 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 1900, loss[loss=0.1924, simple_loss=0.2539, pruned_loss=0.0655, over 4164.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2717, pruned_loss=0.07337, over 951723.09 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:51:47,293 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7733, 2.2289, 5.6329, 5.2483, 4.9283, 5.3961, 4.8329, 4.9835], + device='cuda:2'), covar=tensor([0.5650, 0.4901, 0.0937, 0.1761, 0.0875, 0.1623, 0.0914, 0.1185], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0306, 0.0417, 0.0420, 0.0356, 0.0411, 0.0319, 0.0375], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:52:17,583 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:52:27,677 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4722, 1.6802, 1.4191, 0.8969, 1.1749, 1.1759, 1.3482, 1.1348], + device='cuda:2'), covar=tensor([0.1981, 0.1597, 0.1740, 0.2222, 0.2809, 0.2200, 0.1302, 0.2258], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0220, 0.0176, 0.0207, 0.0213, 0.0186, 0.0168, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:52:30,530 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4583, 3.9038, 0.9360, 2.0783, 2.1359, 2.7094, 2.2707, 1.0135], + device='cuda:2'), covar=tensor([0.1412, 0.0868, 0.2052, 0.1297, 0.1064, 0.1020, 0.1394, 0.2048], + device='cuda:2'), in_proj_covar=tensor([0.0121, 0.0260, 0.0146, 0.0127, 0.0137, 0.0159, 0.0122, 0.0126], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:52:34,610 INFO [finetune.py:976] (2/7) Epoch 6, batch 1950, loss[loss=0.2422, simple_loss=0.3079, pruned_loss=0.08826, over 4806.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2713, pruned_loss=0.07373, over 951129.62 frames. ], batch size: 40, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:52:43,399 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8382, 2.2080, 1.8711, 2.0925, 1.5214, 1.8350, 1.9393, 1.5987], + device='cuda:2'), covar=tensor([0.1643, 0.0984, 0.0872, 0.1100, 0.3281, 0.1231, 0.1435, 0.2013], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0326, 0.0236, 0.0299, 0.0319, 0.0280, 0.0266, 0.0292], + device='cuda:2'), out_proj_covar=tensor([1.2417e-04, 1.3255e-04, 9.5952e-05, 1.2013e-04, 1.3115e-04, 1.1338e-04, + 1.0925e-04, 1.1764e-04], device='cuda:2') +2023-04-26 17:53:00,332 INFO [optim.py:369] (2/7) 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 17:53:06,228 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.8100, 4.8156, 3.3963, 5.5929, 4.8966, 4.8515, 2.5046, 4.7458], + device='cuda:2'), covar=tensor([0.1273, 0.0820, 0.2651, 0.0759, 0.2888, 0.1679, 0.5134, 0.1970], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0220, 0.0255, 0.0312, 0.0303, 0.0256, 0.0275, 0.0276], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:53:06,344 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 17:53:08,042 INFO [finetune.py:976] (2/7) Epoch 6, batch 2000, loss[loss=0.1597, simple_loss=0.2295, pruned_loss=0.04495, over 4906.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2677, pruned_loss=0.07199, over 954475.87 frames. ], batch size: 32, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:53:14,783 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2731, 2.6407, 1.1941, 1.4921, 2.0922, 1.4392, 3.1968, 1.7454], + device='cuda:2'), covar=tensor([0.0577, 0.0785, 0.0856, 0.1037, 0.0414, 0.0851, 0.0205, 0.0539], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 17:53:17,891 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30655.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:53:20,213 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6737, 3.9137, 0.6708, 2.1838, 2.2359, 2.6536, 2.4242, 0.9548], + device='cuda:2'), covar=tensor([0.1316, 0.0815, 0.2267, 0.1211, 0.0991, 0.1044, 0.1258, 0.2148], + device='cuda:2'), in_proj_covar=tensor([0.0121, 0.0260, 0.0146, 0.0127, 0.0137, 0.0159, 0.0123, 0.0126], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:53:41,350 INFO [finetune.py:976] (2/7) Epoch 6, batch 2050, loss[loss=0.1731, simple_loss=0.2475, pruned_loss=0.04941, over 4818.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.264, pruned_loss=0.0706, over 955846.74 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:53:57,184 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4667, 1.6971, 1.3110, 0.9092, 1.1798, 1.1443, 1.2720, 1.0956], + device='cuda:2'), covar=tensor([0.1989, 0.1571, 0.1901, 0.2246, 0.2792, 0.2312, 0.1437, 0.2375], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0219, 0.0175, 0.0205, 0.0211, 0.0185, 0.0167, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 17:53:58,927 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30716.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:53:58,971 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30716.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:54:12,375 INFO [optim.py:369] (2/7) 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 17:54:14,795 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6115, 1.1330, 4.1850, 3.5465, 3.7437, 3.9747, 3.8403, 3.4764], + device='cuda:2'), covar=tensor([0.9528, 0.9035, 0.1713, 0.3512, 0.2481, 0.5272, 0.2484, 0.3161], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0304, 0.0414, 0.0418, 0.0355, 0.0408, 0.0317, 0.0374], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:54:25,815 INFO [finetune.py:976] (2/7) Epoch 6, batch 2100, loss[loss=0.189, simple_loss=0.2685, pruned_loss=0.0547, over 4908.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2643, pruned_loss=0.07095, over 955941.85 frames. ], batch size: 36, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:54:40,395 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7561, 1.5896, 2.0116, 2.1377, 1.9145, 1.6699, 1.7500, 1.8513], + device='cuda:2'), covar=tensor([0.9602, 1.2304, 1.4271, 1.3701, 1.0967, 1.7570, 1.8005, 1.3779], + device='cuda:2'), in_proj_covar=tensor([0.0416, 0.0444, 0.0530, 0.0550, 0.0445, 0.0466, 0.0479, 0.0478], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:54:54,243 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30781.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:54:59,006 INFO [finetune.py:976] (2/7) Epoch 6, batch 2150, loss[loss=0.2078, simple_loss=0.2821, pruned_loss=0.06668, over 4856.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2688, pruned_loss=0.07286, over 957058.79 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:55:10,624 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-04-26 17:55:17,730 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0481, 2.4651, 1.1335, 1.3390, 1.9932, 1.1768, 3.3146, 1.7525], + device='cuda:2'), covar=tensor([0.0706, 0.0622, 0.0761, 0.1364, 0.0490, 0.1076, 0.0260, 0.0614], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 17:55:40,195 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) Epoch 6, batch 2200, loss[loss=0.1732, simple_loss=0.2356, pruned_loss=0.05537, over 4765.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.27, pruned_loss=0.07328, over 956022.18 frames. ], batch size: 26, lr: 3.91e-03, grad_scale: 16.0 +2023-04-26 17:56:23,796 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30865.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:56:50,921 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1710, 1.3876, 1.5216, 1.6533, 1.5488, 1.7348, 1.6015, 1.6182], + device='cuda:2'), covar=tensor([0.6650, 0.9037, 0.7882, 0.7534, 0.8876, 1.2756, 0.9236, 0.8183], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0394, 0.0316, 0.0326, 0.0345, 0.0410, 0.0374, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:56:51,897 INFO [finetune.py:976] (2/7) Epoch 6, batch 2250, loss[loss=0.1911, simple_loss=0.2607, pruned_loss=0.06074, over 4865.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.2713, pruned_loss=0.07313, over 957952.41 frames. ], batch size: 34, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:57:11,900 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30903.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:57:22,865 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30913.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:57:35,496 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30925.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:57:42,533 INFO [optim.py:369] (2/7) 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 17:57:46,820 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 2300, loss[loss=0.1861, simple_loss=0.2269, pruned_loss=0.07269, over 4087.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2703, pruned_loss=0.07191, over 957467.97 frames. ], batch size: 18, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:58:07,546 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30964.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:08,764 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1374, 0.6903, 0.9673, 0.7194, 1.2230, 0.9828, 0.7240, 0.9796], + device='cuda:2'), covar=tensor([0.1756, 0.1541, 0.1893, 0.1557, 0.1000, 0.1401, 0.1871, 0.1945], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0329, 0.0348, 0.0304, 0.0339, 0.0334, 0.0307, 0.0349], + device='cuda:2'), 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:2') +2023-04-26 17:58:21,385 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30986.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:23,529 INFO [finetune.py:976] (2/7) Epoch 6, batch 2350, loss[loss=0.2187, simple_loss=0.2794, pruned_loss=0.07899, over 4867.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2684, pruned_loss=0.07148, over 955126.52 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:58:28,696 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30995.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:35,898 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.2728, 3.2263, 2.4506, 3.8395, 3.2810, 3.3212, 1.3949, 3.2403], + device='cuda:2'), covar=tensor([0.2033, 0.1445, 0.3382, 0.2503, 0.2948, 0.2237, 0.6110, 0.2709], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0220, 0.0255, 0.0314, 0.0303, 0.0255, 0.0275, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 17:58:39,470 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31011.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 17:58:39,516 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31011.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 17:58:42,544 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31016.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:58:49,763 INFO [optim.py:369] (2/7) 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 17:58:56,921 INFO [finetune.py:976] (2/7) Epoch 6, batch 2400, loss[loss=0.1915, simple_loss=0.2594, pruned_loss=0.0618, over 4869.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2673, pruned_loss=0.07202, over 955431.20 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:59:14,973 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31064.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 17:59:19,955 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31072.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 17:59:21,175 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1397, 0.6820, 0.9461, 0.7009, 1.2597, 0.9769, 0.7845, 0.9897], + device='cuda:2'), covar=tensor([0.1604, 0.1630, 0.1980, 0.1545, 0.0987, 0.1415, 0.1740, 0.2043], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0332, 0.0350, 0.0306, 0.0342, 0.0336, 0.0309, 0.0351], + device='cuda:2'), out_proj_covar=tensor([6.6609e-05, 7.0824e-05, 7.5933e-05, 6.3765e-05, 7.2191e-05, 7.2751e-05, + 6.7022e-05, 7.5698e-05], device='cuda:2') +2023-04-26 17:59:30,874 INFO [finetune.py:976] (2/7) Epoch 6, batch 2450, loss[loss=0.2552, simple_loss=0.3129, pruned_loss=0.09877, over 4050.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2661, pruned_loss=0.07222, over 954718.43 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 17:59:31,651 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.84 vs. limit=5.0 +2023-04-26 17:59:52,810 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4748, 2.3932, 2.6742, 3.0225, 2.8341, 2.3527, 2.0879, 2.5604], + device='cuda:2'), covar=tensor([0.0828, 0.0977, 0.0533, 0.0550, 0.0599, 0.0859, 0.0852, 0.0547], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0205, 0.0179, 0.0176, 0.0177, 0.0192, 0.0164, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 17:59:56,081 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-26 17:59:57,657 INFO [optim.py:369] (2/7) 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 17:59:58,459 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6030, 1.2551, 1.6384, 2.0283, 1.7947, 1.6098, 1.6455, 1.7244], + device='cuda:2'), covar=tensor([0.8209, 1.0761, 1.1045, 1.2194, 0.9364, 1.3155, 1.3207, 1.0765], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0439, 0.0523, 0.0543, 0.0439, 0.0461, 0.0474, 0.0472], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:00:04,429 INFO [finetune.py:976] (2/7) Epoch 6, batch 2500, loss[loss=0.1974, simple_loss=0.266, pruned_loss=0.06443, over 4779.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2673, pruned_loss=0.073, over 953609.35 frames. ], batch size: 28, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:00:12,103 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 18:00:37,627 INFO [finetune.py:976] (2/7) Epoch 6, batch 2550, loss[loss=0.2385, simple_loss=0.3034, pruned_loss=0.0868, over 4834.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.272, pruned_loss=0.07457, over 952136.18 frames. ], batch size: 47, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:00:50,234 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5826, 2.2214, 1.5712, 1.4560, 1.1992, 1.2226, 1.6150, 1.1457], + device='cuda:2'), covar=tensor([0.1969, 0.1511, 0.1844, 0.2065, 0.2706, 0.2223, 0.1266, 0.2346], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0220, 0.0177, 0.0207, 0.0212, 0.0186, 0.0168, 0.0193], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:01:00,988 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7098, 1.5955, 0.8917, 1.4093, 1.6355, 1.5623, 1.4599, 1.4798], + device='cuda:2'), covar=tensor([0.0540, 0.0416, 0.0410, 0.0600, 0.0308, 0.0572, 0.0572, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0024, 0.0030, 0.0021, 0.0030, 0.0030, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 18:01:09,724 INFO [optim.py:369] (2/7) 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 18:01:22,514 INFO [finetune.py:976] (2/7) Epoch 6, batch 2600, loss[loss=0.2404, simple_loss=0.3076, pruned_loss=0.08655, over 4813.00 frames. ], tot_loss[loss=0.2104, simple_loss=0.2728, pruned_loss=0.07399, over 953494.61 frames. ], batch size: 40, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:01:51,862 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31259.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:02:14,135 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31281.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:02:24,318 INFO [finetune.py:976] (2/7) Epoch 6, batch 2650, loss[loss=0.1575, simple_loss=0.2309, pruned_loss=0.04207, over 4755.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2737, pruned_loss=0.07408, over 953616.26 frames. ], batch size: 26, lr: 3.91e-03, grad_scale: 32.0 +2023-04-26 18:02:24,978 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31290.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:02:55,886 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31311.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:03:18,343 INFO [optim.py:369] (2/7) 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 18:03:30,923 INFO [finetune.py:976] (2/7) Epoch 6, batch 2700, loss[loss=0.1483, simple_loss=0.2179, pruned_loss=0.03933, over 4703.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2724, pruned_loss=0.07322, over 955279.24 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:04:01,825 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31359.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:04:14,062 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31367.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:04:23,834 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6605, 2.2118, 1.7839, 2.1219, 1.6548, 1.8281, 1.8051, 1.5411], + device='cuda:2'), covar=tensor([0.2336, 0.1388, 0.1093, 0.1428, 0.3376, 0.1431, 0.2075, 0.2887], + device='cuda:2'), in_proj_covar=tensor([0.0303, 0.0324, 0.0234, 0.0295, 0.0316, 0.0277, 0.0263, 0.0288], + device='cuda:2'), out_proj_covar=tensor([1.2307e-04, 1.3179e-04, 9.5085e-05, 1.1827e-04, 1.2997e-04, 1.1188e-04, + 1.0802e-04, 1.1573e-04], device='cuda:2') +2023-04-26 18:04:45,001 INFO [finetune.py:976] (2/7) Epoch 6, batch 2750, loss[loss=0.151, simple_loss=0.2195, pruned_loss=0.04128, over 4866.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2696, pruned_loss=0.07205, over 956064.85 frames. ], batch size: 31, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:05:10,771 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31409.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:05:28,796 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 2800, loss[loss=0.178, simple_loss=0.2331, pruned_loss=0.06148, over 4761.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2654, pruned_loss=0.07075, over 956618.33 frames. ], batch size: 26, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:05:42,711 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6339, 2.2001, 1.5806, 1.5504, 1.2191, 1.2626, 1.6223, 1.1497], + device='cuda:2'), covar=tensor([0.1792, 0.1541, 0.1774, 0.2057, 0.2752, 0.2164, 0.1300, 0.2316], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0220, 0.0176, 0.0206, 0.0212, 0.0186, 0.0168, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:05:57,413 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31470.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:06:09,974 INFO [finetune.py:976] (2/7) Epoch 6, batch 2850, loss[loss=0.1457, simple_loss=0.2152, pruned_loss=0.03807, over 4762.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2648, pruned_loss=0.07095, over 956436.16 frames. ], batch size: 26, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:06:11,328 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31491.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 18:06:18,070 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4753, 1.4821, 0.6541, 1.1906, 1.4401, 1.3640, 1.2787, 1.2861], + device='cuda:2'), covar=tensor([0.0593, 0.0402, 0.0474, 0.0622, 0.0346, 0.0577, 0.0545, 0.0629], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0027, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0049, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 18:06:36,454 INFO [optim.py:369] (2/7) 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 18:06:40,364 INFO [scaling.py:679] (2/7) 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] (2/7) Epoch 6, batch 2900, loss[loss=0.2258, simple_loss=0.2943, pruned_loss=0.07862, over 4835.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2678, pruned_loss=0.0721, over 955726.60 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:06:47,517 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4174, 1.2702, 1.6535, 1.6177, 1.3251, 1.0769, 1.3947, 1.0168], + device='cuda:2'), covar=tensor([0.0654, 0.0694, 0.0539, 0.0570, 0.0845, 0.1165, 0.0627, 0.0762], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0075, 0.0074, 0.0068, 0.0078, 0.0095, 0.0081, 0.0077], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:06:51,763 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31552.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:06:56,008 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31559.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:06,435 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6514, 2.2498, 1.6852, 1.5036, 1.2138, 1.2685, 1.8232, 1.1919], + device='cuda:2'), covar=tensor([0.1848, 0.1590, 0.1805, 0.2120, 0.2777, 0.2278, 0.1197, 0.2307], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0220, 0.0176, 0.0205, 0.0211, 0.0186, 0.0168, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:07:11,690 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31581.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:16,958 INFO [finetune.py:976] (2/7) Epoch 6, batch 2950, loss[loss=0.2125, simple_loss=0.278, pruned_loss=0.07352, over 4901.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2723, pruned_loss=0.07362, over 956905.13 frames. ], batch size: 37, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:07:17,676 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31590.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:28,067 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31607.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:42,550 INFO [optim.py:369] (2/7) 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 18:07:43,219 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31629.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:07:49,138 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 3000, loss[loss=0.1925, simple_loss=0.2728, pruned_loss=0.05613, over 4918.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2746, pruned_loss=0.0748, over 957900.84 frames. ], batch size: 42, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:07:49,694 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 18:07:54,071 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4126, 1.2287, 1.5859, 1.5347, 1.3140, 1.1215, 1.3052, 0.9842], + device='cuda:2'), covar=tensor([0.0673, 0.0816, 0.0629, 0.0665, 0.0832, 0.1179, 0.0707, 0.0777], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0075, 0.0074, 0.0067, 0.0078, 0.0095, 0.0081, 0.0077], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:08:00,214 INFO [finetune.py:1010] (2/7) Epoch 6, validation: loss=0.1565, simple_loss=0.2301, pruned_loss=0.04144, over 2265189.00 frames. +2023-04-26 18:08:00,215 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 18:08:17,852 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31667.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:08:48,665 INFO [finetune.py:976] (2/7) Epoch 6, batch 3050, loss[loss=0.179, simple_loss=0.2541, pruned_loss=0.05193, over 4747.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2753, pruned_loss=0.0746, over 957173.20 frames. ], batch size: 59, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:09:09,598 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6128, 2.0672, 1.5951, 1.4352, 1.3253, 1.3244, 1.6962, 1.2489], + device='cuda:2'), covar=tensor([0.1557, 0.1431, 0.1712, 0.2003, 0.2426, 0.2058, 0.1116, 0.2098], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0220, 0.0176, 0.0206, 0.0211, 0.0187, 0.0168, 0.0193], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:09:10,798 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31704.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:09:20,317 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8845, 1.3498, 1.6385, 1.6425, 1.5925, 1.2815, 0.7443, 1.2512], + device='cuda:2'), covar=tensor([0.3842, 0.4364, 0.2056, 0.3029, 0.3546, 0.3463, 0.5410, 0.3361], + device='cuda:2'), in_proj_covar=tensor([0.0277, 0.0256, 0.0220, 0.0327, 0.0216, 0.0230, 0.0242, 0.0192], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:09:23,337 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31715.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 18:09:40,854 INFO [optim.py:369] (2/7) 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 18:09:53,205 INFO [finetune.py:976] (2/7) Epoch 6, batch 3100, loss[loss=0.1997, simple_loss=0.2733, pruned_loss=0.06303, over 4902.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.2735, pruned_loss=0.07383, over 955773.91 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:10:25,840 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:10:25,887 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:10:56,967 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31788.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:10:57,966 INFO [finetune.py:976] (2/7) Epoch 6, batch 3150, loss[loss=0.173, simple_loss=0.2364, pruned_loss=0.05475, over 4930.00 frames. ], tot_loss[loss=0.2064, simple_loss=0.2692, pruned_loss=0.07179, over 956347.10 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:11:42,881 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:12:02,481 INFO [finetune.py:976] (2/7) Epoch 6, batch 3200, loss[loss=0.2337, simple_loss=0.2775, pruned_loss=0.09493, over 4714.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2658, pruned_loss=0.07065, over 955431.23 frames. ], batch size: 54, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:12:14,985 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31847.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 18:12:22,017 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:12:57,059 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31878.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:13:06,249 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31884.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:13:09,140 INFO [finetune.py:976] (2/7) Epoch 6, batch 3250, loss[loss=0.259, simple_loss=0.3011, pruned_loss=0.1085, over 4727.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.266, pruned_loss=0.07113, over 954891.74 frames. ], batch size: 59, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:13:39,530 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:14:10,187 INFO [finetune.py:976] (2/7) Epoch 6, batch 3300, loss[loss=0.2107, simple_loss=0.2712, pruned_loss=0.07514, over 4754.00 frames. ], tot_loss[loss=0.208, simple_loss=0.2708, pruned_loss=0.07263, over 956513.98 frames. ], batch size: 27, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:14:10,322 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31939.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:14:18,826 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31950.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:14:33,034 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.63 vs. limit=2.0 +2023-04-26 18:14:35,918 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31976.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:14:43,707 INFO [finetune.py:976] (2/7) Epoch 6, batch 3350, loss[loss=0.2227, simple_loss=0.2824, pruned_loss=0.08149, over 4836.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.2731, pruned_loss=0.07405, over 955618.85 frames. ], batch size: 49, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:15:00,873 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=32011.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:15:04,335 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6331, 1.6008, 3.7979, 3.5858, 3.4207, 3.6380, 3.6211, 3.3442], + device='cuda:2'), covar=tensor([0.6130, 0.4527, 0.1177, 0.1696, 0.1057, 0.1752, 0.1128, 0.1455], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0304, 0.0412, 0.0415, 0.0351, 0.0407, 0.0314, 0.0371], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:15:12,029 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 3400, loss[loss=0.1911, simple_loss=0.2632, pruned_loss=0.05947, over 4770.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2727, pruned_loss=0.07392, over 954440.63 frames. ], batch size: 28, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:15:55,465 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32060.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:16:04,739 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 3450, loss[loss=0.1612, simple_loss=0.2335, pruned_loss=0.04449, over 4818.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2704, pruned_loss=0.07225, over 953318.93 frames. ], batch size: 30, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:16:30,211 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-26 18:16:37,413 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:16:47,573 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3596, 2.4011, 2.1917, 2.1759, 2.6731, 2.2103, 3.4773, 2.0062], + device='cuda:2'), covar=tensor([0.4597, 0.2403, 0.4800, 0.4247, 0.2181, 0.2951, 0.1705, 0.4253], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0355, 0.0435, 0.0366, 0.0392, 0.0383, 0.0388, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:16:54,107 INFO [finetune.py:976] (2/7) Epoch 6, batch 3500, loss[loss=0.1812, simple_loss=0.2479, pruned_loss=0.05721, over 4694.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2687, pruned_loss=0.07139, over 953220.26 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:16:57,313 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32144.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:16:59,132 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32147.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 18:17:21,485 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 3550, loss[loss=0.1906, simple_loss=0.249, pruned_loss=0.06604, over 4757.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2662, pruned_loss=0.07087, over 954316.09 frames. ], batch size: 54, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:17:43,431 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32195.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 18:18:29,258 INFO [optim.py:369] (2/7) 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 18:18:39,220 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32234.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:18:47,839 INFO [finetune.py:976] (2/7) Epoch 6, batch 3600, loss[loss=0.1632, simple_loss=0.2359, pruned_loss=0.04528, over 4754.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2629, pruned_loss=0.0695, over 954406.23 frames. ], batch size: 27, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:19:03,024 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6555, 2.0130, 1.7183, 1.8800, 1.4599, 1.6751, 1.6769, 1.4536], + device='cuda:2'), covar=tensor([0.1799, 0.1229, 0.0936, 0.1155, 0.3183, 0.1227, 0.1884, 0.2389], + device='cuda:2'), in_proj_covar=tensor([0.0303, 0.0324, 0.0235, 0.0296, 0.0319, 0.0278, 0.0264, 0.0288], + device='cuda:2'), out_proj_covar=tensor([1.2300e-04, 1.3140e-04, 9.5436e-05, 1.1887e-04, 1.3123e-04, 1.1243e-04, + 1.0868e-04, 1.1578e-04], device='cuda:2') +2023-04-26 18:19:25,147 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 3650, loss[loss=0.1984, simple_loss=0.275, pruned_loss=0.06093, over 4823.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2662, pruned_loss=0.07113, over 956445.84 frames. ], batch size: 40, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:20:05,681 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5139, 1.0302, 1.2843, 1.2094, 1.7193, 1.3962, 1.1355, 1.2791], + device='cuda:2'), covar=tensor([0.1574, 0.1706, 0.2180, 0.1561, 0.0913, 0.1451, 0.1858, 0.2252], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0335, 0.0353, 0.0306, 0.0343, 0.0337, 0.0311, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.6606e-05, 7.1590e-05, 7.6634e-05, 6.3724e-05, 7.2214e-05, 7.2894e-05, + 6.7309e-05, 7.6344e-05], device='cuda:2') +2023-04-26 18:20:16,842 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32306.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:20:19,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5313, 3.4689, 2.6937, 4.1443, 3.5381, 3.5066, 1.7306, 3.5913], + device='cuda:2'), covar=tensor([0.1442, 0.1378, 0.3599, 0.1314, 0.2881, 0.1617, 0.5118, 0.2138], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0219, 0.0254, 0.0310, 0.0300, 0.0253, 0.0275, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:20:43,388 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 3700, loss[loss=0.1985, simple_loss=0.2702, pruned_loss=0.06335, over 4782.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2688, pruned_loss=0.07158, over 956742.21 frames. ], batch size: 51, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:21:06,642 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 18:21:13,661 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1710, 1.4834, 1.2259, 1.6222, 1.4689, 1.8702, 1.2966, 3.4587], + device='cuda:2'), covar=tensor([0.0764, 0.1002, 0.0957, 0.1385, 0.0800, 0.0699, 0.0996, 0.0223], + device='cuda:2'), in_proj_covar=tensor([0.0040, 0.0040, 0.0042, 0.0045, 0.0041, 0.0040, 0.0040, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 18:21:26,772 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32360.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:21:35,823 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5274, 1.2982, 4.1370, 3.8104, 3.6173, 3.9407, 3.9158, 3.5968], + device='cuda:2'), covar=tensor([0.6801, 0.5936, 0.0935, 0.1760, 0.1111, 0.1520, 0.1212, 0.1468], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0307, 0.0414, 0.0418, 0.0353, 0.0411, 0.0316, 0.0373], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:22:04,003 INFO [finetune.py:976] (2/7) Epoch 6, batch 3750, loss[loss=0.2412, simple_loss=0.3167, pruned_loss=0.0829, over 4852.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2708, pruned_loss=0.07224, over 955377.32 frames. ], batch size: 49, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:22:21,625 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32408.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:22:30,669 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5707, 3.4360, 2.9956, 4.1186, 3.2956, 3.6030, 1.8897, 3.6194], + device='cuda:2'), covar=tensor([0.1788, 0.1438, 0.3792, 0.1209, 0.2533, 0.1753, 0.4404, 0.2119], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0220, 0.0256, 0.0312, 0.0303, 0.0255, 0.0277, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:22:35,218 INFO [optim.py:369] (2/7) 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 18:22:39,939 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4826, 3.6761, 0.9140, 2.0308, 2.0070, 2.5893, 2.2135, 0.9793], + device='cuda:2'), covar=tensor([0.1393, 0.1043, 0.1997, 0.1339, 0.1104, 0.1152, 0.1500, 0.2333], + device='cuda:2'), in_proj_covar=tensor([0.0123, 0.0262, 0.0147, 0.0128, 0.0139, 0.0161, 0.0124, 0.0127], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:22:43,808 INFO [finetune.py:976] (2/7) Epoch 6, batch 3800, loss[loss=0.2052, simple_loss=0.2676, pruned_loss=0.07143, over 4898.00 frames. ], tot_loss[loss=0.2085, simple_loss=0.2717, pruned_loss=0.07267, over 953896.18 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:22:53,205 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32444.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:23:07,313 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-26 18:23:21,551 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32479.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:23:22,754 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 18:23:33,784 INFO [finetune.py:976] (2/7) Epoch 6, batch 3850, loss[loss=0.1734, simple_loss=0.2318, pruned_loss=0.05751, over 4926.00 frames. ], tot_loss[loss=0.2063, simple_loss=0.2693, pruned_loss=0.07172, over 953830.74 frames. ], batch size: 33, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:23:41,606 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32492.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:24:18,832 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:24:24,623 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32534.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:24:28,026 INFO [finetune.py:976] (2/7) Epoch 6, batch 3900, loss[loss=0.2077, simple_loss=0.2652, pruned_loss=0.07515, over 4838.00 frames. ], tot_loss[loss=0.2036, simple_loss=0.2664, pruned_loss=0.07039, over 954493.02 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 32.0 +2023-04-26 18:25:09,838 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32571.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:25:14,612 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1530, 2.9557, 2.4056, 2.5754, 2.0372, 2.3774, 2.4669, 2.1065], + device='cuda:2'), covar=tensor([0.2426, 0.1368, 0.0826, 0.1422, 0.3134, 0.1421, 0.2182, 0.2656], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0330, 0.0238, 0.0302, 0.0324, 0.0283, 0.0269, 0.0293], + device='cuda:2'), out_proj_covar=tensor([1.2483e-04, 1.3395e-04, 9.6689e-05, 1.2127e-04, 1.3350e-04, 1.1441e-04, + 1.1058e-04, 1.1792e-04], device='cuda:2') +2023-04-26 18:25:22,276 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32582.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:25:33,444 INFO [finetune.py:976] (2/7) Epoch 6, batch 3950, loss[loss=0.1996, simple_loss=0.2594, pruned_loss=0.06988, over 4741.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2627, pruned_loss=0.06885, over 953562.60 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:25:47,381 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8328, 1.9023, 1.7396, 1.4652, 2.0393, 1.5892, 2.5458, 1.4843], + device='cuda:2'), covar=tensor([0.3970, 0.1864, 0.5179, 0.2995, 0.1830, 0.2635, 0.1640, 0.4776], + device='cuda:2'), in_proj_covar=tensor([0.0349, 0.0357, 0.0437, 0.0367, 0.0396, 0.0385, 0.0387, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:25:50,960 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32606.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:25:58,783 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:26:22,977 INFO [finetune.py:976] (2/7) Epoch 6, batch 4000, loss[loss=0.191, simple_loss=0.2603, pruned_loss=0.06085, over 4910.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2639, pruned_loss=0.06999, over 954619.98 frames. ], batch size: 37, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:26:51,941 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:27:19,403 INFO [finetune.py:976] (2/7) Epoch 6, batch 4050, loss[loss=0.1912, simple_loss=0.2401, pruned_loss=0.07113, over 4711.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2681, pruned_loss=0.07136, over 955945.94 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:27:46,363 INFO [optim.py:369] (2/7) 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 18:27:52,926 INFO [finetune.py:976] (2/7) Epoch 6, batch 4100, loss[loss=0.2374, simple_loss=0.3011, pruned_loss=0.08688, over 4904.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2689, pruned_loss=0.07114, over 952521.85 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:28:21,659 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 18:28:26,745 INFO [finetune.py:976] (2/7) Epoch 6, batch 4150, loss[loss=0.1948, simple_loss=0.2683, pruned_loss=0.06067, over 4916.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2708, pruned_loss=0.0724, over 951507.65 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:29:00,934 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 18:29:05,515 INFO [optim.py:369] (2/7) 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 18:29:12,120 INFO [finetune.py:976] (2/7) Epoch 6, batch 4200, loss[loss=0.1915, simple_loss=0.2483, pruned_loss=0.06737, over 4722.00 frames. ], tot_loss[loss=0.207, simple_loss=0.271, pruned_loss=0.0715, over 953503.81 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:29:14,097 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0715, 0.6968, 0.9320, 0.7396, 1.2002, 0.9560, 0.7557, 0.9475], + device='cuda:2'), covar=tensor([0.1785, 0.1711, 0.2110, 0.1647, 0.0936, 0.1479, 0.2092, 0.2267], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0328, 0.0348, 0.0301, 0.0338, 0.0330, 0.0305, 0.0348], + device='cuda:2'), out_proj_covar=tensor([6.5529e-05, 6.9986e-05, 7.5375e-05, 6.2666e-05, 7.1118e-05, 7.1401e-05, + 6.6000e-05, 7.5008e-05], device='cuda:2') +2023-04-26 18:29:45,429 INFO [finetune.py:976] (2/7) Epoch 6, batch 4250, loss[loss=0.197, simple_loss=0.2639, pruned_loss=0.06501, over 4934.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2683, pruned_loss=0.07071, over 954280.14 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:29:56,118 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 18:29:59,076 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-26 18:30:13,035 INFO [optim.py:369] (2/7) 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 18:30:19,128 INFO [finetune.py:976] (2/7) Epoch 6, batch 4300, loss[loss=0.2479, simple_loss=0.2876, pruned_loss=0.1041, over 3933.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2654, pruned_loss=0.07009, over 954024.03 frames. ], batch size: 17, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:30:24,606 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3622, 1.4840, 1.5513, 2.2116, 2.3094, 1.9917, 1.8738, 1.6846], + device='cuda:2'), covar=tensor([0.1845, 0.2005, 0.2300, 0.1762, 0.1528, 0.2212, 0.2738, 0.2134], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0328, 0.0349, 0.0302, 0.0338, 0.0331, 0.0306, 0.0349], + device='cuda:2'), out_proj_covar=tensor([6.5523e-05, 7.0056e-05, 7.5554e-05, 6.2816e-05, 7.1318e-05, 7.1578e-05, + 6.6222e-05, 7.5218e-05], device='cuda:2') +2023-04-26 18:30:47,973 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=32966.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:31:19,599 INFO [finetune.py:976] (2/7) Epoch 6, batch 4350, loss[loss=0.2215, simple_loss=0.2843, pruned_loss=0.07938, over 4813.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2631, pruned_loss=0.0695, over 954352.05 frames. ], batch size: 41, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:32:11,704 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1224, 1.4659, 1.3342, 1.7477, 1.4879, 1.7052, 1.3454, 3.1252], + device='cuda:2'), covar=tensor([0.0710, 0.0844, 0.0876, 0.1165, 0.0672, 0.0754, 0.0812, 0.0236], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0040, 0.0042, 0.0045, 0.0041, 0.0040, 0.0040, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 18:32:12,354 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33027.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:32:13,441 INFO [optim.py:369] (2/7) 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 18:32:24,806 INFO [finetune.py:976] (2/7) Epoch 6, batch 4400, loss[loss=0.2119, simple_loss=0.2686, pruned_loss=0.07759, over 4908.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2652, pruned_loss=0.07076, over 953883.64 frames. ], batch size: 35, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:33:06,897 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33069.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:33:17,889 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0034, 2.4610, 1.0555, 1.2724, 1.9727, 1.2489, 3.0589, 1.6475], + device='cuda:2'), covar=tensor([0.0734, 0.0578, 0.0812, 0.1438, 0.0474, 0.1090, 0.0379, 0.0691], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0052, 0.0048, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 18:33:20,941 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8745, 2.4562, 1.0611, 1.1932, 1.9337, 1.1368, 3.1943, 1.4537], + device='cuda:2'), covar=tensor([0.0778, 0.0677, 0.0857, 0.1426, 0.0546, 0.1187, 0.0286, 0.0798], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0052, 0.0049, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 18:33:29,654 INFO [finetune.py:976] (2/7) Epoch 6, batch 4450, loss[loss=0.2173, simple_loss=0.287, pruned_loss=0.0738, over 4924.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2682, pruned_loss=0.07148, over 954619.32 frames. ], batch size: 42, lr: 3.90e-03, grad_scale: 16.0 +2023-04-26 18:33:30,412 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5563, 1.2576, 1.8286, 1.7852, 1.3436, 1.2633, 1.4861, 1.0699], + device='cuda:2'), covar=tensor([0.0727, 0.1173, 0.0535, 0.0962, 0.1094, 0.1279, 0.0940, 0.0931], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0076, 0.0074, 0.0068, 0.0079, 0.0096, 0.0081, 0.0078], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:33:40,426 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7102, 1.9771, 1.7829, 1.9593, 1.7150, 1.9891, 1.9310, 1.8300], + device='cuda:2'), covar=tensor([0.6458, 1.0366, 0.8937, 0.7573, 0.9584, 1.3004, 1.0883, 0.9726], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0391, 0.0316, 0.0325, 0.0342, 0.0407, 0.0371, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:34:12,570 INFO [optim.py:369] (2/7) 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 18:34:13,320 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:34:18,682 INFO [finetune.py:976] (2/7) Epoch 6, batch 4500, loss[loss=0.1703, simple_loss=0.2314, pruned_loss=0.05458, over 4564.00 frames. ], tot_loss[loss=0.2077, simple_loss=0.2702, pruned_loss=0.07259, over 950697.38 frames. ], batch size: 20, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:34:29,021 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33155.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:34:52,374 INFO [finetune.py:976] (2/7) Epoch 6, batch 4550, loss[loss=0.2188, simple_loss=0.285, pruned_loss=0.07634, over 4899.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2706, pruned_loss=0.07211, over 952585.68 frames. ], batch size: 43, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:35:01,151 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3659, 1.6660, 1.1881, 1.0411, 1.0922, 1.0529, 1.1416, 0.9867], + device='cuda:2'), covar=tensor([0.2130, 0.1534, 0.2015, 0.2153, 0.2864, 0.2502, 0.1333, 0.2416], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0218, 0.0175, 0.0205, 0.0210, 0.0185, 0.0166, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 18:35:09,482 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:35:25,829 INFO [finetune.py:976] (2/7) Epoch 6, batch 4600, loss[loss=0.1968, simple_loss=0.2559, pruned_loss=0.06879, over 4819.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2698, pruned_loss=0.07131, over 952906.08 frames. ], batch size: 30, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:35:36,337 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 18:35:38,115 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-26 18:35:59,786 INFO [finetune.py:976] (2/7) Epoch 6, batch 4650, loss[loss=0.1891, simple_loss=0.2391, pruned_loss=0.06955, over 4827.00 frames. ], tot_loss[loss=0.2037, simple_loss=0.267, pruned_loss=0.0702, over 954306.54 frames. ], batch size: 39, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:36:07,859 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-04-26 18:36:20,289 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33322.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:36:25,472 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 6, batch 4700, loss[loss=0.1709, simple_loss=0.2325, pruned_loss=0.05468, over 4821.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2638, pruned_loss=0.06911, over 952793.34 frames. ], batch size: 30, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:37:20,897 INFO [finetune.py:976] (2/7) Epoch 6, batch 4750, loss[loss=0.1497, simple_loss=0.2032, pruned_loss=0.04808, over 3995.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2619, pruned_loss=0.06853, over 953633.49 frames. ], batch size: 17, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:38:06,051 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33425.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:38:14,212 INFO [optim.py:369] (2/7) 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 18:38:15,624 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-04-26 18:38:27,282 INFO [finetune.py:976] (2/7) Epoch 6, batch 4800, loss[loss=0.2545, simple_loss=0.3065, pruned_loss=0.1013, over 4819.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2643, pruned_loss=0.06978, over 954119.55 frames. ], batch size: 40, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:38:36,937 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-26 18:38:39,141 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7174, 2.0676, 1.1162, 1.4424, 2.1513, 1.6454, 1.5711, 1.5580], + device='cuda:2'), covar=tensor([0.0575, 0.0386, 0.0359, 0.0581, 0.0271, 0.0577, 0.0552, 0.0681], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 18:39:01,120 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33472.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:39:05,327 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6194, 2.3192, 1.5973, 1.4561, 1.2404, 1.2350, 1.5724, 1.1664], + device='cuda:2'), covar=tensor([0.1944, 0.1389, 0.1806, 0.2179, 0.2726, 0.2148, 0.1305, 0.2309], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0219, 0.0175, 0.0206, 0.0211, 0.0186, 0.0166, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 18:39:12,310 INFO [finetune.py:976] (2/7) Epoch 6, batch 4850, loss[loss=0.2023, simple_loss=0.2653, pruned_loss=0.06964, over 4759.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2682, pruned_loss=0.07119, over 955720.60 frames. ], batch size: 54, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:39:16,594 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8358, 1.6971, 1.9902, 2.2070, 1.9545, 1.7168, 1.7955, 1.8730], + device='cuda:2'), covar=tensor([0.8824, 1.2268, 1.4266, 1.2921, 1.0517, 1.7005, 1.7409, 1.3580], + device='cuda:2'), in_proj_covar=tensor([0.0412, 0.0438, 0.0523, 0.0544, 0.0441, 0.0461, 0.0473, 0.0472], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:39:26,017 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7143, 1.5351, 1.7082, 2.0321, 2.0425, 1.5985, 1.3104, 1.8285], + device='cuda:2'), covar=tensor([0.0801, 0.1115, 0.0722, 0.0594, 0.0574, 0.0911, 0.0920, 0.0552], + device='cuda:2'), in_proj_covar=tensor([0.0202, 0.0206, 0.0180, 0.0179, 0.0180, 0.0196, 0.0165, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:39:27,145 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33511.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:39:38,471 INFO [optim.py:369] (2/7) 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 18:39:41,039 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 4900, loss[loss=0.1972, simple_loss=0.2699, pruned_loss=0.06228, over 4825.00 frames. ], tot_loss[loss=0.2077, simple_loss=0.2704, pruned_loss=0.07246, over 954021.92 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:40:15,607 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-04-26 18:40:18,320 INFO [finetune.py:976] (2/7) Epoch 6, batch 4950, loss[loss=0.206, simple_loss=0.2728, pruned_loss=0.06959, over 4722.00 frames. ], tot_loss[loss=0.208, simple_loss=0.2711, pruned_loss=0.07243, over 953450.62 frames. ], batch size: 59, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:40:40,488 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:40:48,366 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-26 18:40:48,905 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8815, 2.3735, 2.0043, 2.2133, 1.7780, 1.9216, 1.8961, 1.6224], + device='cuda:2'), covar=tensor([0.2034, 0.1127, 0.0880, 0.1259, 0.3271, 0.1224, 0.1813, 0.2720], + device='cuda:2'), in_proj_covar=tensor([0.0302, 0.0320, 0.0234, 0.0293, 0.0314, 0.0275, 0.0262, 0.0286], + device='cuda:2'), out_proj_covar=tensor([1.2258e-04, 1.2985e-04, 9.4741e-05, 1.1776e-04, 1.2921e-04, 1.1103e-04, + 1.0754e-04, 1.1537e-04], device='cuda:2') +2023-04-26 18:40:50,727 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3742, 2.4758, 1.9703, 2.2428, 2.4509, 1.9965, 3.2496, 1.6337], + device='cuda:2'), covar=tensor([0.4568, 0.2288, 0.5295, 0.3814, 0.2557, 0.3084, 0.1762, 0.5399], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0354, 0.0437, 0.0363, 0.0394, 0.0384, 0.0385, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:40:51,672 INFO [finetune.py:976] (2/7) Epoch 6, batch 5000, loss[loss=0.1672, simple_loss=0.2313, pruned_loss=0.05152, over 4856.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2684, pruned_loss=0.07124, over 953312.08 frames. ], batch size: 44, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:41:01,563 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-04-26 18:41:17,890 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2760, 3.4251, 0.7826, 2.0531, 1.8855, 2.4164, 2.0150, 0.9767], + device='cuda:2'), covar=tensor([0.1527, 0.0850, 0.2070, 0.1208, 0.1121, 0.1091, 0.1386, 0.2104], + device='cuda:2'), in_proj_covar=tensor([0.0120, 0.0258, 0.0144, 0.0126, 0.0137, 0.0158, 0.0122, 0.0125], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:41:24,495 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33670.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:41:28,789 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33677.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:41:36,413 INFO [finetune.py:976] (2/7) Epoch 6, batch 5050, loss[loss=0.1797, simple_loss=0.2532, pruned_loss=0.05304, over 4793.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2657, pruned_loss=0.07058, over 953750.29 frames. ], batch size: 29, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:42:01,210 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:42:04,237 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2490, 2.0985, 1.8074, 1.8314, 2.1466, 1.7637, 2.4925, 1.5449], + device='cuda:2'), covar=tensor([0.3258, 0.1330, 0.3697, 0.2606, 0.1598, 0.2061, 0.1627, 0.3863], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0352, 0.0435, 0.0362, 0.0392, 0.0383, 0.0383, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:42:14,977 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 6, batch 5100, loss[loss=0.1796, simple_loss=0.2446, pruned_loss=0.05727, over 4698.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2616, pruned_loss=0.06837, over 953534.22 frames. ], batch size: 23, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:42:51,823 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0177, 2.3933, 1.0084, 1.2201, 1.7430, 1.2056, 3.0126, 1.4232], + device='cuda:2'), covar=tensor([0.0689, 0.0605, 0.0844, 0.1290, 0.0515, 0.0987, 0.0314, 0.0701], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0070, 0.0052, 0.0048, 0.0053, 0.0054, 0.0081, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 18:43:00,671 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33773.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:43:11,292 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 18:43:21,823 INFO [finetune.py:976] (2/7) Epoch 6, batch 5150, loss[loss=0.2193, simple_loss=0.2814, pruned_loss=0.07857, over 4848.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.261, pruned_loss=0.06801, over 953647.27 frames. ], batch size: 49, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:43:42,890 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33811.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:43:54,702 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:44:07,685 INFO [finetune.py:976] (2/7) Epoch 6, batch 5200, loss[loss=0.2376, simple_loss=0.3012, pruned_loss=0.08699, over 4865.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2674, pruned_loss=0.07111, over 953338.28 frames. ], batch size: 31, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:44:38,043 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:45:14,782 INFO [finetune.py:976] (2/7) Epoch 6, batch 5250, loss[loss=0.2305, simple_loss=0.2837, pruned_loss=0.08862, over 4155.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2699, pruned_loss=0.07197, over 950716.38 frames. ], batch size: 65, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:45:14,932 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2541, 1.5920, 2.0801, 2.5307, 2.0883, 1.5485, 1.3137, 1.9431], + device='cuda:2'), covar=tensor([0.3893, 0.4543, 0.2036, 0.3306, 0.3629, 0.3511, 0.5474, 0.3033], + device='cuda:2'), in_proj_covar=tensor([0.0278, 0.0257, 0.0220, 0.0326, 0.0216, 0.0230, 0.0241, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:45:25,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8388, 2.0494, 1.8743, 2.1346, 1.8593, 2.0774, 2.0687, 1.9595], + device='cuda:2'), covar=tensor([0.5907, 0.9853, 0.8040, 0.7015, 0.8692, 1.2342, 1.0116, 0.9286], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0392, 0.0318, 0.0327, 0.0344, 0.0410, 0.0373, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:45:37,129 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7141, 2.5092, 1.6246, 1.7343, 1.3214, 1.2841, 1.7759, 1.2156], + device='cuda:2'), covar=tensor([0.2128, 0.1756, 0.2262, 0.2378, 0.3029, 0.2836, 0.1427, 0.2546], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0218, 0.0174, 0.0205, 0.0210, 0.0186, 0.0166, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 18:45:58,333 INFO [optim.py:369] (2/7) 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 18:46:04,411 INFO [finetune.py:976] (2/7) Epoch 6, batch 5300, loss[loss=0.1869, simple_loss=0.2652, pruned_loss=0.05426, over 4802.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2705, pruned_loss=0.07193, over 952596.60 frames. ], batch size: 51, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:46:04,525 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33939.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:46:38,102 INFO [finetune.py:976] (2/7) Epoch 6, batch 5350, loss[loss=0.1685, simple_loss=0.226, pruned_loss=0.05551, over 4753.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2706, pruned_loss=0.07189, over 954531.87 frames. ], batch size: 27, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:46:46,597 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34000.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:46:57,252 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6518, 3.5132, 2.6279, 4.2332, 3.5652, 3.6956, 1.6482, 3.6453], + device='cuda:2'), covar=tensor([0.1672, 0.1452, 0.4116, 0.1489, 0.3158, 0.1934, 0.5741, 0.2453], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0219, 0.0253, 0.0310, 0.0302, 0.0254, 0.0274, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:46:57,335 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3988, 1.6184, 1.5941, 1.7619, 1.6114, 1.7588, 1.7385, 1.6810], + device='cuda:2'), covar=tensor([0.5348, 0.8386, 0.6883, 0.6306, 0.8005, 1.1776, 0.8441, 0.7603], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0393, 0.0318, 0.0328, 0.0345, 0.0411, 0.0373, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:46:58,539 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7057, 1.5901, 1.7816, 2.0953, 2.0847, 1.7367, 1.3616, 1.8205], + device='cuda:2'), covar=tensor([0.0875, 0.1145, 0.0663, 0.0635, 0.0582, 0.0832, 0.0929, 0.0593], + device='cuda:2'), in_proj_covar=tensor([0.0201, 0.0206, 0.0179, 0.0178, 0.0180, 0.0194, 0.0165, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:47:06,985 INFO [optim.py:369] (2/7) 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 18:47:09,492 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34033.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 18:47:13,056 INFO [finetune.py:976] (2/7) Epoch 6, batch 5400, loss[loss=0.223, simple_loss=0.2826, pruned_loss=0.08164, over 4908.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.268, pruned_loss=0.07134, over 955297.50 frames. ], batch size: 37, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:47:16,178 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34044.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:47:24,611 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-04-26 18:47:29,685 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4016, 3.3349, 0.9098, 1.8830, 1.8329, 2.2969, 1.9179, 0.9744], + device='cuda:2'), covar=tensor([0.1560, 0.1079, 0.1913, 0.1280, 0.1118, 0.1084, 0.1507, 0.1922], + device='cuda:2'), in_proj_covar=tensor([0.0121, 0.0259, 0.0144, 0.0127, 0.0138, 0.0159, 0.0123, 0.0126], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:47:46,754 INFO [finetune.py:976] (2/7) Epoch 6, batch 5450, loss[loss=0.1862, simple_loss=0.2464, pruned_loss=0.06302, over 4870.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2654, pruned_loss=0.07043, over 956841.29 frames. ], batch size: 31, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:47:57,065 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34105.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:48:12,447 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:48:25,398 INFO [finetune.py:976] (2/7) Epoch 6, batch 5500, loss[loss=0.185, simple_loss=0.258, pruned_loss=0.05606, over 4847.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2621, pruned_loss=0.0691, over 954904.45 frames. ], batch size: 44, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:48:29,184 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34145.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:48:37,575 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7286, 1.9977, 0.9831, 1.4976, 2.1317, 1.6426, 1.6039, 1.5669], + device='cuda:2'), covar=tensor([0.0529, 0.0358, 0.0380, 0.0574, 0.0270, 0.0528, 0.0516, 0.0608], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0031, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 18:49:10,262 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6813, 2.2807, 1.9456, 2.1778, 1.6358, 1.9415, 1.8843, 1.5881], + device='cuda:2'), covar=tensor([0.2506, 0.1316, 0.0963, 0.1236, 0.3297, 0.1145, 0.1997, 0.2730], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0323, 0.0236, 0.0296, 0.0317, 0.0278, 0.0265, 0.0288], + device='cuda:2'), out_proj_covar=tensor([1.2388e-04, 1.3114e-04, 9.5635e-05, 1.1888e-04, 1.3056e-04, 1.1228e-04, + 1.0867e-04, 1.1590e-04], device='cuda:2') +2023-04-26 18:49:10,833 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34176.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:49:21,882 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7967, 1.3961, 1.9693, 1.9899, 1.5225, 1.3437, 1.6064, 1.0488], + device='cuda:2'), covar=tensor([0.0593, 0.1113, 0.0627, 0.0699, 0.0903, 0.1308, 0.0814, 0.1019], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0075, 0.0074, 0.0068, 0.0079, 0.0096, 0.0081, 0.0077], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:49:30,902 INFO [finetune.py:976] (2/7) Epoch 6, batch 5550, loss[loss=0.2054, simple_loss=0.2882, pruned_loss=0.06129, over 4862.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2628, pruned_loss=0.06889, over 954674.11 frames. ], batch size: 44, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:49:52,196 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34206.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:50:24,791 INFO [optim.py:369] (2/7) 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 18:50:36,987 INFO [finetune.py:976] (2/7) Epoch 6, batch 5600, loss[loss=0.2025, simple_loss=0.2754, pruned_loss=0.06479, over 4831.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2677, pruned_loss=0.07084, over 954175.46 frames. ], batch size: 49, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:51:28,269 INFO [finetune.py:976] (2/7) Epoch 6, batch 5650, loss[loss=0.2699, simple_loss=0.3282, pruned_loss=0.1058, over 4748.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2712, pruned_loss=0.07148, over 956213.66 frames. ], batch size: 59, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:51:31,888 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34295.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:51:46,981 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-26 18:51:48,255 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 18:51:58,324 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5073, 2.1529, 2.1022, 2.5881, 2.2042, 2.4386, 2.0873, 4.8615], + device='cuda:2'), covar=tensor([0.0541, 0.0622, 0.0693, 0.0981, 0.0559, 0.0465, 0.0626, 0.0118], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 18:52:02,999 INFO [optim.py:369] (2/7) 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 18:52:05,464 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34333.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:52:09,023 INFO [finetune.py:976] (2/7) Epoch 6, batch 5700, loss[loss=0.1966, simple_loss=0.235, pruned_loss=0.07905, over 4284.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2655, pruned_loss=0.07016, over 933398.38 frames. ], batch size: 18, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:52:18,752 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-26 18:52:39,656 INFO [finetune.py:976] (2/7) Epoch 7, batch 0, loss[loss=0.2087, simple_loss=0.2734, pruned_loss=0.07194, over 4773.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2734, pruned_loss=0.07194, over 4773.00 frames. ], batch size: 26, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:52:39,656 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 18:52:48,989 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5423, 2.1119, 1.7607, 2.0488, 1.6942, 1.7293, 1.7048, 1.4645], + device='cuda:2'), covar=tensor([0.2434, 0.1567, 0.1091, 0.1357, 0.3695, 0.1578, 0.1985, 0.2738], + device='cuda:2'), in_proj_covar=tensor([0.0304, 0.0321, 0.0235, 0.0295, 0.0315, 0.0277, 0.0263, 0.0286], + device='cuda:2'), 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:2') +2023-04-26 18:52:50,219 INFO [finetune.py:1010] (2/7) Epoch 7, validation: loss=0.1579, simple_loss=0.2317, pruned_loss=0.04207, over 2265189.00 frames. +2023-04-26 18:52:50,219 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 18:52:59,340 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34381.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:53:00,654 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-04-26 18:53:11,519 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34400.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:53:15,171 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 18:53:23,171 INFO [finetune.py:976] (2/7) Epoch 7, batch 50, loss[loss=0.1812, simple_loss=0.2619, pruned_loss=0.05025, over 4844.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.267, pruned_loss=0.06946, over 216017.83 frames. ], batch size: 44, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:53:32,046 INFO [optim.py:369] (2/7) 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 18:53:56,438 INFO [finetune.py:976] (2/7) Epoch 7, batch 100, loss[loss=0.2232, simple_loss=0.2784, pruned_loss=0.08397, over 4870.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.263, pruned_loss=0.0697, over 380485.78 frames. ], batch size: 31, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:54:19,354 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34501.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:54:21,064 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8647, 2.6042, 1.9605, 1.8442, 1.4240, 1.4781, 2.0611, 1.3376], + device='cuda:2'), covar=tensor([0.1747, 0.1427, 0.1557, 0.1926, 0.2494, 0.1935, 0.1153, 0.2137], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0219, 0.0174, 0.0206, 0.0210, 0.0185, 0.0166, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 18:54:29,756 INFO [finetune.py:976] (2/7) Epoch 7, batch 150, loss[loss=0.1401, simple_loss=0.2075, pruned_loss=0.03637, over 4834.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2565, pruned_loss=0.06623, over 508215.50 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 16.0 +2023-04-26 18:54:39,165 INFO [optim.py:369] (2/7) 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 18:54:43,464 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7133, 3.5638, 0.9499, 2.0445, 2.1102, 2.4550, 2.1388, 0.9413], + device='cuda:2'), covar=tensor([0.1354, 0.1061, 0.2081, 0.1296, 0.1098, 0.1137, 0.1405, 0.2163], + device='cuda:2'), in_proj_covar=tensor([0.0121, 0.0258, 0.0144, 0.0127, 0.0138, 0.0159, 0.0122, 0.0125], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:55:03,727 INFO [finetune.py:976] (2/7) Epoch 7, batch 200, loss[loss=0.1579, simple_loss=0.2336, pruned_loss=0.04108, over 4798.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2592, pruned_loss=0.06858, over 607149.33 frames. ], batch size: 29, lr: 3.89e-03, grad_scale: 32.0 +2023-04-26 18:55:27,246 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34595.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:55:46,306 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9104, 2.4509, 1.9368, 1.6459, 1.3963, 1.4786, 2.0184, 1.3419], + device='cuda:2'), covar=tensor([0.1726, 0.1524, 0.1547, 0.2037, 0.2539, 0.2059, 0.1140, 0.2235], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0220, 0.0175, 0.0206, 0.0210, 0.0186, 0.0167, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 18:56:04,709 INFO [finetune.py:976] (2/7) Epoch 7, batch 250, loss[loss=0.1973, simple_loss=0.2694, pruned_loss=0.06265, over 4904.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2646, pruned_loss=0.07115, over 683181.94 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:56:09,042 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34622.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:56:17,339 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34626.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:56:19,028 INFO [optim.py:369] (2/7) 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 18:56:39,247 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34643.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:56:41,138 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7474, 1.6802, 1.9769, 2.2793, 2.2913, 1.7074, 1.4492, 1.9347], + device='cuda:2'), covar=tensor([0.1102, 0.1277, 0.0764, 0.0703, 0.0616, 0.1190, 0.0966, 0.0739], + device='cuda:2'), in_proj_covar=tensor([0.0205, 0.0209, 0.0183, 0.0180, 0.0182, 0.0197, 0.0167, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:56:51,407 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:57:11,715 INFO [finetune.py:976] (2/7) Epoch 7, batch 300, loss[loss=0.2684, simple_loss=0.3209, pruned_loss=0.108, over 4824.00 frames. ], tot_loss[loss=0.205, simple_loss=0.267, pruned_loss=0.07146, over 742800.18 frames. ], batch size: 51, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:57:14,792 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6067, 1.2564, 1.2975, 1.3640, 1.8619, 1.4126, 1.1745, 1.2727], + device='cuda:2'), covar=tensor([0.1959, 0.1753, 0.2581, 0.1751, 0.0947, 0.2088, 0.2496, 0.2318], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0332, 0.0352, 0.0306, 0.0341, 0.0332, 0.0311, 0.0355], + device='cuda:2'), out_proj_covar=tensor([6.6315e-05, 7.0918e-05, 7.6295e-05, 6.3702e-05, 7.1870e-05, 7.1760e-05, + 6.7265e-05, 7.6398e-05], device='cuda:2') +2023-04-26 18:57:35,528 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34683.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 18:57:37,956 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34687.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:57:42,156 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8423, 3.7882, 2.6755, 4.4851, 3.8765, 3.9099, 1.9884, 3.8220], + device='cuda:2'), covar=tensor([0.1595, 0.1086, 0.3127, 0.1413, 0.3066, 0.1858, 0.5225, 0.2289], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0218, 0.0253, 0.0310, 0.0301, 0.0253, 0.0274, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 18:57:51,397 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34700.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:57:52,055 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2866, 1.6682, 2.0501, 2.7601, 2.1426, 1.5808, 1.4920, 2.1656], + device='cuda:2'), covar=tensor([0.4030, 0.4705, 0.2159, 0.3590, 0.3799, 0.3462, 0.5611, 0.3176], + device='cuda:2'), in_proj_covar=tensor([0.0277, 0.0254, 0.0218, 0.0323, 0.0214, 0.0228, 0.0239, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 18:57:52,626 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0432, 1.3955, 1.2450, 1.6391, 1.3998, 1.5362, 1.2727, 2.4793], + device='cuda:2'), covar=tensor([0.0642, 0.0815, 0.0841, 0.1218, 0.0693, 0.0512, 0.0765, 0.0234], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 18:57:53,207 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8211, 1.3373, 1.4511, 1.4883, 2.0111, 1.6079, 1.3336, 1.3179], + device='cuda:2'), covar=tensor([0.1585, 0.1579, 0.1972, 0.1472, 0.0806, 0.1464, 0.2062, 0.1785], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0332, 0.0352, 0.0305, 0.0341, 0.0332, 0.0310, 0.0354], + device='cuda:2'), 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:2') +2023-04-26 18:58:03,737 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34710.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 18:58:12,738 INFO [finetune.py:976] (2/7) Epoch 7, batch 350, loss[loss=0.2023, simple_loss=0.2633, pruned_loss=0.07064, over 4810.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2691, pruned_loss=0.07225, over 791940.88 frames. ], batch size: 39, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:58:28,035 INFO [optim.py:369] (2/7) 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 18:58:35,663 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 18:58:48,102 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.86 vs. limit=5.0 +2023-04-26 18:58:50,331 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34748.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:59:02,246 INFO [finetune.py:976] (2/7) Epoch 7, batch 400, loss[loss=0.1836, simple_loss=0.2508, pruned_loss=0.05825, over 4804.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2694, pruned_loss=0.07141, over 829854.82 frames. ], batch size: 45, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:59:02,363 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4971, 1.4987, 1.6953, 1.9817, 1.9783, 1.5230, 1.2045, 1.7321], + device='cuda:2'), covar=tensor([0.0997, 0.1213, 0.0766, 0.0654, 0.0640, 0.0993, 0.1038, 0.0664], + device='cuda:2'), in_proj_covar=tensor([0.0204, 0.0209, 0.0183, 0.0180, 0.0182, 0.0197, 0.0167, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 18:59:05,936 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34771.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 18:59:26,210 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34801.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 18:59:36,187 INFO [finetune.py:976] (2/7) Epoch 7, batch 450, loss[loss=0.1816, simple_loss=0.2406, pruned_loss=0.06131, over 4749.00 frames. ], tot_loss[loss=0.205, simple_loss=0.2682, pruned_loss=0.07089, over 858113.62 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 18:59:36,314 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6848, 1.2072, 1.3818, 1.5143, 1.9410, 1.5729, 1.2521, 1.2973], + device='cuda:2'), covar=tensor([0.2063, 0.1776, 0.2343, 0.1579, 0.1015, 0.1648, 0.2554, 0.2069], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0332, 0.0352, 0.0305, 0.0341, 0.0330, 0.0309, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.6284e-05, 7.0778e-05, 7.6250e-05, 6.3380e-05, 7.1759e-05, 7.1328e-05, + 6.6914e-05, 7.6125e-05], device='cuda:2') +2023-04-26 18:59:45,537 INFO [optim.py:369] (2/7) 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 18:59:58,665 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:00:09,434 INFO [finetune.py:976] (2/7) Epoch 7, batch 500, loss[loss=0.1979, simple_loss=0.2587, pruned_loss=0.06857, over 4896.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2654, pruned_loss=0.07015, over 879455.99 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:00:21,971 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-26 19:00:42,234 INFO [finetune.py:976] (2/7) Epoch 7, batch 550, loss[loss=0.1602, simple_loss=0.226, pruned_loss=0.04723, over 4733.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2616, pruned_loss=0.06907, over 897603.53 frames. ], batch size: 23, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:00:51,085 INFO [optim.py:369] (2/7) 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 19:01:10,912 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34949.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:01:32,182 INFO [finetune.py:976] (2/7) Epoch 7, batch 600, loss[loss=0.2223, simple_loss=0.2939, pruned_loss=0.0754, over 4930.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2627, pruned_loss=0.06982, over 908399.85 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:01:41,900 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3304, 1.5494, 1.5711, 1.7415, 1.6076, 1.7499, 1.7188, 1.6614], + device='cuda:2'), covar=tensor([0.6238, 0.9121, 0.7768, 0.6878, 0.8535, 1.2151, 0.9663, 0.7793], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0393, 0.0317, 0.0328, 0.0344, 0.0409, 0.0373, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 19:01:45,331 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34978.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 19:01:54,262 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34982.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:02:27,946 INFO [finetune.py:976] (2/7) Epoch 7, batch 650, loss[loss=0.2033, simple_loss=0.2675, pruned_loss=0.06953, over 4131.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2671, pruned_loss=0.07113, over 918972.84 frames. ], batch size: 65, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:02:42,342 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 7, batch 700, loss[loss=0.1827, simple_loss=0.256, pruned_loss=0.05465, over 4789.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2686, pruned_loss=0.07149, over 926205.14 frames. ], batch size: 29, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:03:19,116 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35066.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 19:03:47,927 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35092.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:04:01,354 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-04-26 19:04:19,900 INFO [finetune.py:976] (2/7) Epoch 7, batch 750, loss[loss=0.2108, simple_loss=0.2732, pruned_loss=0.0742, over 4804.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.27, pruned_loss=0.07208, over 933087.84 frames. ], batch size: 45, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:04:33,721 INFO [optim.py:369] (2/7) 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 19:04:51,749 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4046, 2.4885, 2.0129, 2.2141, 2.5852, 2.0663, 3.5387, 1.7914], + device='cuda:2'), covar=tensor([0.4214, 0.2533, 0.5172, 0.4084, 0.2077, 0.3240, 0.1485, 0.4687], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0353, 0.0436, 0.0363, 0.0390, 0.0382, 0.0386, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:05:05,764 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 800, loss[loss=0.2459, simple_loss=0.279, pruned_loss=0.1064, over 4228.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2699, pruned_loss=0.07214, over 937136.39 frames. ], batch size: 18, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:06:20,321 INFO [finetune.py:976] (2/7) Epoch 7, batch 850, loss[loss=0.1933, simple_loss=0.2624, pruned_loss=0.06206, over 4840.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.267, pruned_loss=0.07092, over 941495.68 frames. ], batch size: 44, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:06:22,175 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1026, 1.3509, 1.1926, 1.6168, 1.4158, 1.4173, 1.2697, 2.4054], + device='cuda:2'), covar=tensor([0.0647, 0.0820, 0.0878, 0.1221, 0.0682, 0.0557, 0.0778, 0.0252], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 19:06:32,698 INFO [optim.py:369] (2/7) 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 19:06:45,872 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-26 19:06:58,279 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35249.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:07:20,078 INFO [finetune.py:976] (2/7) Epoch 7, batch 900, loss[loss=0.1538, simple_loss=0.2219, pruned_loss=0.04283, over 4923.00 frames. ], tot_loss[loss=0.2023, simple_loss=0.2644, pruned_loss=0.07013, over 944062.21 frames. ], batch size: 46, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:07:31,629 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4186, 1.0868, 0.3996, 1.1689, 1.0982, 1.3423, 1.2478, 1.2394], + device='cuda:2'), covar=tensor([0.0590, 0.0440, 0.0500, 0.0613, 0.0357, 0.0572, 0.0568, 0.0652], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 19:07:38,509 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35278.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:07:40,986 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35282.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:08:02,071 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35297.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:08:14,476 INFO [finetune.py:976] (2/7) Epoch 7, batch 950, loss[loss=0.2248, simple_loss=0.2856, pruned_loss=0.08198, over 4816.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2622, pruned_loss=0.06905, over 946000.41 frames. ], batch size: 38, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:08:27,232 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35326.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:08:29,016 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35345.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:09:20,441 INFO [finetune.py:976] (2/7) Epoch 7, batch 1000, loss[loss=0.2526, simple_loss=0.3005, pruned_loss=0.1024, over 4097.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2638, pruned_loss=0.06971, over 944930.83 frames. ], batch size: 65, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:09:20,537 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35366.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:09:29,970 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35374.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:10:05,443 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0763, 2.7146, 1.1159, 1.3016, 1.8529, 1.1660, 3.5352, 1.8315], + device='cuda:2'), covar=tensor([0.0708, 0.0725, 0.0965, 0.1307, 0.0553, 0.1151, 0.0236, 0.0656], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0068, 0.0051, 0.0048, 0.0052, 0.0053, 0.0080, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 19:10:06,712 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35406.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:10:15,440 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-26 19:10:18,215 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 1050, loss[loss=0.1685, simple_loss=0.2316, pruned_loss=0.05272, over 4776.00 frames. ], tot_loss[loss=0.205, simple_loss=0.268, pruned_loss=0.07098, over 946347.53 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:10:39,105 INFO [optim.py:369] (2/7) 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 19:10:48,431 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35435.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:11:02,533 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35448.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:11:31,123 INFO [finetune.py:976] (2/7) Epoch 7, batch 1100, loss[loss=0.1798, simple_loss=0.2429, pruned_loss=0.05833, over 4783.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2708, pruned_loss=0.0721, over 949851.84 frames. ], batch size: 29, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:11:44,678 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6946, 1.0498, 1.3982, 1.5598, 1.5631, 1.6690, 1.3989, 1.4853], + device='cuda:2'), covar=tensor([0.5574, 0.7805, 0.6966, 0.6783, 0.8022, 1.1310, 0.7300, 0.6822], + device='cuda:2'), in_proj_covar=tensor([0.0320, 0.0391, 0.0316, 0.0327, 0.0343, 0.0407, 0.0371, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 19:11:57,243 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9599, 1.9706, 1.7888, 1.6333, 2.1414, 1.6665, 2.6553, 1.6151], + device='cuda:2'), covar=tensor([0.3858, 0.1845, 0.4424, 0.3443, 0.1649, 0.2753, 0.1325, 0.4165], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0349, 0.0431, 0.0361, 0.0386, 0.0379, 0.0382, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:12:09,916 INFO [finetune.py:976] (2/7) Epoch 7, batch 1150, loss[loss=0.1675, simple_loss=0.2403, pruned_loss=0.04731, over 4859.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2718, pruned_loss=0.07265, over 951240.18 frames. ], batch size: 31, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:12:18,334 INFO [optim.py:369] (2/7) 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 19:12:19,080 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35530.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:12:43,336 INFO [finetune.py:976] (2/7) Epoch 7, batch 1200, loss[loss=0.1812, simple_loss=0.2468, pruned_loss=0.05786, over 4894.00 frames. ], tot_loss[loss=0.2077, simple_loss=0.2705, pruned_loss=0.07247, over 952706.41 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:13:00,089 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35591.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:13:17,334 INFO [finetune.py:976] (2/7) Epoch 7, batch 1250, loss[loss=0.2378, simple_loss=0.2787, pruned_loss=0.09848, over 4909.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.2664, pruned_loss=0.07089, over 953262.10 frames. ], batch size: 32, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:13:26,236 INFO [optim.py:369] (2/7) 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 19:13:42,754 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6862, 1.9327, 0.9693, 1.4062, 2.0415, 1.5641, 1.4325, 1.5508], + device='cuda:2'), covar=tensor([0.0544, 0.0378, 0.0389, 0.0578, 0.0286, 0.0565, 0.0566, 0.0600], + device='cuda:2'), in_proj_covar=tensor([0.0031, 0.0026, 0.0024, 0.0031, 0.0021, 0.0030, 0.0030, 0.0031], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 19:13:51,306 INFO [finetune.py:976] (2/7) Epoch 7, batch 1300, loss[loss=0.2177, simple_loss=0.2705, pruned_loss=0.08245, over 4762.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2639, pruned_loss=0.06999, over 953819.77 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:14:01,121 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-04-26 19:14:03,987 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4277, 1.8203, 2.2330, 2.8920, 2.2887, 1.8193, 1.6775, 2.0988], + device='cuda:2'), covar=tensor([0.4330, 0.4271, 0.2001, 0.3722, 0.3807, 0.3400, 0.5178, 0.3422], + device='cuda:2'), in_proj_covar=tensor([0.0280, 0.0256, 0.0219, 0.0325, 0.0215, 0.0229, 0.0239, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 19:14:14,094 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35701.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:14:35,192 INFO [finetune.py:976] (2/7) Epoch 7, batch 1350, loss[loss=0.1925, simple_loss=0.263, pruned_loss=0.06103, over 4822.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2639, pruned_loss=0.06996, over 955203.55 frames. ], batch size: 40, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:14:43,892 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6804, 2.0520, 1.7746, 1.9809, 1.4751, 1.7259, 1.8219, 1.5358], + device='cuda:2'), covar=tensor([0.1797, 0.1346, 0.1010, 0.1056, 0.3376, 0.1313, 0.1770, 0.2216], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0325, 0.0236, 0.0300, 0.0318, 0.0279, 0.0266, 0.0289], + device='cuda:2'), out_proj_covar=tensor([1.2505e-04, 1.3157e-04, 9.5751e-05, 1.2031e-04, 1.3089e-04, 1.1269e-04, + 1.0916e-04, 1.1611e-04], device='cuda:2') +2023-04-26 19:14:54,290 INFO [optim.py:369] (2/7) 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 19:14:54,986 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35730.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:15:17,733 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35748.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:15:17,955 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-26 19:15:40,834 INFO [finetune.py:976] (2/7) Epoch 7, batch 1400, loss[loss=0.2311, simple_loss=0.2898, pruned_loss=0.08618, over 4873.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2665, pruned_loss=0.07092, over 954280.06 frames. ], batch size: 34, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:15:44,426 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35771.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:16:06,188 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35796.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:16:17,413 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-26 19:16:25,145 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-26 19:16:26,246 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35811.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:16:30,132 INFO [finetune.py:976] (2/7) Epoch 7, batch 1450, loss[loss=0.2097, simple_loss=0.2672, pruned_loss=0.0761, over 4747.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2682, pruned_loss=0.07121, over 953531.47 frames. ], batch size: 27, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:16:50,309 INFO [optim.py:369] (2/7) 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 19:16:58,478 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35832.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:17:41,157 INFO [finetune.py:976] (2/7) Epoch 7, batch 1500, loss[loss=0.2396, simple_loss=0.2939, pruned_loss=0.09262, over 4853.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2705, pruned_loss=0.0721, over 954630.63 frames. ], batch size: 31, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:17:46,048 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35872.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:17:55,859 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35886.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:18:31,248 INFO [finetune.py:976] (2/7) Epoch 7, batch 1550, loss[loss=0.1965, simple_loss=0.2633, pruned_loss=0.06484, over 4908.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2695, pruned_loss=0.07113, over 953892.21 frames. ], batch size: 46, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:18:51,453 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 7, batch 1600, loss[loss=0.1748, simple_loss=0.2402, pruned_loss=0.0547, over 4839.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2667, pruned_loss=0.07, over 954804.47 frames. ], batch size: 44, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:20:22,134 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-26 19:20:30,767 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36001.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:20:45,229 INFO [finetune.py:976] (2/7) Epoch 7, batch 1650, loss[loss=0.2324, simple_loss=0.2817, pruned_loss=0.09161, over 4829.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2648, pruned_loss=0.06961, over 953392.10 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:20:59,791 INFO [optim.py:369] (2/7) 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 19:21:00,971 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36030.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:21:26,273 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36049.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:21:28,204 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5591, 1.0920, 1.3327, 1.2486, 1.7210, 1.3703, 1.1014, 1.3078], + device='cuda:2'), covar=tensor([0.1720, 0.1699, 0.2373, 0.1549, 0.1016, 0.1662, 0.2310, 0.2197], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0331, 0.0354, 0.0303, 0.0340, 0.0330, 0.0309, 0.0356], + device='cuda:2'), out_proj_covar=tensor([6.6114e-05, 7.0521e-05, 7.6686e-05, 6.2935e-05, 7.1550e-05, 7.1280e-05, + 6.6810e-05, 7.6555e-05], device='cuda:2') +2023-04-26 19:21:48,193 INFO [finetune.py:976] (2/7) Epoch 7, batch 1700, loss[loss=0.1843, simple_loss=0.2461, pruned_loss=0.06128, over 4890.00 frames. ], tot_loss[loss=0.2015, simple_loss=0.2635, pruned_loss=0.06975, over 955219.30 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:21:56,085 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36078.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:22:21,090 INFO [finetune.py:976] (2/7) Epoch 7, batch 1750, loss[loss=0.1992, simple_loss=0.2755, pruned_loss=0.06144, over 4909.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.2647, pruned_loss=0.07023, over 956230.45 frames. ], batch size: 43, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:22:28,799 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36127.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:22:29,955 INFO [optim.py:369] (2/7) 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 19:22:54,678 INFO [finetune.py:976] (2/7) Epoch 7, batch 1800, loss[loss=0.186, simple_loss=0.252, pruned_loss=0.05998, over 4905.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2677, pruned_loss=0.07065, over 955567.63 frames. ], batch size: 36, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:22:55,364 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36167.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:23:05,437 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36181.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:23:08,501 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36186.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:23:23,506 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4263, 1.7587, 1.6640, 1.8595, 1.7894, 1.9836, 1.7524, 1.6876], + device='cuda:2'), covar=tensor([0.5998, 0.7787, 0.7561, 0.6110, 0.7460, 0.9891, 0.8510, 0.8095], + device='cuda:2'), in_proj_covar=tensor([0.0319, 0.0390, 0.0316, 0.0327, 0.0342, 0.0405, 0.0371, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 19:23:44,253 INFO [finetune.py:976] (2/7) Epoch 7, batch 1850, loss[loss=0.2054, simple_loss=0.2672, pruned_loss=0.07182, over 4872.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2695, pruned_loss=0.07117, over 956389.46 frames. ], batch size: 31, lr: 3.88e-03, grad_scale: 32.0 +2023-04-26 19:24:03,616 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36234.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:24:19,057 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36242.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:24:45,933 INFO [finetune.py:976] (2/7) Epoch 7, batch 1900, loss[loss=0.2117, simple_loss=0.2707, pruned_loss=0.07635, over 4861.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2702, pruned_loss=0.07102, over 957278.35 frames. ], batch size: 34, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:24:51,726 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 19:25:19,583 INFO [finetune.py:976] (2/7) Epoch 7, batch 1950, loss[loss=0.123, simple_loss=0.2058, pruned_loss=0.02009, over 4808.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2681, pruned_loss=0.06993, over 958772.68 frames. ], batch size: 41, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:25:20,281 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5752, 1.3673, 4.2274, 3.9240, 3.6996, 3.9603, 3.9091, 3.7041], + device='cuda:2'), covar=tensor([0.6729, 0.5669, 0.0927, 0.1562, 0.1023, 0.1473, 0.1753, 0.1316], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0304, 0.0408, 0.0412, 0.0350, 0.0407, 0.0316, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:25:27,367 INFO [optim.py:369] (2/7) 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 19:25:30,368 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0610, 3.8819, 1.2733, 2.2691, 2.6044, 2.9336, 2.3670, 1.4095], + device='cuda:2'), covar=tensor([0.1284, 0.1076, 0.2037, 0.1311, 0.0936, 0.0999, 0.1406, 0.1760], + device='cuda:2'), in_proj_covar=tensor([0.0120, 0.0257, 0.0145, 0.0127, 0.0137, 0.0158, 0.0122, 0.0125], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 19:26:09,641 INFO [finetune.py:976] (2/7) Epoch 7, batch 2000, loss[loss=0.198, simple_loss=0.2628, pruned_loss=0.06659, over 4852.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.2655, pruned_loss=0.06938, over 956342.13 frames. ], batch size: 47, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:27:00,248 INFO [finetune.py:976] (2/7) Epoch 7, batch 2050, loss[loss=0.2242, simple_loss=0.2722, pruned_loss=0.0881, over 4901.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2631, pruned_loss=0.0686, over 957361.80 frames. ], batch size: 43, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:27:07,143 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36427.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:27:08,255 INFO [optim.py:369] (2/7) 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 19:27:08,344 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5402, 1.1528, 4.0367, 3.5086, 3.7290, 3.8305, 3.8212, 3.3817], + device='cuda:2'), covar=tensor([0.9408, 0.8824, 0.1722, 0.3300, 0.2073, 0.3492, 0.2299, 0.3103], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0304, 0.0409, 0.0413, 0.0350, 0.0408, 0.0317, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:27:08,379 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8418, 1.7964, 2.0135, 2.3184, 2.1376, 1.8025, 1.4447, 1.8918], + device='cuda:2'), covar=tensor([0.1014, 0.1132, 0.0660, 0.0630, 0.0708, 0.0992, 0.0964, 0.0656], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0206, 0.0180, 0.0177, 0.0180, 0.0194, 0.0162, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:27:43,883 INFO [finetune.py:976] (2/7) Epoch 7, batch 2100, loss[loss=0.1739, simple_loss=0.2449, pruned_loss=0.05145, over 4756.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2623, pruned_loss=0.0683, over 958034.45 frames. ], batch size: 26, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:27:45,077 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36467.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:27:48,126 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6432, 1.3104, 4.2336, 3.9347, 3.6895, 3.9058, 3.9267, 3.7586], + device='cuda:2'), covar=tensor([0.6934, 0.5667, 0.1049, 0.1608, 0.1223, 0.1426, 0.2099, 0.1439], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0306, 0.0411, 0.0417, 0.0353, 0.0411, 0.0319, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:27:49,908 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36475.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:28:17,169 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 2150, loss[loss=0.1904, simple_loss=0.2581, pruned_loss=0.06135, over 4769.00 frames. ], tot_loss[loss=0.2017, simple_loss=0.2648, pruned_loss=0.06929, over 957877.49 frames. ], batch size: 28, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:28:25,935 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) Epoch 7, batch 2200, loss[loss=0.1589, simple_loss=0.214, pruned_loss=0.05185, over 4710.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2672, pruned_loss=0.0703, over 956477.73 frames. ], batch size: 23, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:29:28,460 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36605.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:29:30,317 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36608.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:29:30,373 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7549, 1.6139, 1.9622, 2.1650, 1.9041, 1.6832, 1.7840, 1.8155], + device='cuda:2'), covar=tensor([0.8000, 1.1634, 1.2189, 1.0826, 0.9590, 1.5034, 1.5200, 1.3194], + device='cuda:2'), in_proj_covar=tensor([0.0414, 0.0435, 0.0519, 0.0541, 0.0441, 0.0462, 0.0473, 0.0471], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:29:38,198 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36612.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:29:41,092 INFO [finetune.py:976] (2/7) Epoch 7, batch 2250, loss[loss=0.2052, simple_loss=0.2738, pruned_loss=0.06825, over 4794.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2675, pruned_loss=0.0703, over 957068.14 frames. ], batch size: 45, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:30:00,404 INFO [optim.py:369] (2/7) 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 19:30:02,569 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 19:30:15,059 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6556, 1.1864, 1.3122, 1.3070, 1.8292, 1.4630, 1.1665, 1.2754], + device='cuda:2'), covar=tensor([0.1514, 0.1278, 0.1863, 0.1297, 0.0709, 0.1195, 0.1728, 0.1707], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0332, 0.0357, 0.0307, 0.0342, 0.0333, 0.0310, 0.0356], + device='cuda:2'), out_proj_covar=tensor([6.6194e-05, 7.0812e-05, 7.7325e-05, 6.3763e-05, 7.2009e-05, 7.1905e-05, + 6.7121e-05, 7.6687e-05], device='cuda:2') +2023-04-26 19:30:40,443 INFO [finetune.py:976] (2/7) Epoch 7, batch 2300, loss[loss=0.1882, simple_loss=0.2576, pruned_loss=0.05938, over 4756.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2674, pruned_loss=0.06979, over 955521.11 frames. ], batch size: 28, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:30:40,581 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36666.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:30:42,398 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36669.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:30:45,356 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36673.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:30:52,470 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1182, 1.4984, 1.9661, 2.5118, 1.8432, 1.5374, 1.2263, 1.7622], + device='cuda:2'), covar=tensor([0.4011, 0.4373, 0.2043, 0.3229, 0.3623, 0.3387, 0.5304, 0.3080], + device='cuda:2'), in_proj_covar=tensor([0.0279, 0.0253, 0.0217, 0.0321, 0.0214, 0.0228, 0.0238, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 19:31:02,260 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36700.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:31:15,639 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8505, 1.3840, 1.4422, 1.5880, 2.0481, 1.6277, 1.3930, 1.3917], + device='cuda:2'), covar=tensor([0.1615, 0.1843, 0.2135, 0.1505, 0.0926, 0.1852, 0.2246, 0.2195], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0331, 0.0355, 0.0305, 0.0341, 0.0331, 0.0309, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.5808e-05, 7.0455e-05, 7.6947e-05, 6.3370e-05, 7.1811e-05, 7.1554e-05, + 6.6786e-05, 7.6226e-05], device='cuda:2') +2023-04-26 19:31:23,676 INFO [finetune.py:976] (2/7) Epoch 7, batch 2350, loss[loss=0.1917, simple_loss=0.2434, pruned_loss=0.06999, over 4761.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2651, pruned_loss=0.0692, over 954435.31 frames. ], batch size: 28, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:31:38,221 INFO [optim.py:369] (2/7) 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 19:31:45,446 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1424, 2.6239, 1.0825, 1.5079, 2.1596, 1.2418, 3.5750, 1.8196], + device='cuda:2'), covar=tensor([0.0700, 0.0666, 0.0798, 0.1261, 0.0499, 0.1026, 0.0223, 0.0659], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0051, 0.0048, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 19:31:56,345 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9616, 2.6390, 2.2130, 2.3287, 1.8481, 2.1601, 2.0969, 1.7443], + device='cuda:2'), covar=tensor([0.1893, 0.1066, 0.0838, 0.1220, 0.2909, 0.1062, 0.1921, 0.2282], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0322, 0.0235, 0.0296, 0.0317, 0.0274, 0.0264, 0.0286], + device='cuda:2'), out_proj_covar=tensor([1.2364e-04, 1.3035e-04, 9.5030e-05, 1.1883e-04, 1.3047e-04, 1.1084e-04, + 1.0816e-04, 1.1513e-04], device='cuda:2') +2023-04-26 19:32:21,176 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36761.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 19:32:30,354 INFO [finetune.py:976] (2/7) Epoch 7, batch 2400, loss[loss=0.2302, simple_loss=0.2805, pruned_loss=0.08993, over 4835.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2632, pruned_loss=0.06873, over 955085.23 frames. ], batch size: 33, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:32:56,365 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36805.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:33:03,827 INFO [finetune.py:976] (2/7) Epoch 7, batch 2450, loss[loss=0.1746, simple_loss=0.251, pruned_loss=0.04907, over 4806.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2609, pruned_loss=0.06817, over 954525.61 frames. ], batch size: 39, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:33:12,663 INFO [optim.py:369] (2/7) 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 19:33:18,545 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36837.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:33:29,481 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4006, 1.5820, 1.6088, 1.8320, 1.6516, 1.7313, 1.7910, 1.7361], + device='cuda:2'), covar=tensor([0.6354, 0.8608, 0.6999, 0.6677, 0.7821, 1.1294, 0.8238, 0.7688], + device='cuda:2'), in_proj_covar=tensor([0.0322, 0.0393, 0.0319, 0.0329, 0.0347, 0.0410, 0.0373, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 19:33:37,052 INFO [finetune.py:976] (2/7) Epoch 7, batch 2500, loss[loss=0.2135, simple_loss=0.2685, pruned_loss=0.07927, over 4823.00 frames. ], tot_loss[loss=0.2017, simple_loss=0.2636, pruned_loss=0.06992, over 956044.94 frames. ], batch size: 30, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:33:37,194 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36866.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:33:51,058 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36885.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:01,340 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.4006, 4.3186, 3.0403, 5.0294, 4.2947, 4.3370, 1.8460, 4.3849], + device='cuda:2'), covar=tensor([0.1540, 0.1012, 0.3569, 0.0970, 0.2852, 0.1656, 0.5384, 0.1889], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0219, 0.0252, 0.0310, 0.0304, 0.0253, 0.0274, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 19:34:07,395 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 2550, loss[loss=0.2105, simple_loss=0.2871, pruned_loss=0.06695, over 4900.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2669, pruned_loss=0.07049, over 955470.54 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:34:13,959 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-26 19:34:20,142 INFO [optim.py:369] (2/7) 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 19:34:32,613 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9758, 2.4089, 1.0576, 1.3144, 1.7998, 1.1868, 3.0342, 1.5849], + device='cuda:2'), covar=tensor([0.0697, 0.0566, 0.0740, 0.1330, 0.0507, 0.1081, 0.0271, 0.0672], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 19:34:45,093 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36958.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:52,522 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36964.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:34:55,491 INFO [finetune.py:976] (2/7) Epoch 7, batch 2600, loss[loss=0.2142, simple_loss=0.2706, pruned_loss=0.07883, over 4870.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2684, pruned_loss=0.07093, over 955412.32 frames. ], batch size: 34, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:35:00,961 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36968.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:35:03,354 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1796, 1.6390, 1.5352, 2.0362, 1.8466, 2.2100, 1.4881, 4.3332], + device='cuda:2'), covar=tensor([0.0676, 0.0846, 0.0823, 0.1221, 0.0690, 0.0566, 0.0798, 0.0119], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 19:35:04,483 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36972.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:35:14,200 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.76 vs. limit=5.0 +2023-04-26 19:35:48,148 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37004.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:36:01,123 INFO [finetune.py:976] (2/7) Epoch 7, batch 2650, loss[loss=0.1956, simple_loss=0.2562, pruned_loss=0.06751, over 4777.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2681, pruned_loss=0.07038, over 954855.50 frames. ], batch size: 26, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:36:03,039 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37019.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:36:09,955 INFO [optim.py:369] (2/7) 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 19:36:28,798 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37056.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 19:36:34,290 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37065.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:36:34,793 INFO [finetune.py:976] (2/7) Epoch 7, batch 2700, loss[loss=0.1756, simple_loss=0.2402, pruned_loss=0.05546, over 4930.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2678, pruned_loss=0.07057, over 956201.56 frames. ], batch size: 38, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:36:44,430 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1036, 1.5016, 1.3504, 1.7800, 1.5529, 1.8557, 1.3346, 3.2703], + device='cuda:2'), covar=tensor([0.0664, 0.0737, 0.0760, 0.1107, 0.0611, 0.0483, 0.0730, 0.0168], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 19:36:57,344 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0763, 2.4583, 0.9193, 1.3107, 1.6540, 1.2760, 3.0497, 1.5494], + device='cuda:2'), covar=tensor([0.0654, 0.0536, 0.0752, 0.1249, 0.0522, 0.0953, 0.0242, 0.0659], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 19:37:08,964 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4474, 1.3985, 1.8097, 1.7842, 1.4037, 1.2027, 1.5151, 1.0045], + device='cuda:2'), covar=tensor([0.0781, 0.0851, 0.0552, 0.0722, 0.0890, 0.1414, 0.0737, 0.0903], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0074, 0.0072, 0.0067, 0.0077, 0.0095, 0.0080, 0.0076], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 19:37:19,208 INFO [finetune.py:976] (2/7) Epoch 7, batch 2750, loss[loss=0.2127, simple_loss=0.2759, pruned_loss=0.07479, over 4910.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2649, pruned_loss=0.06931, over 956661.71 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:37:32,276 INFO [optim.py:369] (2/7) 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 19:38:17,435 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37161.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 19:38:20,451 INFO [finetune.py:976] (2/7) Epoch 7, batch 2800, loss[loss=0.1743, simple_loss=0.2339, pruned_loss=0.05732, over 4869.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2611, pruned_loss=0.06745, over 954910.47 frames. ], batch size: 31, lr: 3.87e-03, grad_scale: 64.0 +2023-04-26 19:39:00,020 INFO [finetune.py:976] (2/7) Epoch 7, batch 2850, loss[loss=0.1654, simple_loss=0.2395, pruned_loss=0.04559, over 4764.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2608, pruned_loss=0.06743, over 957694.55 frames. ], batch size: 54, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:39:08,542 INFO [optim.py:369] (2/7) 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 19:39:30,810 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37261.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:32,613 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 2900, loss[loss=0.3232, simple_loss=0.3706, pruned_loss=0.1379, over 4814.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.2654, pruned_loss=0.06989, over 956910.51 frames. ], batch size: 51, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:39:34,410 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37267.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:35,010 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:39:35,054 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:02,582 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37309.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:10,642 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37312.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:11,892 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37314.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:13,061 INFO [finetune.py:976] (2/7) Epoch 7, batch 2950, loss[loss=0.1975, simple_loss=0.2679, pruned_loss=0.06354, over 4903.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2682, pruned_loss=0.07083, over 954579.54 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:40:13,122 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37316.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:40:32,668 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 19:41:05,672 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37356.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 19:41:08,606 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37360.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:41:17,428 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37364.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:41:18,550 INFO [finetune.py:976] (2/7) Epoch 7, batch 3000, loss[loss=0.1854, simple_loss=0.2607, pruned_loss=0.0551, over 4800.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2703, pruned_loss=0.0722, over 954385.27 frames. ], batch size: 29, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:41:18,550 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 19:41:26,458 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2100, 1.5531, 1.4512, 1.8109, 1.5742, 1.6771, 1.3985, 3.1180], + device='cuda:2'), covar=tensor([0.0667, 0.0839, 0.0802, 0.1240, 0.0705, 0.0497, 0.0754, 0.0200], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0045, 0.0041, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 19:41:30,401 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4085, 1.4555, 3.8842, 3.5910, 3.5105, 3.7275, 3.7844, 3.4118], + device='cuda:2'), covar=tensor([0.6638, 0.5075, 0.1291, 0.2147, 0.1297, 0.1237, 0.0875, 0.1913], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0304, 0.0410, 0.0415, 0.0351, 0.0407, 0.0318, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:41:32,895 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1800, 2.5744, 1.0591, 1.3956, 1.8829, 1.3795, 3.0593, 1.5762], + device='cuda:2'), covar=tensor([0.0644, 0.0491, 0.0719, 0.1310, 0.0489, 0.0940, 0.0239, 0.0688], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0068, 0.0051, 0.0048, 0.0053, 0.0053, 0.0079, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 19:41:33,484 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7209, 2.1010, 1.8667, 2.0099, 1.6078, 1.8244, 1.8324, 1.4860], + device='cuda:2'), covar=tensor([0.1769, 0.1019, 0.0798, 0.0994, 0.2637, 0.0947, 0.1610, 0.2228], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0327, 0.0237, 0.0299, 0.0321, 0.0278, 0.0267, 0.0290], + device='cuda:2'), out_proj_covar=tensor([1.2544e-04, 1.3233e-04, 9.5907e-05, 1.2006e-04, 1.3213e-04, 1.1224e-04, + 1.0947e-04, 1.1667e-04], device='cuda:2') +2023-04-26 19:41:33,717 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6513, 2.1165, 1.8134, 2.0137, 1.6655, 1.7597, 1.7642, 1.3936], + device='cuda:2'), covar=tensor([0.1969, 0.1203, 0.0944, 0.1233, 0.3187, 0.1286, 0.1889, 0.2604], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0327, 0.0237, 0.0299, 0.0321, 0.0278, 0.0267, 0.0290], + device='cuda:2'), out_proj_covar=tensor([1.2544e-04, 1.3233e-04, 9.5907e-05, 1.2006e-04, 1.3213e-04, 1.1224e-04, + 1.0947e-04, 1.1667e-04], device='cuda:2') +2023-04-26 19:41:40,583 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 19:42:11,806 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7271, 2.1150, 0.9922, 1.4775, 2.2477, 1.6367, 1.5519, 1.6705], + device='cuda:2'), covar=tensor([0.0543, 0.0370, 0.0348, 0.0585, 0.0253, 0.0552, 0.0536, 0.0605], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0038, 0.0048, 0.0048, 0.0049], + device='cuda:2') +2023-04-26 19:42:15,396 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 3050, loss[loss=0.2308, simple_loss=0.3006, pruned_loss=0.08053, over 4907.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2721, pruned_loss=0.07298, over 952957.22 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:42:30,236 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37425.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:42:34,155 INFO [optim.py:369] (2/7) 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 19:42:54,537 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37461.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:42:57,505 INFO [finetune.py:976] (2/7) Epoch 7, batch 3100, loss[loss=0.1933, simple_loss=0.2598, pruned_loss=0.06347, over 4819.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2694, pruned_loss=0.07112, over 953308.83 frames. ], batch size: 41, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:43:11,689 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8702, 1.2120, 4.8173, 4.2430, 4.3570, 4.4715, 4.2065, 3.9554], + device='cuda:2'), covar=tensor([0.8992, 0.8908, 0.1363, 0.2817, 0.1920, 0.2669, 0.2641, 0.3029], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0305, 0.0411, 0.0415, 0.0352, 0.0408, 0.0319, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:43:23,832 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3297, 1.6221, 1.5808, 2.1532, 2.3950, 2.0707, 1.9298, 1.7920], + device='cuda:2'), covar=tensor([0.1583, 0.1848, 0.2146, 0.1742, 0.1283, 0.1805, 0.2228, 0.1878], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0330, 0.0356, 0.0304, 0.0342, 0.0331, 0.0309, 0.0355], + device='cuda:2'), out_proj_covar=tensor([6.5983e-05, 7.0354e-05, 7.7085e-05, 6.3154e-05, 7.1972e-05, 7.1307e-05, + 6.6685e-05, 7.6223e-05], device='cuda:2') +2023-04-26 19:43:26,074 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37509.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 19:43:31,235 INFO [finetune.py:976] (2/7) Epoch 7, batch 3150, loss[loss=0.2155, simple_loss=0.2707, pruned_loss=0.08011, over 4755.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2667, pruned_loss=0.06993, over 955019.26 frames. ], batch size: 26, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:43:52,227 INFO [optim.py:369] (2/7) 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 19:44:36,861 INFO [finetune.py:976] (2/7) Epoch 7, batch 3200, loss[loss=0.1977, simple_loss=0.2667, pruned_loss=0.06433, over 4734.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2625, pruned_loss=0.06836, over 954822.59 frames. ], batch size: 59, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:44:37,546 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37567.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:44:59,050 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-26 19:45:10,334 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-26 19:45:42,231 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 7, batch 3250, loss[loss=0.2014, simple_loss=0.2808, pruned_loss=0.06102, over 4904.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2625, pruned_loss=0.06812, over 953934.34 frames. ], batch size: 43, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:45:55,035 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-26 19:45:55,396 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37624.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:45:56,673 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0283, 1.6342, 1.9553, 2.3611, 2.3591, 2.0342, 1.6239, 1.9749], + device='cuda:2'), covar=tensor([0.0896, 0.1331, 0.0718, 0.0603, 0.0597, 0.0839, 0.0883, 0.0655], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0204, 0.0181, 0.0177, 0.0179, 0.0193, 0.0162, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:46:03,764 INFO [optim.py:369] (2/7) 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 19:46:41,391 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37660.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:46:42,586 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37662.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:46:45,459 INFO [finetune.py:976] (2/7) Epoch 7, batch 3300, loss[loss=0.195, simple_loss=0.2671, pruned_loss=0.06146, over 4100.00 frames. ], tot_loss[loss=0.2032, simple_loss=0.2665, pruned_loss=0.06997, over 951679.68 frames. ], batch size: 65, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:47:48,146 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37708.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:47:57,368 INFO [finetune.py:976] (2/7) Epoch 7, batch 3350, loss[loss=0.199, simple_loss=0.2721, pruned_loss=0.06291, over 4799.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2682, pruned_loss=0.07056, over 951319.59 frames. ], batch size: 45, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:47:59,883 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37720.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:48:12,882 INFO [optim.py:369] (2/7) 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 19:48:36,824 INFO [finetune.py:976] (2/7) Epoch 7, batch 3400, loss[loss=0.1897, simple_loss=0.2654, pruned_loss=0.05699, over 4744.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2692, pruned_loss=0.07079, over 952312.41 frames. ], batch size: 54, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:48:59,244 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2969, 2.9368, 0.9351, 1.6800, 1.7480, 2.2362, 1.7600, 0.9675], + device='cuda:2'), covar=tensor([0.1358, 0.0889, 0.1724, 0.1269, 0.1058, 0.0862, 0.1470, 0.2046], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0254, 0.0142, 0.0124, 0.0135, 0.0155, 0.0120, 0.0124], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 19:49:36,094 INFO [finetune.py:976] (2/7) Epoch 7, batch 3450, loss[loss=0.2867, simple_loss=0.3135, pruned_loss=0.1299, over 4220.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2696, pruned_loss=0.07081, over 953421.54 frames. ], batch size: 66, lr: 3.87e-03, grad_scale: 32.0 +2023-04-26 19:49:55,704 INFO [optim.py:369] (2/7) 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 19:50:28,528 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 19:50:37,111 INFO [finetune.py:976] (2/7) Epoch 7, batch 3500, loss[loss=0.1726, simple_loss=0.2331, pruned_loss=0.05604, over 4821.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2667, pruned_loss=0.07003, over 952736.42 frames. ], batch size: 38, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:51:17,010 INFO [finetune.py:976] (2/7) Epoch 7, batch 3550, loss[loss=0.1445, simple_loss=0.2109, pruned_loss=0.03908, over 4833.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2637, pruned_loss=0.06887, over 953119.44 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:51:25,314 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 19:51:27,517 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37924.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:51:36,394 INFO [optim.py:369] (2/7) 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 19:51:38,853 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0745, 2.1257, 2.0023, 1.8323, 2.2318, 1.6302, 2.9308, 1.5728], + device='cuda:2'), covar=tensor([0.4067, 0.1877, 0.4413, 0.3505, 0.1962, 0.2947, 0.1457, 0.4965], + device='cuda:2'), in_proj_covar=tensor([0.0347, 0.0352, 0.0432, 0.0362, 0.0389, 0.0382, 0.0384, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:51:39,450 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5713, 3.0975, 1.0034, 1.5522, 2.2069, 1.6270, 4.3471, 2.0557], + device='cuda:2'), covar=tensor([0.0644, 0.0841, 0.0925, 0.1394, 0.0579, 0.1006, 0.0240, 0.0669], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0052, 0.0053, 0.0080, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 19:52:00,860 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0729, 2.9718, 2.0101, 2.0109, 1.4507, 1.5001, 2.1576, 1.3801], + device='cuda:2'), covar=tensor([0.1769, 0.1505, 0.1613, 0.1982, 0.2664, 0.2129, 0.1167, 0.2308], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0220, 0.0175, 0.0207, 0.0210, 0.0187, 0.0166, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 19:52:01,352 INFO [finetune.py:976] (2/7) Epoch 7, batch 3600, loss[loss=0.1755, simple_loss=0.2399, pruned_loss=0.05552, over 4821.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2608, pruned_loss=0.06782, over 954079.49 frames. ], batch size: 30, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:52:05,029 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37972.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:52:09,414 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37979.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:52:18,257 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9837, 2.4249, 1.0298, 1.3274, 1.7973, 1.2610, 3.2307, 1.6965], + device='cuda:2'), covar=tensor([0.0697, 0.0689, 0.0797, 0.1288, 0.0539, 0.1009, 0.0215, 0.0647], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 19:52:21,706 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2653, 3.0093, 0.9489, 1.6441, 1.6219, 2.1221, 1.7657, 0.9829], + device='cuda:2'), covar=tensor([0.1468, 0.1079, 0.1833, 0.1394, 0.1193, 0.0999, 0.1517, 0.1874], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0255, 0.0143, 0.0125, 0.0136, 0.0156, 0.0120, 0.0123], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 19:52:30,654 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3786, 1.2298, 1.6355, 1.6463, 1.3121, 1.1222, 1.3466, 0.8886], + device='cuda:2'), covar=tensor([0.0776, 0.0979, 0.0556, 0.0736, 0.0888, 0.1563, 0.0762, 0.1031], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0074, 0.0072, 0.0067, 0.0076, 0.0095, 0.0080, 0.0076], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 19:52:35,876 INFO [finetune.py:976] (2/7) Epoch 7, batch 3650, loss[loss=0.1981, simple_loss=0.2609, pruned_loss=0.06765, over 4769.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2622, pruned_loss=0.06893, over 952490.57 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:52:38,360 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 19:52:51,018 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38040.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:52:51,613 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0870, 2.5607, 0.9453, 1.3256, 1.8012, 1.2555, 3.5665, 1.7765], + device='cuda:2'), covar=tensor([0.0703, 0.0748, 0.0891, 0.1387, 0.0605, 0.1128, 0.0282, 0.0673], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0069, 0.0051, 0.0048, 0.0053, 0.0054, 0.0080, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-26 19:53:08,983 INFO [finetune.py:976] (2/7) Epoch 7, batch 3700, loss[loss=0.1997, simple_loss=0.2705, pruned_loss=0.06439, over 4924.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2661, pruned_loss=0.06982, over 953066.75 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:53:10,240 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38068.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:53:41,668 INFO [finetune.py:976] (2/7) Epoch 7, batch 3750, loss[loss=0.1695, simple_loss=0.2254, pruned_loss=0.05681, over 4705.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2663, pruned_loss=0.06969, over 952839.24 frames. ], batch size: 23, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:53:50,668 INFO [optim.py:369] (2/7) 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 19:54:31,493 INFO [finetune.py:976] (2/7) Epoch 7, batch 3800, loss[loss=0.1909, simple_loss=0.2388, pruned_loss=0.07148, over 3786.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2688, pruned_loss=0.07104, over 954253.82 frames. ], batch size: 16, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:55:36,933 INFO [finetune.py:976] (2/7) Epoch 7, batch 3850, loss[loss=0.1577, simple_loss=0.2102, pruned_loss=0.05255, over 4087.00 frames. ], tot_loss[loss=0.2036, simple_loss=0.2676, pruned_loss=0.06977, over 954041.61 frames. ], batch size: 17, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:55:56,090 INFO [optim.py:369] (2/7) 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 19:56:41,864 INFO [finetune.py:976] (2/7) Epoch 7, batch 3900, loss[loss=0.1755, simple_loss=0.2465, pruned_loss=0.05222, over 4824.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2644, pruned_loss=0.06868, over 953906.05 frames. ], batch size: 38, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:57:45,016 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9502, 1.9151, 2.2329, 2.4314, 2.4027, 1.9403, 1.7598, 2.0653], + device='cuda:2'), covar=tensor([0.0956, 0.1078, 0.0567, 0.0664, 0.0618, 0.0911, 0.0821, 0.0621], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0205, 0.0182, 0.0178, 0.0180, 0.0193, 0.0162, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:57:45,597 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9689, 3.8383, 2.8849, 4.5808, 3.9734, 4.0129, 1.6415, 3.9463], + device='cuda:2'), covar=tensor([0.1765, 0.1322, 0.3254, 0.1396, 0.3483, 0.1878, 0.5965, 0.2285], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0220, 0.0253, 0.0310, 0.0305, 0.0255, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 19:57:47,850 INFO [finetune.py:976] (2/7) Epoch 7, batch 3950, loss[loss=0.1696, simple_loss=0.2363, pruned_loss=0.05143, over 4760.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2617, pruned_loss=0.06813, over 955442.71 frames. ], batch size: 27, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:58:09,819 INFO [optim.py:369] (2/7) 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 19:58:18,668 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38335.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:58:43,702 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38360.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 19:58:51,440 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3648, 1.6357, 2.1036, 2.8145, 2.8185, 2.2375, 1.8935, 2.3025], + device='cuda:2'), covar=tensor([0.1098, 0.1664, 0.0908, 0.0744, 0.0711, 0.1167, 0.1076, 0.0867], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0205, 0.0182, 0.0178, 0.0180, 0.0194, 0.0162, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 19:58:53,666 INFO [finetune.py:976] (2/7) Epoch 7, batch 4000, loss[loss=0.2053, simple_loss=0.2737, pruned_loss=0.06841, over 4906.00 frames. ], tot_loss[loss=0.1989, simple_loss=0.2614, pruned_loss=0.06826, over 955072.80 frames. ], batch size: 35, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 19:59:55,621 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 20:00:00,679 INFO [finetune.py:976] (2/7) Epoch 7, batch 4050, loss[loss=0.2427, simple_loss=0.3089, pruned_loss=0.08822, over 4812.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2649, pruned_loss=0.06949, over 952354.86 frames. ], batch size: 51, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:00:10,074 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38421.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:00:21,448 INFO [optim.py:369] (2/7) 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 20:01:01,406 INFO [finetune.py:976] (2/7) Epoch 7, batch 4100, loss[loss=0.1881, simple_loss=0.2533, pruned_loss=0.06141, over 4744.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2661, pruned_loss=0.06945, over 951752.09 frames. ], batch size: 27, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:01:18,498 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 20:02:13,353 INFO [finetune.py:976] (2/7) Epoch 7, batch 4150, loss[loss=0.162, simple_loss=0.2286, pruned_loss=0.04764, over 4738.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2674, pruned_loss=0.06958, over 952342.25 frames. ], batch size: 27, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:02:28,364 INFO [optim.py:369] (2/7) 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 20:02:52,791 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-26 20:02:57,367 INFO [finetune.py:976] (2/7) Epoch 7, batch 4200, loss[loss=0.2137, simple_loss=0.2755, pruned_loss=0.07591, over 4776.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2674, pruned_loss=0.06935, over 952731.59 frames. ], batch size: 45, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:02:58,739 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7644, 2.0022, 1.8355, 2.0534, 1.9077, 2.0868, 1.9477, 1.8647], + device='cuda:2'), covar=tensor([0.6086, 0.9736, 0.8705, 0.7286, 0.8504, 1.1849, 1.0174, 0.9265], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0392, 0.0316, 0.0327, 0.0342, 0.0407, 0.0370, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:03:29,907 INFO [finetune.py:976] (2/7) Epoch 7, batch 4250, loss[loss=0.1892, simple_loss=0.2498, pruned_loss=0.06428, over 4817.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2651, pruned_loss=0.06818, over 953058.84 frames. ], batch size: 30, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:03:35,994 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-26 20:03:40,411 INFO [optim.py:369] (2/7) 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 20:03:42,860 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8304, 1.9439, 1.8883, 1.4749, 2.1711, 1.6293, 2.6125, 1.6335], + device='cuda:2'), covar=tensor([0.3658, 0.1576, 0.4722, 0.3000, 0.1389, 0.2486, 0.1342, 0.4463], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0353, 0.0432, 0.0363, 0.0389, 0.0382, 0.0383, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:03:44,043 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38635.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:04:03,668 INFO [finetune.py:976] (2/7) Epoch 7, batch 4300, loss[loss=0.2236, simple_loss=0.2738, pruned_loss=0.08666, over 4224.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2624, pruned_loss=0.06755, over 953919.73 frames. ], batch size: 65, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:04:26,427 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38683.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:05:03,900 INFO [finetune.py:976] (2/7) Epoch 7, batch 4350, loss[loss=0.1723, simple_loss=0.2338, pruned_loss=0.05543, over 4778.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2591, pruned_loss=0.06664, over 951940.73 frames. ], batch size: 29, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:05:08,718 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38716.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:05:18,157 INFO [optim.py:369] (2/7) 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 20:05:48,690 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 20:05:52,729 INFO [finetune.py:976] (2/7) Epoch 7, batch 4400, loss[loss=0.2032, simple_loss=0.2724, pruned_loss=0.06697, over 4929.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2599, pruned_loss=0.06649, over 955534.61 frames. ], batch size: 38, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:06:12,852 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6998, 3.6125, 2.7366, 4.2844, 3.7520, 3.6972, 1.8178, 3.5072], + device='cuda:2'), covar=tensor([0.1848, 0.1394, 0.3487, 0.1778, 0.3825, 0.2135, 0.5770, 0.2755], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0219, 0.0254, 0.0311, 0.0304, 0.0255, 0.0276, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:06:23,893 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-04-26 20:06:26,655 INFO [finetune.py:976] (2/7) Epoch 7, batch 4450, loss[loss=0.2044, simple_loss=0.2785, pruned_loss=0.06512, over 4924.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2644, pruned_loss=0.06812, over 955057.92 frames. ], batch size: 36, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:06:36,508 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 20:07:06,742 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9151, 3.7213, 2.8529, 4.4715, 3.8876, 3.9209, 1.7714, 3.7756], + device='cuda:2'), covar=tensor([0.1539, 0.1367, 0.2802, 0.1523, 0.3193, 0.1798, 0.5736, 0.2390], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0218, 0.0254, 0.0310, 0.0304, 0.0254, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:07:37,018 INFO [finetune.py:976] (2/7) Epoch 7, batch 4500, loss[loss=0.206, simple_loss=0.2665, pruned_loss=0.07276, over 4894.00 frames. ], tot_loss[loss=0.2024, simple_loss=0.2666, pruned_loss=0.06916, over 955542.27 frames. ], batch size: 32, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:07:59,562 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38884.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:08:12,772 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7028, 3.6635, 2.7949, 4.2753, 3.7060, 3.6523, 1.7382, 3.6290], + device='cuda:2'), covar=tensor([0.1505, 0.1182, 0.3218, 0.1357, 0.2415, 0.1758, 0.5408, 0.2176], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0252, 0.0310, 0.0302, 0.0253, 0.0274, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:08:44,164 INFO [finetune.py:976] (2/7) Epoch 7, batch 4550, loss[loss=0.2002, simple_loss=0.2739, pruned_loss=0.0633, over 4890.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2686, pruned_loss=0.06958, over 956359.28 frames. ], batch size: 32, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:08:58,069 INFO [optim.py:369] (2/7) 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 20:09:04,823 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2791, 1.5454, 1.3574, 1.6555, 1.5544, 1.8900, 1.3141, 3.3023], + device='cuda:2'), covar=tensor([0.0645, 0.0746, 0.0782, 0.1175, 0.0623, 0.0613, 0.0767, 0.0166], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 20:09:50,305 INFO [finetune.py:976] (2/7) Epoch 7, batch 4600, loss[loss=0.1713, simple_loss=0.2432, pruned_loss=0.04971, over 4753.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2676, pruned_loss=0.06917, over 956721.65 frames. ], batch size: 27, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:10:11,629 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3906, 1.6061, 1.4647, 1.7969, 1.6964, 1.8289, 1.4959, 3.2083], + device='cuda:2'), covar=tensor([0.0625, 0.0742, 0.0760, 0.1073, 0.0583, 0.0735, 0.0751, 0.0192], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 20:10:32,150 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7361, 1.2336, 1.3832, 1.2639, 1.8777, 1.4800, 1.1921, 1.3324], + device='cuda:2'), covar=tensor([0.1524, 0.1432, 0.2399, 0.1385, 0.0873, 0.1408, 0.2038, 0.2057], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0327, 0.0355, 0.0302, 0.0339, 0.0326, 0.0306, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.5225e-05, 6.9484e-05, 7.6951e-05, 6.2704e-05, 7.1404e-05, 7.0156e-05, + 6.6245e-05, 7.5974e-05], device='cuda:2') +2023-04-26 20:10:54,977 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1043, 1.8213, 2.1585, 2.4853, 2.5791, 2.0320, 1.7255, 2.1203], + device='cuda:2'), covar=tensor([0.0917, 0.1159, 0.0664, 0.0623, 0.0551, 0.0952, 0.0896, 0.0613], + device='cuda:2'), in_proj_covar=tensor([0.0201, 0.0206, 0.0184, 0.0180, 0.0180, 0.0195, 0.0164, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:10:56,079 INFO [finetune.py:976] (2/7) Epoch 7, batch 4650, loss[loss=0.2212, simple_loss=0.2703, pruned_loss=0.08604, over 4825.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2652, pruned_loss=0.06864, over 955518.62 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:10:56,168 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39016.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:11:16,118 INFO [optim.py:369] (2/7) 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 20:12:02,801 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 4700, loss[loss=0.2553, simple_loss=0.2982, pruned_loss=0.1063, over 4711.00 frames. ], tot_loss[loss=0.1988, simple_loss=0.2621, pruned_loss=0.06779, over 954587.98 frames. ], batch size: 23, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:12:24,018 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5733, 1.8609, 2.3609, 2.8977, 2.2018, 1.8785, 1.7437, 2.0904], + device='cuda:2'), covar=tensor([0.3993, 0.4457, 0.1904, 0.3659, 0.3824, 0.3455, 0.5217, 0.3357], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0255, 0.0220, 0.0325, 0.0215, 0.0230, 0.0239, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 20:13:04,271 INFO [finetune.py:976] (2/7) Epoch 7, batch 4750, loss[loss=0.1796, simple_loss=0.2408, pruned_loss=0.05922, over 4746.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2601, pruned_loss=0.06768, over 954855.51 frames. ], batch size: 23, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:13:24,487 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 7, batch 4800, loss[loss=0.2089, simple_loss=0.2745, pruned_loss=0.07168, over 4787.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2631, pruned_loss=0.06846, over 955151.36 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 32.0 +2023-04-26 20:14:17,548 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6209, 1.2639, 4.3581, 3.8578, 3.8231, 3.9118, 3.9916, 3.7030], + device='cuda:2'), covar=tensor([0.8585, 0.8060, 0.1400, 0.2774, 0.2018, 0.3285, 0.2134, 0.2535], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0306, 0.0408, 0.0413, 0.0349, 0.0405, 0.0317, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:14:31,378 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39179.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:14:40,467 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39185.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:14:42,263 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5771, 1.2399, 1.6979, 1.9634, 1.6690, 1.5587, 1.6240, 1.6817], + device='cuda:2'), covar=tensor([0.7796, 1.0535, 1.0921, 1.0923, 0.9339, 1.2553, 1.2467, 1.1488], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0432, 0.0514, 0.0536, 0.0441, 0.0459, 0.0473, 0.0469], + device='cuda:2'), out_proj_covar=tensor([9.9960e-05, 1.0699e-04, 1.1594e-04, 1.2698e-04, 1.0732e-04, 1.1127e-04, + 1.1382e-04, 1.1421e-04], device='cuda:2') +2023-04-26 20:14:59,992 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5446, 1.1387, 0.4247, 1.1768, 1.2238, 1.3887, 1.2573, 1.2437], + device='cuda:2'), covar=tensor([0.0598, 0.0471, 0.0517, 0.0630, 0.0330, 0.0588, 0.0590, 0.0661], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:2') +2023-04-26 20:15:12,277 INFO [finetune.py:976] (2/7) Epoch 7, batch 4850, loss[loss=0.2055, simple_loss=0.2718, pruned_loss=0.06959, over 4816.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2656, pruned_loss=0.0688, over 951281.20 frames. ], batch size: 51, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:15:21,783 INFO [optim.py:369] (2/7) 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 20:15:25,637 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.08 vs. limit=5.0 +2023-04-26 20:15:31,636 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39246.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:15:45,481 INFO [finetune.py:976] (2/7) Epoch 7, batch 4900, loss[loss=0.1987, simple_loss=0.2696, pruned_loss=0.06392, over 4903.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2665, pruned_loss=0.06956, over 949637.06 frames. ], batch size: 43, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:16:09,153 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39284.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:16:31,481 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1728, 1.4709, 1.3058, 1.6119, 1.5470, 1.6856, 1.3514, 3.0478], + device='cuda:2'), covar=tensor([0.0649, 0.0817, 0.0823, 0.1276, 0.0660, 0.0544, 0.0765, 0.0171], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0040, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 20:16:50,964 INFO [finetune.py:976] (2/7) Epoch 7, batch 4950, loss[loss=0.1847, simple_loss=0.2591, pruned_loss=0.05519, over 4781.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2687, pruned_loss=0.07008, over 951782.40 frames. ], batch size: 51, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:17:06,094 INFO [optim.py:369] (2/7) 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 20:17:15,564 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7131, 1.4566, 1.9781, 1.9760, 1.5148, 1.2559, 1.6579, 1.0206], + device='cuda:2'), covar=tensor([0.0634, 0.1159, 0.0505, 0.0852, 0.0868, 0.1525, 0.0821, 0.1034], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0074, 0.0072, 0.0067, 0.0077, 0.0096, 0.0080, 0.0076], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 20:17:25,334 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 5000, loss[loss=0.164, simple_loss=0.2333, pruned_loss=0.04733, over 4876.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2669, pruned_loss=0.06939, over 953392.50 frames. ], batch size: 32, lr: 3.86e-03, grad_scale: 64.0 +2023-04-26 20:17:50,486 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7071, 1.2963, 1.2823, 1.5433, 2.0167, 1.5603, 1.3597, 1.2943], + device='cuda:2'), covar=tensor([0.1842, 0.1763, 0.2140, 0.1371, 0.0834, 0.1721, 0.2429, 0.2208], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0326, 0.0354, 0.0300, 0.0339, 0.0326, 0.0307, 0.0353], + device='cuda:2'), out_proj_covar=tensor([6.5171e-05, 6.9240e-05, 7.6655e-05, 6.2174e-05, 7.1339e-05, 7.0203e-05, + 6.6274e-05, 7.5908e-05], device='cuda:2') +2023-04-26 20:17:54,066 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5847, 1.9090, 2.3271, 3.0035, 2.3036, 1.8776, 1.5360, 2.2553], + device='cuda:2'), covar=tensor([0.3867, 0.4107, 0.2041, 0.3424, 0.3897, 0.3170, 0.5172, 0.2839], + device='cuda:2'), in_proj_covar=tensor([0.0279, 0.0253, 0.0219, 0.0322, 0.0214, 0.0229, 0.0237, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 20:18:22,935 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4597, 2.4806, 2.7693, 2.9600, 2.8527, 2.3857, 1.9115, 2.5148], + device='cuda:2'), covar=tensor([0.0900, 0.0861, 0.0490, 0.0600, 0.0598, 0.0965, 0.0852, 0.0581], + device='cuda:2'), in_proj_covar=tensor([0.0201, 0.0205, 0.0183, 0.0180, 0.0180, 0.0195, 0.0164, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:18:23,437 INFO [finetune.py:976] (2/7) Epoch 7, batch 5050, loss[loss=0.1685, simple_loss=0.2369, pruned_loss=0.05004, over 4908.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2632, pruned_loss=0.06832, over 952910.47 frames. ], batch size: 36, lr: 3.85e-03, grad_scale: 64.0 +2023-04-26 20:18:33,400 INFO [optim.py:369] (2/7) 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 20:18:54,467 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39463.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:18:56,691 INFO [finetune.py:976] (2/7) Epoch 7, batch 5100, loss[loss=0.2104, simple_loss=0.2698, pruned_loss=0.0755, over 4925.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2595, pruned_loss=0.06678, over 952433.79 frames. ], batch size: 36, lr: 3.85e-03, grad_scale: 64.0 +2023-04-26 20:19:05,634 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39479.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:19:11,839 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7218, 2.3014, 1.7528, 1.5739, 1.2715, 1.3336, 1.7815, 1.2240], + device='cuda:2'), covar=tensor([0.1747, 0.1563, 0.1657, 0.2166, 0.2705, 0.2148, 0.1187, 0.2270], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0220, 0.0174, 0.0206, 0.0209, 0.0187, 0.0165, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 20:19:29,889 INFO [finetune.py:976] (2/7) Epoch 7, batch 5150, loss[loss=0.2266, simple_loss=0.2869, pruned_loss=0.08313, over 4868.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2606, pruned_loss=0.06765, over 953033.90 frames. ], batch size: 34, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:19:35,820 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39527.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:19:40,054 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 20:19:40,452 INFO [optim.py:369] (2/7) 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 20:19:47,586 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39541.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:19:55,415 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5663, 3.5237, 2.7482, 4.1867, 3.5693, 3.5806, 1.6264, 3.5196], + device='cuda:2'), covar=tensor([0.1807, 0.1364, 0.3444, 0.1579, 0.2995, 0.2005, 0.5639, 0.2389], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0219, 0.0253, 0.0310, 0.0303, 0.0255, 0.0276, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:20:03,309 INFO [finetune.py:976] (2/7) Epoch 7, batch 5200, loss[loss=0.2103, simple_loss=0.2797, pruned_loss=0.07045, over 4816.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.265, pruned_loss=0.06882, over 952315.90 frames. ], batch size: 39, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:20:37,216 INFO [finetune.py:976] (2/7) Epoch 7, batch 5250, loss[loss=0.1778, simple_loss=0.2598, pruned_loss=0.04789, over 4755.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2676, pruned_loss=0.06928, over 953908.69 frames. ], batch size: 28, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:20:47,796 INFO [optim.py:369] (2/7) 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 20:20:53,192 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39639.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:20:53,791 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39640.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:21:09,390 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.02 vs. limit=5.0 +2023-04-26 20:21:10,370 INFO [finetune.py:976] (2/7) Epoch 7, batch 5300, loss[loss=0.1577, simple_loss=0.2409, pruned_loss=0.03724, over 4822.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2688, pruned_loss=0.06982, over 953440.39 frames. ], batch size: 47, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:21:14,226 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 20:21:50,864 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39700.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 20:22:11,716 INFO [finetune.py:976] (2/7) Epoch 7, batch 5350, loss[loss=0.2149, simple_loss=0.2927, pruned_loss=0.06856, over 4912.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2684, pruned_loss=0.06913, over 954295.46 frames. ], batch size: 33, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:22:31,529 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 7, batch 5400, loss[loss=0.1695, simple_loss=0.2271, pruned_loss=0.05591, over 4024.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2653, pruned_loss=0.0683, over 953291.52 frames. ], batch size: 17, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:23:17,750 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39767.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:24:18,140 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39815.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:24:18,617 INFO [finetune.py:976] (2/7) Epoch 7, batch 5450, loss[loss=0.2084, simple_loss=0.2681, pruned_loss=0.07433, over 4780.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2631, pruned_loss=0.06788, over 953833.55 frames. ], batch size: 26, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:24:20,494 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39819.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:24:30,443 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6955, 1.0887, 1.5699, 2.0567, 1.7902, 1.5809, 1.5534, 1.6608], + device='cuda:2'), covar=tensor([0.7203, 0.9426, 0.9283, 1.0375, 0.8505, 1.1213, 1.1680, 0.9294], + device='cuda:2'), in_proj_covar=tensor([0.0410, 0.0428, 0.0513, 0.0533, 0.0439, 0.0457, 0.0470, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9549e-05, 1.0623e-04, 1.1562e-04, 1.2640e-04, 1.0676e-04, 1.1069e-04, + 1.1310e-04, 1.1364e-04], device='cuda:2') +2023-04-26 20:24:31,592 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 20:24:40,439 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 5500, loss[loss=0.1869, simple_loss=0.2497, pruned_loss=0.06208, over 4833.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2603, pruned_loss=0.06733, over 954938.82 frames. ], batch size: 33, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:25:04,133 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39876.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:25:12,054 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 7, batch 5550, loss[loss=0.1844, simple_loss=0.2604, pruned_loss=0.05424, over 4795.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2605, pruned_loss=0.06734, over 954458.10 frames. ], batch size: 51, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:25:31,878 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8969, 2.7286, 1.9144, 1.9437, 1.4264, 1.4153, 2.0444, 1.3565], + device='cuda:2'), covar=tensor([0.1661, 0.1571, 0.1575, 0.2013, 0.2576, 0.2068, 0.1112, 0.2201], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0219, 0.0174, 0.0206, 0.0208, 0.0187, 0.0164, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 20:25:40,910 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:25:48,811 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-26 20:26:02,973 INFO [finetune.py:976] (2/7) Epoch 7, batch 5600, loss[loss=0.2196, simple_loss=0.2766, pruned_loss=0.08129, over 4761.00 frames. ], tot_loss[loss=0.2015, simple_loss=0.2652, pruned_loss=0.06887, over 956462.02 frames. ], batch size: 26, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:26:14,988 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-26 20:26:15,867 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39988.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:26:19,991 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39995.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:26:31,038 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6647, 1.2853, 1.7008, 2.0440, 1.7676, 1.6072, 1.6732, 1.7160], + device='cuda:2'), covar=tensor([0.7508, 1.0010, 0.9943, 1.0664, 0.9053, 1.1984, 1.2206, 1.0381], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0426, 0.0512, 0.0533, 0.0438, 0.0458, 0.0469, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9256e-05, 1.0573e-04, 1.1551e-04, 1.2640e-04, 1.0662e-04, 1.1083e-04, + 1.1288e-04, 1.1355e-04], device='cuda:2') +2023-04-26 20:26:38,931 INFO [finetune.py:976] (2/7) Epoch 7, batch 5650, loss[loss=0.231, simple_loss=0.2887, pruned_loss=0.08662, over 4891.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2669, pruned_loss=0.06873, over 956756.38 frames. ], batch size: 32, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:26:53,974 INFO [optim.py:369] (2/7) 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 20:27:20,303 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2482, 2.6839, 1.5863, 2.1010, 2.7695, 2.1168, 2.0759, 2.2053], + device='cuda:2'), covar=tensor([0.0451, 0.0319, 0.0307, 0.0517, 0.0210, 0.0504, 0.0501, 0.0540], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 20:27:27,070 INFO [finetune.py:976] (2/7) Epoch 7, batch 5700, loss[loss=0.1852, simple_loss=0.2342, pruned_loss=0.06809, over 4161.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.263, pruned_loss=0.068, over 940041.67 frames. ], batch size: 18, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:27:48,126 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-26 20:28:14,052 INFO [finetune.py:976] (2/7) Epoch 8, batch 0, loss[loss=0.2095, simple_loss=0.2783, pruned_loss=0.07039, over 4813.00 frames. ], tot_loss[loss=0.2095, simple_loss=0.2783, pruned_loss=0.07039, over 4813.00 frames. ], batch size: 38, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:28:14,053 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 20:28:25,065 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3435, 1.1909, 1.5672, 1.4761, 1.2335, 1.0517, 1.2808, 0.8806], + device='cuda:2'), covar=tensor([0.0696, 0.0740, 0.0594, 0.0689, 0.0867, 0.1181, 0.0674, 0.0799], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0074, 0.0073, 0.0067, 0.0077, 0.0096, 0.0080, 0.0075], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 20:28:30,560 INFO [finetune.py:1010] (2/7) Epoch 8, validation: loss=0.1574, simple_loss=0.2299, pruned_loss=0.04247, over 2265189.00 frames. +2023-04-26 20:28:30,560 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 20:28:44,874 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6516, 1.7409, 0.9053, 1.4005, 1.8908, 1.5085, 1.4150, 1.5446], + device='cuda:2'), covar=tensor([0.0532, 0.0395, 0.0377, 0.0598, 0.0271, 0.0579, 0.0519, 0.0627], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:2') +2023-04-26 20:29:03,013 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40119.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:03,223 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-26 20:29:05,430 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40123.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:10,789 INFO [optim.py:369] (2/7) 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 20:29:20,984 INFO [finetune.py:976] (2/7) Epoch 8, batch 50, loss[loss=0.1889, simple_loss=0.2526, pruned_loss=0.06256, over 4801.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2684, pruned_loss=0.06871, over 217427.09 frames. ], batch size: 51, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:29:24,041 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5406, 1.4830, 0.8427, 1.2928, 1.6714, 1.4138, 1.3514, 1.3626], + device='cuda:2'), covar=tensor([0.0522, 0.0370, 0.0404, 0.0548, 0.0302, 0.0520, 0.0501, 0.0604], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 20:29:36,059 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40167.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:38,549 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40171.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:29:41,239 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-26 20:29:42,172 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0126, 3.9709, 2.8192, 4.6340, 4.1400, 3.9724, 1.7449, 3.8889], + device='cuda:2'), covar=tensor([0.1907, 0.1032, 0.3294, 0.1244, 0.2269, 0.1886, 0.6243, 0.2234], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0218, 0.0253, 0.0310, 0.0303, 0.0253, 0.0274, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:29:54,375 INFO [finetune.py:976] (2/7) Epoch 8, batch 100, loss[loss=0.1552, simple_loss=0.2243, pruned_loss=0.04304, over 4788.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2605, pruned_loss=0.06639, over 380606.64 frames. ], batch size: 29, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:29:55,590 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40195.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:01,242 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9738, 1.4419, 1.5389, 1.6151, 2.1639, 1.7142, 1.4960, 1.5557], + device='cuda:2'), covar=tensor([0.1642, 0.1999, 0.2361, 0.1351, 0.1017, 0.2137, 0.2503, 0.2274], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0328, 0.0357, 0.0303, 0.0341, 0.0328, 0.0309, 0.0354], + device='cuda:2'), 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:2') +2023-04-26 20:30:10,505 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40218.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:18,313 INFO [optim.py:369] (2/7) 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 20:30:28,096 INFO [finetune.py:976] (2/7) Epoch 8, batch 150, loss[loss=0.165, simple_loss=0.239, pruned_loss=0.04544, over 4815.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2578, pruned_loss=0.06616, over 508169.42 frames. ], batch size: 38, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:30:36,433 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40256.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:46,173 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40272.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:30:50,426 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 200, loss[loss=0.1752, simple_loss=0.228, pruned_loss=0.06122, over 4726.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2589, pruned_loss=0.0678, over 606680.20 frames. ], batch size: 23, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:31:02,420 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40295.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 20:31:24,757 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 20:31:26,598 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 8, batch 250, loss[loss=0.2006, simple_loss=0.2796, pruned_loss=0.06085, over 4828.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2612, pruned_loss=0.06805, over 683875.73 frames. ], batch size: 47, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:31:44,556 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6593, 1.2504, 1.7482, 2.0604, 1.7818, 1.6033, 1.6364, 1.7322], + device='cuda:2'), covar=tensor([0.6541, 0.9523, 0.9019, 0.9554, 0.7594, 1.1011, 1.1048, 0.9755], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0424, 0.0511, 0.0532, 0.0439, 0.0457, 0.0468, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9165e-05, 1.0525e-04, 1.1526e-04, 1.2613e-04, 1.0665e-04, 1.1074e-04, + 1.1269e-04, 1.1359e-04], device='cuda:2') +2023-04-26 20:31:49,977 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1637, 1.2387, 1.3371, 1.5788, 1.6329, 1.2208, 0.9626, 1.3659], + device='cuda:2'), covar=tensor([0.0920, 0.1283, 0.0859, 0.0659, 0.0603, 0.0930, 0.0921, 0.0672], + device='cuda:2'), in_proj_covar=tensor([0.0201, 0.0208, 0.0184, 0.0181, 0.0181, 0.0196, 0.0164, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:31:57,220 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-04-26 20:32:05,768 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40391.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:32:07,908 INFO [finetune.py:976] (2/7) Epoch 8, batch 300, loss[loss=0.2347, simple_loss=0.2996, pruned_loss=0.08491, over 4823.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2646, pruned_loss=0.06845, over 742494.03 frames. ], batch size: 40, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:32:27,218 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 8, batch 350, loss[loss=0.1761, simple_loss=0.2346, pruned_loss=0.05882, over 4771.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2665, pruned_loss=0.06956, over 791276.44 frames. ], batch size: 26, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:33:27,711 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:33:27,740 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:33:37,662 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3862, 1.6321, 1.4719, 1.8975, 1.7831, 1.8584, 1.4684, 3.8476], + device='cuda:2'), covar=tensor([0.0572, 0.0677, 0.0770, 0.1102, 0.0564, 0.0646, 0.0703, 0.0162], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0040, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 20:33:39,481 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8026, 2.8329, 2.2589, 3.2650, 2.8382, 2.8490, 1.1697, 2.7486], + device='cuda:2'), covar=tensor([0.2416, 0.1801, 0.3642, 0.3162, 0.3439, 0.2524, 0.5973, 0.3189], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0218, 0.0253, 0.0309, 0.0303, 0.0254, 0.0274, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:33:52,003 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 400, loss[loss=0.2201, simple_loss=0.2823, pruned_loss=0.07898, over 4756.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2677, pruned_loss=0.0701, over 827006.96 frames. ], batch size: 59, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:34:20,604 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0196, 1.9026, 2.2883, 2.5980, 2.5974, 2.1379, 1.7808, 2.1245], + device='cuda:2'), covar=tensor([0.1080, 0.1139, 0.0621, 0.0662, 0.0571, 0.0866, 0.0859, 0.0738], + device='cuda:2'), in_proj_covar=tensor([0.0203, 0.0210, 0.0186, 0.0183, 0.0182, 0.0197, 0.0166, 0.0191], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:34:32,789 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40519.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:34:45,208 INFO [optim.py:369] (2/7) 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 20:35:05,026 INFO [finetune.py:976] (2/7) Epoch 8, batch 450, loss[loss=0.2062, simple_loss=0.2598, pruned_loss=0.07632, over 4855.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2655, pruned_loss=0.06912, over 854112.29 frames. ], batch size: 44, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:35:16,807 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40551.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:35:18,071 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 20:35:18,617 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40553.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:35:35,918 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2844, 1.6062, 2.1683, 2.6080, 2.0739, 1.6412, 1.4268, 1.9455], + device='cuda:2'), covar=tensor([0.3952, 0.4170, 0.2002, 0.3453, 0.3491, 0.3329, 0.5261, 0.2815], + device='cuda:2'), in_proj_covar=tensor([0.0278, 0.0252, 0.0218, 0.0321, 0.0213, 0.0228, 0.0236, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 20:35:38,784 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40574.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:35:56,422 INFO [finetune.py:976] (2/7) Epoch 8, batch 500, loss[loss=0.1745, simple_loss=0.2489, pruned_loss=0.05002, over 4873.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2627, pruned_loss=0.06796, over 877447.69 frames. ], batch size: 34, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:36:37,617 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40628.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:36:39,368 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 550, loss[loss=0.1884, simple_loss=0.2464, pruned_loss=0.06523, over 4900.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2596, pruned_loss=0.06665, over 895776.90 frames. ], batch size: 43, lr: 3.85e-03, grad_scale: 32.0 +2023-04-26 20:36:56,970 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40651.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:37:47,174 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40686.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:37:52,044 INFO [finetune.py:976] (2/7) Epoch 8, batch 600, loss[loss=0.2194, simple_loss=0.2915, pruned_loss=0.07362, over 4791.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2599, pruned_loss=0.06622, over 910065.98 frames. ], batch size: 29, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:38:21,432 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40712.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:38:22,134 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-04-26 20:38:44,482 INFO [optim.py:369] (2/7) 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 20:38:54,534 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6011, 1.4162, 1.8741, 1.8972, 1.4415, 1.1748, 1.6167, 0.9517], + device='cuda:2'), covar=tensor([0.0707, 0.1190, 0.0531, 0.0879, 0.1142, 0.1528, 0.0813, 0.1015], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0074, 0.0072, 0.0067, 0.0076, 0.0096, 0.0080, 0.0075], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 20:39:02,720 INFO [finetune.py:976] (2/7) Epoch 8, batch 650, loss[loss=0.2451, simple_loss=0.3153, pruned_loss=0.08749, over 4818.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.262, pruned_loss=0.06639, over 921444.73 frames. ], batch size: 40, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:39:13,999 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 20:39:36,225 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0020, 2.6348, 2.0935, 1.9213, 1.6478, 1.6904, 2.1116, 1.5900], + device='cuda:2'), covar=tensor([0.1574, 0.1487, 0.1549, 0.1788, 0.2324, 0.2001, 0.1090, 0.2010], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0217, 0.0173, 0.0205, 0.0206, 0.0186, 0.0163, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 20:40:08,309 INFO [finetune.py:976] (2/7) Epoch 8, batch 700, loss[loss=0.1375, simple_loss=0.2152, pruned_loss=0.02991, over 4756.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2643, pruned_loss=0.06751, over 924743.34 frames. ], batch size: 28, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:40:21,978 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-04-26 20:40:24,129 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9566, 1.8514, 2.1920, 2.3304, 1.7957, 1.4271, 1.9660, 1.0661], + device='cuda:2'), covar=tensor([0.0633, 0.0965, 0.0608, 0.0791, 0.0848, 0.1349, 0.0872, 0.1036], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0074, 0.0072, 0.0067, 0.0076, 0.0095, 0.0080, 0.0075], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 20:40:55,212 INFO [optim.py:369] (2/7) 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 20:41:14,344 INFO [finetune.py:976] (2/7) Epoch 8, batch 750, loss[loss=0.2307, simple_loss=0.288, pruned_loss=0.08668, over 4894.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2662, pruned_loss=0.06827, over 933231.58 frames. ], batch size: 32, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:41:16,854 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:17,636 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-26 20:41:24,328 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40851.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:38,680 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40865.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:44,668 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40874.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:41:57,696 INFO [finetune.py:976] (2/7) Epoch 8, batch 800, loss[loss=0.1779, simple_loss=0.2446, pruned_loss=0.05564, over 4826.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2652, pruned_loss=0.06727, over 937607.50 frames. ], batch size: 49, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:42:00,808 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40899.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:01,458 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40900.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:16,692 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40922.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:19,187 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40926.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:20,388 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40928.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:22,146 INFO [optim.py:369] (2/7) 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 20:42:31,005 INFO [finetune.py:976] (2/7) Epoch 8, batch 850, loss[loss=0.1654, simple_loss=0.2278, pruned_loss=0.05147, over 4856.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2625, pruned_loss=0.0664, over 943246.13 frames. ], batch size: 49, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:42:41,369 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40961.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:42:48,400 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 20:42:52,275 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40976.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:42:58,801 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40986.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:03,444 INFO [finetune.py:976] (2/7) Epoch 8, batch 900, loss[loss=0.1831, simple_loss=0.2404, pruned_loss=0.0629, over 4795.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2604, pruned_loss=0.06632, over 944432.50 frames. ], batch size: 29, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:43:05,767 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40997.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:12,066 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 20:43:30,875 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41034.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:35,078 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-26 20:43:37,319 INFO [finetune.py:976] (2/7) Epoch 8, batch 950, loss[loss=0.3066, simple_loss=0.3457, pruned_loss=0.1337, over 4288.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2593, pruned_loss=0.0665, over 945910.61 frames. ], batch size: 65, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:43:38,029 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41045.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:43:46,393 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 1000, loss[loss=0.2453, simple_loss=0.3119, pruned_loss=0.0894, over 4805.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2612, pruned_loss=0.06712, over 947754.22 frames. ], batch size: 41, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:44:16,009 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41102.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:18,378 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41106.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:23,861 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9718, 1.7047, 1.8838, 2.2325, 2.2405, 1.9376, 1.7142, 1.9503], + device='cuda:2'), covar=tensor([0.0789, 0.1063, 0.0622, 0.0557, 0.0631, 0.0779, 0.0807, 0.0623], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0206, 0.0183, 0.0180, 0.0179, 0.0193, 0.0163, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:44:35,825 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 1050, loss[loss=0.1734, simple_loss=0.2271, pruned_loss=0.05986, over 4346.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2648, pruned_loss=0.06819, over 949200.92 frames. ], batch size: 19, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:44:46,608 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41148.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:49,638 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6777, 2.0246, 1.6449, 1.8190, 1.5147, 1.5862, 1.5700, 1.3911], + device='cuda:2'), covar=tensor([0.2030, 0.1280, 0.1106, 0.1530, 0.3375, 0.1540, 0.2147, 0.2661], + device='cuda:2'), in_proj_covar=tensor([0.0303, 0.0323, 0.0234, 0.0296, 0.0320, 0.0275, 0.0263, 0.0287], + device='cuda:2'), out_proj_covar=tensor([1.2275e-04, 1.3062e-04, 9.4610e-05, 1.1864e-04, 1.3140e-04, 1.1121e-04, + 1.0799e-04, 1.1521e-04], device='cuda:2') +2023-04-26 20:44:54,941 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1619, 2.4252, 0.9855, 1.2897, 1.8394, 1.2732, 3.1238, 1.7130], + device='cuda:2'), covar=tensor([0.0594, 0.0585, 0.0705, 0.1232, 0.0458, 0.0929, 0.0341, 0.0616], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 20:44:56,158 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41163.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:44:56,742 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2868, 3.0179, 0.8247, 1.6844, 1.7047, 2.2036, 1.7966, 0.8870], + device='cuda:2'), covar=tensor([0.1414, 0.1050, 0.1930, 0.1229, 0.1070, 0.0959, 0.1371, 0.1781], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0254, 0.0142, 0.0124, 0.0136, 0.0155, 0.0120, 0.0123], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 20:45:27,977 INFO [finetune.py:976] (2/7) Epoch 8, batch 1100, loss[loss=0.2153, simple_loss=0.2694, pruned_loss=0.08058, over 4397.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.265, pruned_loss=0.06797, over 952109.30 frames. ], batch size: 19, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:45:29,273 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41196.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:45:31,994 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 20:45:45,865 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7034, 2.4060, 1.7117, 1.7856, 1.2151, 1.2963, 1.8380, 1.1862], + device='cuda:2'), covar=tensor([0.2138, 0.1620, 0.1897, 0.2112, 0.3007, 0.2658, 0.1235, 0.2500], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0217, 0.0172, 0.0204, 0.0206, 0.0185, 0.0163, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 20:45:50,755 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41221.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:45:53,186 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9249, 2.9395, 2.2828, 3.3373, 2.8652, 2.9460, 1.1951, 2.7631], + device='cuda:2'), covar=tensor([0.1968, 0.1424, 0.2986, 0.2617, 0.2814, 0.2056, 0.5756, 0.2888], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0220, 0.0253, 0.0311, 0.0304, 0.0255, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:45:58,234 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 1150, loss[loss=0.1964, simple_loss=0.2684, pruned_loss=0.06217, over 4831.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2665, pruned_loss=0.06801, over 954687.74 frames. ], batch size: 47, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:46:27,344 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 20:46:36,336 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41256.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:46:50,025 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-26 20:47:24,548 INFO [finetune.py:976] (2/7) Epoch 8, batch 1200, loss[loss=0.1514, simple_loss=0.2311, pruned_loss=0.03585, over 4766.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2646, pruned_loss=0.06761, over 953587.95 frames. ], batch size: 26, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:47:44,109 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41307.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:47:59,149 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 1250, loss[loss=0.1704, simple_loss=0.2362, pruned_loss=0.05233, over 4854.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.261, pruned_loss=0.066, over 954966.72 frames. ], batch size: 47, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:48:14,611 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41353.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:48:15,822 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41355.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:48:36,662 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1568, 2.7603, 2.3305, 2.4808, 1.9509, 2.1417, 2.2666, 1.8673], + device='cuda:2'), covar=tensor([0.2233, 0.1248, 0.0878, 0.1340, 0.3163, 0.1454, 0.2402, 0.2918], + device='cuda:2'), in_proj_covar=tensor([0.0301, 0.0321, 0.0233, 0.0294, 0.0319, 0.0273, 0.0262, 0.0285], + device='cuda:2'), out_proj_covar=tensor([1.2208e-04, 1.2957e-04, 9.4091e-05, 1.1761e-04, 1.3090e-04, 1.1030e-04, + 1.0727e-04, 1.1450e-04], device='cuda:2') +2023-04-26 20:48:42,247 INFO [finetune.py:976] (2/7) Epoch 8, batch 1300, loss[loss=0.2062, simple_loss=0.2701, pruned_loss=0.07109, over 4868.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2575, pruned_loss=0.06447, over 954587.08 frames. ], batch size: 34, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:48:46,732 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41401.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:49:05,847 INFO [optim.py:369] (2/7) 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 20:49:15,122 INFO [finetune.py:976] (2/7) Epoch 8, batch 1350, loss[loss=0.2469, simple_loss=0.3014, pruned_loss=0.09618, over 4794.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2584, pruned_loss=0.06526, over 955180.78 frames. ], batch size: 45, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:49:20,565 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-26 20:49:23,868 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3775, 1.6412, 1.4599, 1.8786, 1.6333, 1.9403, 1.4708, 3.6064], + device='cuda:2'), covar=tensor([0.0631, 0.0735, 0.0778, 0.1099, 0.0631, 0.0524, 0.0732, 0.0191], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0041, 0.0040, 0.0039, 0.0060], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0017], + device='cuda:2') +2023-04-26 20:49:25,037 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:49:27,474 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9903, 1.2122, 1.4498, 1.6345, 1.5268, 1.6574, 1.5540, 1.5382], + device='cuda:2'), covar=tensor([0.6193, 0.8005, 0.6951, 0.6372, 0.7588, 1.1614, 0.7814, 0.7130], + device='cuda:2'), in_proj_covar=tensor([0.0324, 0.0389, 0.0319, 0.0329, 0.0343, 0.0408, 0.0369, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:49:48,138 INFO [finetune.py:976] (2/7) Epoch 8, batch 1400, loss[loss=0.2164, simple_loss=0.2679, pruned_loss=0.08249, over 4899.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2608, pruned_loss=0.06572, over 956138.80 frames. ], batch size: 35, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:50:06,403 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41521.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:50:10,639 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9910, 0.9866, 1.2565, 1.1769, 1.0063, 0.8824, 0.9706, 0.6310], + device='cuda:2'), covar=tensor([0.0631, 0.0872, 0.0539, 0.0684, 0.0867, 0.1311, 0.0615, 0.0850], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0074, 0.0072, 0.0067, 0.0076, 0.0096, 0.0080, 0.0075], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 20:50:12,930 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 1450, loss[loss=0.203, simple_loss=0.2629, pruned_loss=0.07159, over 4818.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2639, pruned_loss=0.06678, over 956710.09 frames. ], batch size: 39, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:50:30,570 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41556.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 20:50:30,711 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-26 20:50:38,925 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41569.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:50:54,404 INFO [finetune.py:976] (2/7) Epoch 8, batch 1500, loss[loss=0.2412, simple_loss=0.3102, pruned_loss=0.08609, over 4817.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.265, pruned_loss=0.06703, over 956828.10 frames. ], batch size: 38, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:50:59,630 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6667, 1.0144, 1.4948, 2.0272, 1.7421, 1.5623, 1.5791, 1.6358], + device='cuda:2'), covar=tensor([0.6917, 0.9723, 1.0232, 1.0367, 0.9273, 1.1953, 1.1396, 0.9505], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0426, 0.0507, 0.0529, 0.0437, 0.0456, 0.0467, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9199e-05, 1.0551e-04, 1.1469e-04, 1.2538e-04, 1.0615e-04, 1.1051e-04, + 1.1254e-04, 1.1310e-04], device='cuda:2') +2023-04-26 20:51:02,368 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41604.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:51:11,882 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3324, 1.6434, 1.5705, 1.7800, 1.6523, 2.1090, 1.4238, 3.7646], + device='cuda:2'), covar=tensor([0.0663, 0.0792, 0.0779, 0.1273, 0.0690, 0.0576, 0.0812, 0.0118], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0041, 0.0040, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 20:51:25,122 INFO [optim.py:369] (2/7) 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 20:51:35,877 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4416, 1.3336, 4.1506, 3.8890, 3.6620, 3.9857, 3.9732, 3.6210], + device='cuda:2'), covar=tensor([0.7113, 0.5726, 0.1014, 0.1620, 0.0989, 0.1421, 0.1102, 0.1294], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0307, 0.0410, 0.0414, 0.0349, 0.0408, 0.0318, 0.0371], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:51:38,881 INFO [finetune.py:976] (2/7) Epoch 8, batch 1550, loss[loss=0.1881, simple_loss=0.2558, pruned_loss=0.06023, over 4866.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2657, pruned_loss=0.06732, over 955179.94 frames. ], batch size: 34, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:51:51,079 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41653.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:51:58,450 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41654.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:52:02,017 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5331, 0.6601, 1.2812, 1.8550, 1.5929, 1.3995, 1.3960, 1.5018], + device='cuda:2'), covar=tensor([0.6764, 0.9192, 0.9269, 1.0076, 0.8366, 1.0865, 1.1402, 0.9671], + device='cuda:2'), in_proj_covar=tensor([0.0406, 0.0424, 0.0504, 0.0526, 0.0435, 0.0454, 0.0465, 0.0463], + device='cuda:2'), out_proj_covar=tensor([9.8659e-05, 1.0498e-04, 1.1405e-04, 1.2486e-04, 1.0577e-04, 1.0996e-04, + 1.1194e-04, 1.1260e-04], device='cuda:2') +2023-04-26 20:52:31,817 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3342, 1.1990, 1.5765, 1.5136, 1.2393, 1.1081, 1.3034, 0.8759], + device='cuda:2'), covar=tensor([0.0613, 0.0711, 0.0437, 0.0521, 0.0817, 0.1052, 0.0603, 0.0644], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0073, 0.0072, 0.0067, 0.0076, 0.0095, 0.0079, 0.0074], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 20:52:47,106 INFO [finetune.py:976] (2/7) Epoch 8, batch 1600, loss[loss=0.1724, simple_loss=0.2274, pruned_loss=0.05874, over 4027.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2638, pruned_loss=0.06721, over 954799.58 frames. ], batch size: 17, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:52:56,991 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:52:57,017 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:53:12,205 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41715.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 20:53:22,287 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 1650, loss[loss=0.1892, simple_loss=0.2247, pruned_loss=0.07688, over 4217.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2606, pruned_loss=0.06633, over 955236.38 frames. ], batch size: 18, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:53:33,799 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41749.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:53:35,635 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8488, 3.7918, 2.7733, 4.4476, 3.8682, 3.8969, 1.7841, 3.7031], + device='cuda:2'), covar=tensor([0.1577, 0.1098, 0.2899, 0.1388, 0.2595, 0.1552, 0.5651, 0.2253], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0220, 0.0252, 0.0310, 0.0303, 0.0253, 0.0275, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 20:53:40,312 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41758.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:53:56,878 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3670, 2.1523, 2.4560, 2.7450, 2.7180, 2.1776, 1.8230, 2.2397], + device='cuda:2'), covar=tensor([0.0832, 0.1014, 0.0544, 0.0563, 0.0632, 0.0908, 0.0893, 0.0664], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0205, 0.0182, 0.0179, 0.0179, 0.0192, 0.0161, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:54:03,919 INFO [finetune.py:976] (2/7) Epoch 8, batch 1700, loss[loss=0.1782, simple_loss=0.2352, pruned_loss=0.06057, over 4709.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2588, pruned_loss=0.06532, over 955931.93 frames. ], batch size: 23, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:54:11,549 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41806.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:54:24,095 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9803, 1.6368, 1.9355, 2.2833, 2.0598, 1.8944, 1.9369, 1.9257], + device='cuda:2'), covar=tensor([0.6994, 0.8549, 0.9909, 0.9266, 0.7764, 1.0789, 1.1345, 0.8881], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0427, 0.0509, 0.0530, 0.0439, 0.0457, 0.0470, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9408e-05, 1.0587e-04, 1.1501e-04, 1.2577e-04, 1.0668e-04, 1.1062e-04, + 1.1308e-04, 1.1353e-04], device='cuda:2') +2023-04-26 20:54:29,248 INFO [optim.py:369] (2/7) 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 20:54:31,870 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 20:54:37,045 INFO [finetune.py:976] (2/7) Epoch 8, batch 1750, loss[loss=0.2097, simple_loss=0.2867, pruned_loss=0.06635, over 4747.00 frames. ], tot_loss[loss=0.197, simple_loss=0.261, pruned_loss=0.06646, over 955992.72 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:54:53,648 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7002, 1.7912, 0.7862, 1.3351, 1.8425, 1.5618, 1.4542, 1.4960], + device='cuda:2'), covar=tensor([0.0527, 0.0403, 0.0402, 0.0591, 0.0290, 0.0578, 0.0560, 0.0629], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 20:55:00,120 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2521, 1.6390, 2.0834, 2.7990, 2.0474, 1.6865, 1.4550, 1.8991], + device='cuda:2'), covar=tensor([0.3739, 0.4093, 0.2015, 0.3137, 0.3809, 0.3197, 0.5103, 0.2949], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0253, 0.0219, 0.0323, 0.0215, 0.0230, 0.0237, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 20:55:03,135 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1470, 2.6709, 1.3509, 1.4841, 2.1842, 1.3792, 3.2212, 1.7827], + device='cuda:2'), covar=tensor([0.0609, 0.0712, 0.0739, 0.1131, 0.0404, 0.0901, 0.0185, 0.0604], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0052, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 20:55:10,266 INFO [finetune.py:976] (2/7) Epoch 8, batch 1800, loss[loss=0.2476, simple_loss=0.3097, pruned_loss=0.09276, over 4226.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2638, pruned_loss=0.06735, over 952463.45 frames. ], batch size: 65, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:55:18,155 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41906.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:55:35,583 INFO [optim.py:369] (2/7) 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 20:55:43,854 INFO [finetune.py:976] (2/7) Epoch 8, batch 1850, loss[loss=0.1889, simple_loss=0.2562, pruned_loss=0.06081, over 4752.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2644, pruned_loss=0.06749, over 949146.26 frames. ], batch size: 26, lr: 3.84e-03, grad_scale: 64.0 +2023-04-26 20:55:46,374 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41948.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 20:55:52,270 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0689, 2.0169, 1.7499, 1.8135, 2.2839, 1.6423, 2.7916, 1.5553], + device='cuda:2'), covar=tensor([0.3925, 0.2046, 0.5597, 0.2882, 0.1708, 0.2811, 0.1134, 0.4764], + device='cuda:2'), in_proj_covar=tensor([0.0350, 0.0354, 0.0437, 0.0366, 0.0392, 0.0386, 0.0385, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:55:58,803 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41967.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 20:56:02,852 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0333, 3.0741, 1.8496, 2.4505, 1.5183, 1.5257, 2.1761, 1.4968], + device='cuda:2'), covar=tensor([0.2124, 0.1592, 0.1984, 0.1837, 0.3050, 0.2607, 0.1325, 0.2365], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0219, 0.0174, 0.0206, 0.0208, 0.0186, 0.0164, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 20:56:17,000 INFO [finetune.py:976] (2/7) Epoch 8, batch 1900, loss[loss=0.2177, simple_loss=0.2662, pruned_loss=0.08458, over 4145.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2665, pruned_loss=0.06805, over 951301.06 frames. ], batch size: 66, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:56:21,978 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3072, 1.3102, 1.1430, 1.7426, 1.3933, 1.8392, 1.2441, 3.4141], + device='cuda:2'), covar=tensor([0.0696, 0.0919, 0.0996, 0.1218, 0.0791, 0.0553, 0.0909, 0.0239], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0041, 0.0040, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 20:56:27,974 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42009.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 20:56:28,516 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42010.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 20:56:49,878 INFO [optim.py:369] (2/7) 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 20:57:08,687 INFO [finetune.py:976] (2/7) Epoch 8, batch 1950, loss[loss=0.1636, simple_loss=0.2297, pruned_loss=0.04872, over 4742.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2645, pruned_loss=0.06732, over 951903.85 frames. ], batch size: 59, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:57:10,603 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6066, 1.4475, 1.6547, 2.0069, 1.9923, 1.5997, 1.3639, 1.7325], + device='cuda:2'), covar=tensor([0.0866, 0.1129, 0.0704, 0.0557, 0.0551, 0.0821, 0.0779, 0.0609], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0207, 0.0183, 0.0180, 0.0180, 0.0193, 0.0162, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 20:58:13,786 INFO [finetune.py:976] (2/7) Epoch 8, batch 2000, loss[loss=0.2043, simple_loss=0.2596, pruned_loss=0.07447, over 4841.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2618, pruned_loss=0.06658, over 952597.04 frames. ], batch size: 47, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:59:06,776 INFO [optim.py:369] (2/7) 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 20:59:17,300 INFO [finetune.py:976] (2/7) Epoch 8, batch 2050, loss[loss=0.1708, simple_loss=0.23, pruned_loss=0.05575, over 4749.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2584, pruned_loss=0.06535, over 954162.40 frames. ], batch size: 28, lr: 3.84e-03, grad_scale: 32.0 +2023-04-26 20:59:18,044 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1957, 2.5879, 1.1209, 1.4426, 2.0595, 1.2403, 3.3908, 1.7869], + device='cuda:2'), covar=tensor([0.0604, 0.0702, 0.0746, 0.1267, 0.0461, 0.0977, 0.0259, 0.0615], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0047, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 20:59:37,186 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3219, 3.2393, 2.5092, 3.8167, 3.2152, 3.3922, 1.5969, 3.2491], + device='cuda:2'), covar=tensor([0.1993, 0.1379, 0.3124, 0.2078, 0.2790, 0.2102, 0.5075, 0.2587], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0216, 0.0248, 0.0307, 0.0299, 0.0248, 0.0270, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:00:12,508 INFO [finetune.py:976] (2/7) Epoch 8, batch 2100, loss[loss=0.2142, simple_loss=0.2724, pruned_loss=0.07799, over 4144.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2588, pruned_loss=0.06592, over 952383.18 frames. ], batch size: 65, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:00:24,425 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 21:00:27,905 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8630, 2.2171, 2.0495, 2.1889, 1.9456, 2.2961, 2.1164, 2.0427], + device='cuda:2'), covar=tensor([0.5361, 0.8673, 0.7582, 0.6475, 0.7918, 0.9705, 0.9339, 0.8188], + device='cuda:2'), in_proj_covar=tensor([0.0321, 0.0385, 0.0316, 0.0327, 0.0341, 0.0405, 0.0368, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:00:38,851 INFO [optim.py:369] (2/7) 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 21:00:41,473 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1242, 1.5870, 2.0213, 2.2673, 1.9495, 1.5834, 1.1349, 1.6619], + device='cuda:2'), covar=tensor([0.4512, 0.4255, 0.1988, 0.3390, 0.3592, 0.3456, 0.5317, 0.3145], + device='cuda:2'), in_proj_covar=tensor([0.0280, 0.0253, 0.0219, 0.0322, 0.0214, 0.0229, 0.0235, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 21:00:46,719 INFO [finetune.py:976] (2/7) Epoch 8, batch 2150, loss[loss=0.2182, simple_loss=0.2768, pruned_loss=0.07974, over 4838.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2623, pruned_loss=0.06708, over 954158.84 frames. ], batch size: 30, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:00:58,118 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42262.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:01:20,321 INFO [finetune.py:976] (2/7) Epoch 8, batch 2200, loss[loss=0.1782, simple_loss=0.2458, pruned_loss=0.05526, over 4728.00 frames. ], tot_loss[loss=0.2017, simple_loss=0.2659, pruned_loss=0.06875, over 952678.51 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:01:26,912 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42304.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:01:30,540 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42310.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 21:01:44,809 INFO [optim.py:369] (2/7) 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 21:01:52,460 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6035, 1.5374, 0.8408, 1.2829, 1.7770, 1.4822, 1.3949, 1.4317], + device='cuda:2'), covar=tensor([0.0514, 0.0405, 0.0376, 0.0581, 0.0285, 0.0541, 0.0512, 0.0594], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 21:01:52,972 INFO [finetune.py:976] (2/7) Epoch 8, batch 2250, loss[loss=0.2034, simple_loss=0.28, pruned_loss=0.06342, over 4924.00 frames. ], tot_loss[loss=0.2037, simple_loss=0.2677, pruned_loss=0.06981, over 953913.51 frames. ], batch size: 41, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:01:55,427 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0378, 1.0040, 1.2231, 1.1068, 1.0010, 0.8829, 0.9612, 0.5789], + device='cuda:2'), covar=tensor([0.0574, 0.0767, 0.0536, 0.0755, 0.0766, 0.1370, 0.0575, 0.0902], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0076, 0.0096, 0.0080, 0.0075], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:02:02,510 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42358.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:02:11,129 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3710, 1.3598, 1.3888, 1.0793, 1.4142, 1.2063, 1.7745, 1.2665], + device='cuda:2'), covar=tensor([0.4087, 0.1755, 0.5365, 0.2737, 0.1526, 0.2115, 0.1560, 0.5217], + device='cuda:2'), in_proj_covar=tensor([0.0347, 0.0352, 0.0434, 0.0363, 0.0389, 0.0384, 0.0384, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:02:15,804 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42379.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:02:24,544 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 21:02:26,180 INFO [finetune.py:976] (2/7) Epoch 8, batch 2300, loss[loss=0.2181, simple_loss=0.2836, pruned_loss=0.07629, over 4774.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2681, pruned_loss=0.06923, over 953249.76 frames. ], batch size: 28, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:02:50,914 INFO [optim.py:369] (2/7) 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 21:02:56,800 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42440.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:02:57,383 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3881, 1.7646, 1.6968, 2.1799, 1.9126, 2.0531, 1.5223, 4.4825], + device='cuda:2'), covar=tensor([0.0562, 0.0738, 0.0724, 0.1090, 0.0613, 0.0516, 0.0720, 0.0116], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0041, 0.0040, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:02:59,687 INFO [finetune.py:976] (2/7) Epoch 8, batch 2350, loss[loss=0.1719, simple_loss=0.2363, pruned_loss=0.05373, over 4751.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2651, pruned_loss=0.06808, over 955209.36 frames. ], batch size: 28, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:03:22,261 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-04-26 21:03:23,901 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5461, 1.6345, 1.4425, 1.1812, 1.1457, 1.2156, 1.4827, 1.1408], + device='cuda:2'), covar=tensor([0.1835, 0.1486, 0.1700, 0.1749, 0.2552, 0.2085, 0.1159, 0.2142], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0219, 0.0173, 0.0206, 0.0207, 0.0186, 0.0164, 0.0190], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 21:03:48,469 INFO [finetune.py:976] (2/7) Epoch 8, batch 2400, loss[loss=0.1797, simple_loss=0.2355, pruned_loss=0.06199, over 4829.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2618, pruned_loss=0.06747, over 954930.57 frames. ], batch size: 25, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:03:56,832 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-26 21:04:08,560 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42506.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:04:21,383 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8768, 1.7361, 4.6978, 4.4045, 4.1718, 4.4782, 4.3102, 4.1467], + device='cuda:2'), covar=tensor([0.6817, 0.5668, 0.1197, 0.1838, 0.1230, 0.1871, 0.1301, 0.1541], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0308, 0.0409, 0.0416, 0.0352, 0.0407, 0.0318, 0.0370], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:04:40,674 INFO [optim.py:369] (2/7) 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 21:04:54,584 INFO [finetune.py:976] (2/7) Epoch 8, batch 2450, loss[loss=0.2144, simple_loss=0.2754, pruned_loss=0.0767, over 4902.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2583, pruned_loss=0.0661, over 954255.66 frames. ], batch size: 37, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:05:24,359 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42562.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:05:27,979 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42567.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:05:55,812 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 21:06:00,481 INFO [finetune.py:976] (2/7) Epoch 8, batch 2500, loss[loss=0.1704, simple_loss=0.2504, pruned_loss=0.04523, over 4801.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2593, pruned_loss=0.06604, over 953255.89 frames. ], batch size: 45, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:06:20,330 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1228, 1.3370, 1.1611, 1.5920, 1.4172, 1.4855, 1.2704, 2.4230], + device='cuda:2'), covar=tensor([0.0642, 0.0824, 0.0883, 0.1257, 0.0687, 0.0535, 0.0782, 0.0258], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0041, 0.0040, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:06:20,336 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) 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 21:07:08,191 INFO [finetune.py:976] (2/7) Epoch 8, batch 2550, loss[loss=0.1731, simple_loss=0.244, pruned_loss=0.05112, over 4762.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.263, pruned_loss=0.06687, over 952830.33 frames. ], batch size: 54, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:07:11,904 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42652.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:07:29,131 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1558, 3.1612, 2.6554, 3.6545, 3.0980, 3.1727, 1.7035, 3.1055], + device='cuda:2'), covar=tensor([0.2056, 0.1313, 0.4095, 0.2394, 0.2767, 0.1984, 0.4950, 0.2975], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0250, 0.0308, 0.0298, 0.0250, 0.0271, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:07:41,702 INFO [finetune.py:976] (2/7) Epoch 8, batch 2600, loss[loss=0.2256, simple_loss=0.2833, pruned_loss=0.08391, over 4790.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2647, pruned_loss=0.0676, over 953793.11 frames. ], batch size: 51, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:07:53,064 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42710.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:08:07,493 INFO [optim.py:369] (2/7) 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 21:08:09,417 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42735.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:08:11,386 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.29 vs. limit=5.0 +2023-04-26 21:08:13,076 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3014, 1.7378, 1.5785, 2.0977, 1.8672, 1.9652, 1.5404, 4.4059], + device='cuda:2'), covar=tensor([0.0616, 0.0781, 0.0798, 0.1189, 0.0677, 0.0568, 0.0798, 0.0116], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0041, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:08:14,788 INFO [finetune.py:976] (2/7) Epoch 8, batch 2650, loss[loss=0.1951, simple_loss=0.2576, pruned_loss=0.0663, over 4829.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2667, pruned_loss=0.06826, over 953393.38 frames. ], batch size: 49, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:08:48,362 INFO [finetune.py:976] (2/7) Epoch 8, batch 2700, loss[loss=0.162, simple_loss=0.2265, pruned_loss=0.04876, over 4230.00 frames. ], tot_loss[loss=0.2, simple_loss=0.265, pruned_loss=0.06747, over 952944.24 frames. ], batch size: 18, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:09:14,554 INFO [optim.py:369] (2/7) 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 21:09:17,226 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-26 21:09:21,927 INFO [finetune.py:976] (2/7) Epoch 8, batch 2750, loss[loss=0.2036, simple_loss=0.2561, pruned_loss=0.07553, over 4760.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2625, pruned_loss=0.06693, over 953792.06 frames. ], batch size: 27, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:09:31,576 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:09:34,273 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42862.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:09:57,308 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42879.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:10:17,377 INFO [finetune.py:976] (2/7) Epoch 8, batch 2800, loss[loss=0.1535, simple_loss=0.2214, pruned_loss=0.0428, over 4936.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2589, pruned_loss=0.06567, over 954113.84 frames. ], batch size: 38, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:10:20,591 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 21:10:29,412 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9869, 3.8991, 2.8246, 4.5330, 4.0855, 3.8984, 1.7756, 3.8451], + device='cuda:2'), covar=tensor([0.1619, 0.1061, 0.2975, 0.1413, 0.2520, 0.1616, 0.5803, 0.2285], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0216, 0.0249, 0.0308, 0.0299, 0.0250, 0.0272, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:10:51,623 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42920.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:11:10,641 INFO [optim.py:369] (2/7) 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 21:11:16,204 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 2850, loss[loss=0.2358, simple_loss=0.3079, pruned_loss=0.08189, over 4822.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2579, pruned_loss=0.06574, over 951882.09 frames. ], batch size: 39, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:11:27,069 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1905, 2.7135, 1.1639, 1.3966, 2.0403, 1.2182, 3.5514, 1.7789], + device='cuda:2'), covar=tensor([0.0663, 0.0710, 0.0839, 0.1314, 0.0511, 0.1043, 0.0265, 0.0613], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 21:11:56,701 INFO [finetune.py:976] (2/7) Epoch 8, batch 2900, loss[loss=0.2302, simple_loss=0.2992, pruned_loss=0.08062, over 4738.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2617, pruned_loss=0.06721, over 952618.40 frames. ], batch size: 54, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:11:56,882 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 21:12:15,327 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43005.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:12:28,425 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9903, 2.5049, 1.0009, 1.4346, 2.1201, 1.2212, 3.3917, 1.8091], + device='cuda:2'), covar=tensor([0.0717, 0.0868, 0.0951, 0.1281, 0.0473, 0.1044, 0.0228, 0.0604], + device='cuda:2'), in_proj_covar=tensor([0.0053, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 21:12:51,487 INFO [optim.py:369] (2/7) 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 21:12:59,361 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43035.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:12:59,415 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6786, 1.2721, 1.7449, 1.9503, 1.7062, 1.6251, 1.6996, 1.7454], + device='cuda:2'), covar=tensor([0.8820, 1.2238, 1.2335, 1.5388, 1.0377, 1.4344, 1.5037, 1.3440], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0425, 0.0507, 0.0529, 0.0437, 0.0457, 0.0469, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9149e-05, 1.0540e-04, 1.1456e-04, 1.2563e-04, 1.0641e-04, 1.1042e-04, + 1.1295e-04, 1.1304e-04], device='cuda:2') +2023-04-26 21:13:10,705 INFO [finetune.py:976] (2/7) Epoch 8, batch 2950, loss[loss=0.2283, simple_loss=0.2854, pruned_loss=0.08559, over 4911.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2643, pruned_loss=0.0675, over 952110.70 frames. ], batch size: 36, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:13:10,808 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7185, 1.9863, 1.0080, 1.4449, 2.0271, 1.6333, 1.6002, 1.6790], + device='cuda:2'), covar=tensor([0.0551, 0.0374, 0.0349, 0.0559, 0.0260, 0.0512, 0.0516, 0.0596], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 21:13:37,396 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43083.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:13:37,496 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6032, 1.1880, 1.7051, 2.0928, 1.7368, 1.6210, 1.6485, 1.6347], + device='cuda:2'), covar=tensor([0.6932, 0.9538, 0.9634, 0.9769, 0.8836, 1.1597, 1.1798, 1.0432], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0425, 0.0507, 0.0529, 0.0438, 0.0457, 0.0469, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9219e-05, 1.0552e-04, 1.1470e-04, 1.2564e-04, 1.0654e-04, 1.1058e-04, + 1.1307e-04, 1.1322e-04], device='cuda:2') +2023-04-26 21:13:44,491 INFO [finetune.py:976] (2/7) Epoch 8, batch 3000, loss[loss=0.2138, simple_loss=0.2813, pruned_loss=0.07311, over 4830.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2673, pruned_loss=0.06887, over 954126.81 frames. ], batch size: 30, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:13:44,491 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 21:13:50,383 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4133, 1.4251, 3.8075, 3.5339, 3.4610, 3.7086, 3.7417, 3.4187], + device='cuda:2'), covar=tensor([0.6580, 0.4606, 0.1139, 0.1758, 0.1120, 0.1311, 0.0750, 0.1494], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0309, 0.0411, 0.0415, 0.0351, 0.0407, 0.0318, 0.0372], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:13:51,391 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8218, 2.1444, 1.9161, 2.0620, 1.6857, 1.7738, 1.8403, 1.4926], + device='cuda:2'), covar=tensor([0.1976, 0.1358, 0.0872, 0.1164, 0.3699, 0.1338, 0.1791, 0.2619], + device='cuda:2'), in_proj_covar=tensor([0.0302, 0.0322, 0.0233, 0.0296, 0.0320, 0.0276, 0.0262, 0.0287], + device='cuda:2'), 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:2') +2023-04-26 21:13:54,961 INFO [finetune.py:1010] (2/7) Epoch 8, validation: loss=0.1551, simple_loss=0.2273, pruned_loss=0.04149, over 2265189.00 frames. +2023-04-26 21:13:54,962 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 21:14:18,550 INFO [optim.py:369] (2/7) 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 21:14:27,800 INFO [finetune.py:976] (2/7) Epoch 8, batch 3050, loss[loss=0.2058, simple_loss=0.2726, pruned_loss=0.06956, over 4827.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.268, pruned_loss=0.06853, over 954814.22 frames. ], batch size: 47, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:14:29,553 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1057, 1.5326, 1.3656, 1.7746, 1.5823, 1.9461, 1.3627, 3.6442], + device='cuda:2'), covar=tensor([0.0704, 0.0801, 0.0839, 0.1227, 0.0688, 0.0546, 0.0823, 0.0142], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0041, 0.0044, 0.0041, 0.0040, 0.0040, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:14:36,088 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.59 vs. limit=5.0 +2023-04-26 21:14:39,667 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43162.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:14:59,302 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8613, 1.1184, 1.4515, 1.6068, 1.5593, 1.7069, 1.5306, 1.5025], + device='cuda:2'), covar=tensor([0.4029, 0.5701, 0.5386, 0.5010, 0.5848, 0.8270, 0.5819, 0.5224], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0386, 0.0317, 0.0326, 0.0343, 0.0406, 0.0367, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:15:00,984 INFO [finetune.py:976] (2/7) Epoch 8, batch 3100, loss[loss=0.1718, simple_loss=0.2314, pruned_loss=0.05616, over 4850.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2651, pruned_loss=0.06791, over 954089.06 frames. ], batch size: 49, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:15:11,983 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43215.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:15:25,428 INFO [optim.py:369] (2/7) 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 21:15:26,164 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6934, 1.6438, 0.8901, 1.3592, 1.7191, 1.5969, 1.5011, 1.4773], + device='cuda:2'), covar=tensor([0.0540, 0.0391, 0.0374, 0.0580, 0.0293, 0.0555, 0.0522, 0.0613], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 21:15:27,879 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43235.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:15:28,561 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9364, 1.8194, 2.0874, 2.2379, 2.0813, 1.7774, 1.8705, 1.8847], + device='cuda:2'), covar=tensor([0.6388, 0.8811, 1.0011, 0.8491, 0.7281, 1.1977, 1.2921, 1.1458], + device='cuda:2'), in_proj_covar=tensor([0.0407, 0.0423, 0.0506, 0.0528, 0.0437, 0.0457, 0.0468, 0.0465], + device='cuda:2'), out_proj_covar=tensor([9.8917e-05, 1.0499e-04, 1.1436e-04, 1.2525e-04, 1.0618e-04, 1.1049e-04, + 1.1283e-04, 1.1290e-04], device='cuda:2') +2023-04-26 21:15:34,183 INFO [finetune.py:976] (2/7) Epoch 8, batch 3150, loss[loss=0.2357, simple_loss=0.293, pruned_loss=0.08915, over 4874.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.261, pruned_loss=0.06655, over 954454.48 frames. ], batch size: 34, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:16:09,910 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1696, 1.6555, 1.5263, 1.9273, 1.7908, 2.0130, 1.4826, 3.8442], + device='cuda:2'), covar=tensor([0.0604, 0.0705, 0.0731, 0.1043, 0.0591, 0.0657, 0.0741, 0.0131], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:16:18,114 INFO [finetune.py:976] (2/7) Epoch 8, batch 3200, loss[loss=0.1916, simple_loss=0.2566, pruned_loss=0.06329, over 4758.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2571, pruned_loss=0.06498, over 955175.46 frames. ], batch size: 27, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:16:30,931 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43305.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:16:31,525 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9771, 3.8540, 2.8906, 4.5781, 3.9929, 3.9215, 1.6699, 3.8645], + device='cuda:2'), covar=tensor([0.1642, 0.1328, 0.3206, 0.1467, 0.3730, 0.1676, 0.6025, 0.2412], + device='cuda:2'), in_proj_covar=tensor([0.0241, 0.0215, 0.0247, 0.0307, 0.0297, 0.0249, 0.0269, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:17:03,559 INFO [optim.py:369] (2/7) 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 21:17:14,648 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7046, 3.7371, 1.3844, 1.9944, 2.0973, 2.7151, 2.2591, 1.1677], + device='cuda:2'), covar=tensor([0.1396, 0.0972, 0.1827, 0.1394, 0.1108, 0.1030, 0.1425, 0.1868], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0252, 0.0141, 0.0123, 0.0135, 0.0154, 0.0120, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:17:20,616 INFO [finetune.py:976] (2/7) Epoch 8, batch 3250, loss[loss=0.173, simple_loss=0.227, pruned_loss=0.05951, over 4224.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2579, pruned_loss=0.0653, over 954898.24 frames. ], batch size: 18, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:17:21,546 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-04-26 21:17:33,415 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43353.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:18:26,604 INFO [finetune.py:976] (2/7) Epoch 8, batch 3300, loss[loss=0.1634, simple_loss=0.2299, pruned_loss=0.04845, over 4443.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2602, pruned_loss=0.06568, over 953997.79 frames. ], batch size: 19, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:18:59,197 INFO [optim.py:369] (2/7) 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 21:19:06,530 INFO [finetune.py:976] (2/7) Epoch 8, batch 3350, loss[loss=0.1533, simple_loss=0.2299, pruned_loss=0.03836, over 4861.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2618, pruned_loss=0.06579, over 956250.47 frames. ], batch size: 34, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:19:26,580 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8145, 2.5101, 1.8424, 1.7544, 1.2427, 1.3250, 1.9756, 1.3345], + device='cuda:2'), covar=tensor([0.1959, 0.1709, 0.1908, 0.2112, 0.2872, 0.2322, 0.1337, 0.2376], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0218, 0.0173, 0.0206, 0.0206, 0.0185, 0.0164, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 21:19:40,384 INFO [finetune.py:976] (2/7) Epoch 8, batch 3400, loss[loss=0.1699, simple_loss=0.2415, pruned_loss=0.04916, over 4846.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2636, pruned_loss=0.06653, over 955479.05 frames. ], batch size: 49, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:19:54,942 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43514.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:19:55,533 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43515.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:06,649 INFO [optim.py:369] (2/7) 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 21:20:08,581 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43535.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:13,907 INFO [finetune.py:976] (2/7) Epoch 8, batch 3450, loss[loss=0.1745, simple_loss=0.2426, pruned_loss=0.05321, over 4722.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2638, pruned_loss=0.06591, over 958322.73 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:20:26,909 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43563.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:35,627 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43583.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:20:47,606 INFO [finetune.py:976] (2/7) Epoch 8, batch 3500, loss[loss=0.205, simple_loss=0.2567, pruned_loss=0.07668, over 4896.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2612, pruned_loss=0.06573, over 956018.26 frames. ], batch size: 35, lr: 3.83e-03, grad_scale: 32.0 +2023-04-26 21:21:13,732 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 3550, loss[loss=0.2002, simple_loss=0.2486, pruned_loss=0.07593, over 4898.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2587, pruned_loss=0.06517, over 956434.37 frames. ], batch size: 36, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:22:22,680 INFO [finetune.py:976] (2/7) Epoch 8, batch 3600, loss[loss=0.1672, simple_loss=0.2318, pruned_loss=0.05128, over 4774.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2541, pruned_loss=0.06309, over 955716.22 frames. ], batch size: 27, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:22:55,181 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5204, 1.2702, 4.1621, 3.8807, 3.6536, 3.8200, 3.7265, 3.6640], + device='cuda:2'), covar=tensor([0.7668, 0.6009, 0.1145, 0.2079, 0.1482, 0.1587, 0.2721, 0.1551], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0306, 0.0407, 0.0410, 0.0349, 0.0403, 0.0316, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:23:04,440 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5519, 1.2074, 1.6613, 2.0510, 1.7155, 1.5518, 1.6100, 1.6048], + device='cuda:2'), covar=tensor([0.6867, 0.9199, 0.9205, 0.9313, 0.7778, 1.0889, 1.0937, 1.0201], + device='cuda:2'), in_proj_covar=tensor([0.0407, 0.0423, 0.0506, 0.0528, 0.0436, 0.0456, 0.0467, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9015e-05, 1.0484e-04, 1.1449e-04, 1.2520e-04, 1.0611e-04, 1.1024e-04, + 1.1249e-04, 1.1291e-04], device='cuda:2') +2023-04-26 21:23:15,354 INFO [optim.py:369] (2/7) 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 21:23:34,071 INFO [finetune.py:976] (2/7) Epoch 8, batch 3650, loss[loss=0.2506, simple_loss=0.3202, pruned_loss=0.0905, over 4265.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2573, pruned_loss=0.06449, over 955076.98 frames. ], batch size: 65, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:23:36,618 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3945, 1.2469, 1.6907, 1.6024, 1.3035, 1.1465, 1.3884, 0.9559], + device='cuda:2'), covar=tensor([0.0712, 0.0893, 0.0510, 0.0785, 0.0920, 0.1423, 0.0664, 0.0851], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0073, 0.0072, 0.0067, 0.0076, 0.0095, 0.0079, 0.0075], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:23:37,209 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1233, 1.9776, 2.2152, 2.5177, 2.6193, 2.0343, 1.7638, 2.2072], + device='cuda:2'), covar=tensor([0.0889, 0.1022, 0.0519, 0.0577, 0.0522, 0.0902, 0.0848, 0.0612], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0205, 0.0180, 0.0177, 0.0178, 0.0190, 0.0160, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:23:45,307 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2253, 1.4986, 1.3188, 1.4358, 1.2258, 1.2260, 1.2794, 1.1452], + device='cuda:2'), covar=tensor([0.1943, 0.1408, 0.1004, 0.1386, 0.4019, 0.1406, 0.1816, 0.2178], + device='cuda:2'), in_proj_covar=tensor([0.0301, 0.0321, 0.0232, 0.0294, 0.0319, 0.0275, 0.0261, 0.0285], + device='cuda:2'), out_proj_covar=tensor([1.2210e-04, 1.2960e-04, 9.3512e-05, 1.1763e-04, 1.3103e-04, 1.1122e-04, + 1.0699e-04, 1.1466e-04], device='cuda:2') +2023-04-26 21:24:00,774 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2069, 1.4185, 1.5447, 1.7578, 1.6003, 1.7032, 1.6802, 1.6257], + device='cuda:2'), covar=tensor([0.5883, 0.7225, 0.6365, 0.5905, 0.7235, 1.0253, 0.7358, 0.6922], + device='cuda:2'), in_proj_covar=tensor([0.0325, 0.0389, 0.0319, 0.0330, 0.0346, 0.0409, 0.0370, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:24:29,558 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 3700, loss[loss=0.1937, simple_loss=0.2746, pruned_loss=0.05635, over 4810.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2618, pruned_loss=0.06593, over 954463.83 frames. ], batch size: 39, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:25:03,746 INFO [optim.py:369] (2/7) 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 21:25:11,090 INFO [finetune.py:976] (2/7) Epoch 8, batch 3750, loss[loss=0.2299, simple_loss=0.2877, pruned_loss=0.08604, over 4807.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2641, pruned_loss=0.06665, over 955745.87 frames. ], batch size: 38, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:25:14,039 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43848.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:25:27,356 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43870.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:25:39,449 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6039, 3.9844, 0.7281, 2.0964, 2.0606, 2.6246, 2.3370, 1.0433], + device='cuda:2'), covar=tensor([0.1479, 0.0899, 0.2470, 0.1401, 0.1205, 0.1104, 0.1353, 0.2168], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0256, 0.0143, 0.0125, 0.0136, 0.0156, 0.0121, 0.0122], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:25:44,744 INFO [finetune.py:976] (2/7) Epoch 8, batch 3800, loss[loss=0.1744, simple_loss=0.2463, pruned_loss=0.05127, over 4844.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2659, pruned_loss=0.06741, over 958037.21 frames. ], batch size: 49, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:25:47,322 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0818, 2.4970, 1.1006, 1.4470, 1.8731, 1.2866, 3.3840, 1.7612], + device='cuda:2'), covar=tensor([0.0697, 0.0637, 0.0799, 0.1338, 0.0520, 0.1026, 0.0360, 0.0669], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0007], + device='cuda:2') +2023-04-26 21:25:54,498 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43909.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:26:09,807 INFO [optim.py:369] (2/7) 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 21:26:18,043 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6883, 3.5733, 2.5785, 4.2383, 3.7496, 3.5787, 1.5907, 3.6231], + device='cuda:2'), covar=tensor([0.1642, 0.1211, 0.3098, 0.1903, 0.2830, 0.2284, 0.5884, 0.2251], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0215, 0.0248, 0.0306, 0.0298, 0.0249, 0.0269, 0.0267], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:26:18,579 INFO [finetune.py:976] (2/7) Epoch 8, batch 3850, loss[loss=0.1591, simple_loss=0.2211, pruned_loss=0.04852, over 4805.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2637, pruned_loss=0.0658, over 958876.81 frames. ], batch size: 25, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:26:25,312 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4855, 0.6050, 1.2595, 1.8878, 1.6409, 1.3903, 1.3366, 1.4260], + device='cuda:2'), covar=tensor([0.6109, 0.8347, 0.8077, 0.8781, 0.7273, 0.9606, 0.9696, 0.8659], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0424, 0.0508, 0.0530, 0.0438, 0.0457, 0.0470, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9473e-05, 1.0504e-04, 1.1485e-04, 1.2563e-04, 1.0648e-04, 1.1039e-04, + 1.1298e-04, 1.1329e-04], device='cuda:2') +2023-04-26 21:26:28,335 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43970.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:27:02,337 INFO [finetune.py:976] (2/7) Epoch 8, batch 3900, loss[loss=0.1706, simple_loss=0.2347, pruned_loss=0.05322, over 4812.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2606, pruned_loss=0.06524, over 956842.20 frames. ], batch size: 51, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:27:14,295 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.2076, 3.1155, 2.3633, 3.6807, 3.1034, 3.2452, 1.4452, 3.2133], + device='cuda:2'), covar=tensor([0.2009, 0.1427, 0.3184, 0.2182, 0.3076, 0.2167, 0.5418, 0.2653], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0216, 0.0248, 0.0307, 0.0299, 0.0250, 0.0270, 0.0268], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:27:35,501 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 8, batch 3950, loss[loss=0.147, simple_loss=0.2113, pruned_loss=0.04135, over 4750.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2567, pruned_loss=0.06397, over 957079.27 frames. ], batch size: 23, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:28:58,843 INFO [finetune.py:976] (2/7) Epoch 8, batch 4000, loss[loss=0.2102, simple_loss=0.2697, pruned_loss=0.0754, over 4906.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2555, pruned_loss=0.0638, over 957590.83 frames. ], batch size: 32, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:29:17,896 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0884, 2.5162, 0.9708, 1.4553, 1.8514, 1.1720, 3.4013, 1.7559], + device='cuda:2'), covar=tensor([0.0687, 0.0844, 0.0816, 0.1255, 0.0540, 0.1069, 0.0336, 0.0616], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 21:29:49,240 INFO [optim.py:369] (2/7) 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 21:30:02,121 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44143.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:30:02,666 INFO [finetune.py:976] (2/7) Epoch 8, batch 4050, loss[loss=0.1779, simple_loss=0.2554, pruned_loss=0.05018, over 4901.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2603, pruned_loss=0.06587, over 956832.59 frames. ], batch size: 43, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:30:10,780 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-26 21:30:35,440 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44170.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:30:42,087 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 21:30:46,850 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 21:31:07,887 INFO [finetune.py:976] (2/7) Epoch 8, batch 4100, loss[loss=0.2414, simple_loss=0.3103, pruned_loss=0.08629, over 4743.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2638, pruned_loss=0.0672, over 955567.81 frames. ], batch size: 54, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:31:28,941 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8991, 1.7380, 2.0201, 2.2944, 2.4080, 1.8745, 1.4576, 2.0662], + device='cuda:2'), covar=tensor([0.0937, 0.1119, 0.0618, 0.0630, 0.0491, 0.0863, 0.0940, 0.0614], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0202, 0.0178, 0.0175, 0.0175, 0.0187, 0.0159, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:31:41,630 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44218.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:32:00,834 INFO [optim.py:369] (2/7) 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 21:32:03,527 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-26 21:32:13,834 INFO [finetune.py:976] (2/7) Epoch 8, batch 4150, loss[loss=0.2037, simple_loss=0.2577, pruned_loss=0.07485, over 4771.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2651, pruned_loss=0.06765, over 953914.17 frames. ], batch size: 26, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:32:34,619 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44265.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:32:50,348 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6294, 1.4153, 1.9210, 1.8837, 1.4380, 1.3033, 1.6213, 1.0459], + device='cuda:2'), covar=tensor([0.0670, 0.0928, 0.0467, 0.0771, 0.0913, 0.1263, 0.0729, 0.0909], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0073, 0.0072, 0.0066, 0.0076, 0.0095, 0.0078, 0.0074], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:32:52,669 INFO [finetune.py:976] (2/7) Epoch 8, batch 4200, loss[loss=0.1845, simple_loss=0.2483, pruned_loss=0.06035, over 4777.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2651, pruned_loss=0.06717, over 952727.26 frames. ], batch size: 51, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:32:56,237 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1620, 1.7036, 1.4510, 1.9072, 1.7665, 1.9521, 1.5205, 3.7871], + device='cuda:2'), covar=tensor([0.0661, 0.0749, 0.0806, 0.1159, 0.0639, 0.0559, 0.0732, 0.0142], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:33:08,257 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 21:33:24,806 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 21:33:26,340 INFO [finetune.py:976] (2/7) Epoch 8, batch 4250, loss[loss=0.1823, simple_loss=0.2511, pruned_loss=0.05672, over 4789.00 frames. ], tot_loss[loss=0.198, simple_loss=0.263, pruned_loss=0.0665, over 953114.04 frames. ], batch size: 29, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:33:38,249 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-04-26 21:34:00,167 INFO [finetune.py:976] (2/7) Epoch 8, batch 4300, loss[loss=0.2111, simple_loss=0.2725, pruned_loss=0.07486, over 4916.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2603, pruned_loss=0.06603, over 951900.84 frames. ], batch size: 32, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:34:26,193 INFO [optim.py:369] (2/7) 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 21:34:44,407 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 4350, loss[loss=0.1792, simple_loss=0.2327, pruned_loss=0.0628, over 4893.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2558, pruned_loss=0.06366, over 954027.22 frames. ], batch size: 32, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:35:46,508 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44491.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:35:48,246 INFO [finetune.py:976] (2/7) Epoch 8, batch 4400, loss[loss=0.1604, simple_loss=0.2149, pruned_loss=0.05296, over 3263.00 frames. ], tot_loss[loss=0.192, simple_loss=0.256, pruned_loss=0.06404, over 951297.31 frames. ], batch size: 14, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:36:11,296 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44527.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:36:14,644 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 8, batch 4450, loss[loss=0.2461, simple_loss=0.2956, pruned_loss=0.09828, over 4230.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2612, pruned_loss=0.06624, over 951597.91 frames. ], batch size: 65, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:36:42,762 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44550.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:36:58,010 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44565.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:37:30,486 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 4500, loss[loss=0.2272, simple_loss=0.2997, pruned_loss=0.07734, over 4801.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.265, pruned_loss=0.06769, over 951262.89 frames. ], batch size: 45, lr: 3.82e-03, grad_scale: 64.0 +2023-04-26 21:37:49,430 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8037, 2.1496, 1.1073, 1.6542, 2.4477, 1.7858, 1.7482, 1.8260], + device='cuda:2'), covar=tensor([0.0508, 0.0359, 0.0328, 0.0569, 0.0233, 0.0542, 0.0486, 0.0593], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0049, 0.0037, 0.0048, 0.0048, 0.0049], + device='cuda:2') +2023-04-26 21:37:57,807 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44611.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:37:58,974 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44613.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:38:00,234 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44615.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:38:30,726 INFO [optim.py:369] (2/7) 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 21:38:43,666 INFO [finetune.py:976] (2/7) Epoch 8, batch 4550, loss[loss=0.2029, simple_loss=0.2819, pruned_loss=0.06197, over 4725.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2664, pruned_loss=0.06801, over 952890.10 frames. ], batch size: 59, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:38:55,367 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44663.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:39:17,128 INFO [finetune.py:976] (2/7) Epoch 8, batch 4600, loss[loss=0.1765, simple_loss=0.2464, pruned_loss=0.05325, over 4848.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2655, pruned_loss=0.06742, over 953228.95 frames. ], batch size: 49, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:39:43,235 INFO [optim.py:369] (2/7) 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 21:39:55,901 INFO [finetune.py:976] (2/7) Epoch 8, batch 4650, loss[loss=0.2039, simple_loss=0.2654, pruned_loss=0.07119, over 4774.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2623, pruned_loss=0.06655, over 954574.61 frames. ], batch size: 54, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:40:29,730 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-26 21:41:01,773 INFO [finetune.py:976] (2/7) Epoch 8, batch 4700, loss[loss=0.171, simple_loss=0.2402, pruned_loss=0.05095, over 4906.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2595, pruned_loss=0.06552, over 954946.62 frames. ], batch size: 32, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:41:14,583 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8431, 1.3954, 1.6839, 1.7813, 1.6367, 1.3506, 0.7692, 1.3743], + device='cuda:2'), covar=tensor([0.3626, 0.3667, 0.1806, 0.2311, 0.2764, 0.2920, 0.4714, 0.2511], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0251, 0.0219, 0.0321, 0.0214, 0.0229, 0.0236, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 21:41:26,394 INFO [optim.py:369] (2/7) 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 21:41:35,060 INFO [finetune.py:976] (2/7) Epoch 8, batch 4750, loss[loss=0.1804, simple_loss=0.2411, pruned_loss=0.05986, over 4837.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.257, pruned_loss=0.06483, over 955505.98 frames. ], batch size: 30, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:41:39,340 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-26 21:41:55,736 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-04-26 21:41:58,691 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7956, 2.3670, 1.9987, 2.1448, 1.7090, 2.0345, 2.1309, 1.6487], + device='cuda:2'), covar=tensor([0.2050, 0.1142, 0.0819, 0.1248, 0.3074, 0.1107, 0.1966, 0.2285], + device='cuda:2'), in_proj_covar=tensor([0.0301, 0.0322, 0.0230, 0.0293, 0.0318, 0.0274, 0.0260, 0.0284], + device='cuda:2'), out_proj_covar=tensor([1.2203e-04, 1.2988e-04, 9.2841e-05, 1.1740e-04, 1.3072e-04, 1.1055e-04, + 1.0660e-04, 1.1394e-04], device='cuda:2') +2023-04-26 21:41:59,216 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 4800, loss[loss=0.2017, simple_loss=0.2651, pruned_loss=0.06915, over 4906.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2603, pruned_loss=0.06672, over 953689.35 frames. ], batch size: 32, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:42:10,800 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6310, 1.3726, 1.2667, 1.4461, 1.9255, 1.4966, 1.3446, 1.2319], + device='cuda:2'), covar=tensor([0.1835, 0.1444, 0.1981, 0.1469, 0.0796, 0.1834, 0.2045, 0.2123], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0327, 0.0357, 0.0305, 0.0342, 0.0328, 0.0311, 0.0360], + device='cuda:2'), out_proj_covar=tensor([6.6053e-05, 6.9380e-05, 7.7292e-05, 6.3122e-05, 7.1964e-05, 7.0563e-05, + 6.6929e-05, 7.7339e-05], device='cuda:2') +2023-04-26 21:42:16,074 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 21:42:40,903 INFO [finetune.py:976] (2/7) Epoch 8, batch 4850, loss[loss=0.2002, simple_loss=0.275, pruned_loss=0.06266, over 4906.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2647, pruned_loss=0.06851, over 951560.71 frames. ], batch size: 43, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:43:40,816 INFO [finetune.py:976] (2/7) Epoch 8, batch 4900, loss[loss=0.2605, simple_loss=0.3196, pruned_loss=0.1008, over 4850.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.2665, pruned_loss=0.06939, over 953210.32 frames. ], batch size: 31, lr: 3.82e-03, grad_scale: 32.0 +2023-04-26 21:44:34,741 INFO [optim.py:369] (2/7) 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 21:44:46,771 INFO [finetune.py:976] (2/7) Epoch 8, batch 4950, loss[loss=0.2093, simple_loss=0.2822, pruned_loss=0.06821, over 4879.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2674, pruned_loss=0.06909, over 952392.51 frames. ], batch size: 32, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:45:08,583 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45056.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:45:09,179 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0205, 1.4207, 5.3520, 4.9530, 4.6563, 5.2079, 4.6249, 4.7200], + device='cuda:2'), covar=tensor([0.6142, 0.6143, 0.0829, 0.1500, 0.0864, 0.1137, 0.1237, 0.1204], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0305, 0.0405, 0.0409, 0.0345, 0.0402, 0.0314, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:45:35,340 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 5000, loss[loss=0.2057, simple_loss=0.2641, pruned_loss=0.07359, over 4891.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2649, pruned_loss=0.06783, over 953652.75 frames. ], batch size: 32, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:46:02,940 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8168, 2.4747, 1.7730, 1.6272, 1.3243, 1.4107, 1.8750, 1.3335], + device='cuda:2'), covar=tensor([0.1728, 0.1401, 0.1575, 0.1961, 0.2470, 0.2073, 0.1111, 0.2094], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0218, 0.0173, 0.0206, 0.0206, 0.0186, 0.0163, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 21:46:04,174 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45117.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 21:46:13,679 INFO [optim.py:369] (2/7) 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 21:46:16,315 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 21:46:19,279 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 5050, loss[loss=0.1706, simple_loss=0.2347, pruned_loss=0.05322, over 4919.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2617, pruned_loss=0.06671, over 952092.12 frames. ], batch size: 36, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:46:20,656 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 21:46:45,865 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.44 vs. limit=5.0 +2023-04-26 21:46:47,001 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45183.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:46:53,634 INFO [finetune.py:976] (2/7) Epoch 8, batch 5100, loss[loss=0.1804, simple_loss=0.2492, pruned_loss=0.0558, over 4923.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2585, pruned_loss=0.06559, over 952151.26 frames. ], batch size: 43, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:47:02,089 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45206.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:47:13,920 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8462, 2.3475, 2.7762, 3.5225, 2.8681, 2.2619, 2.1780, 2.9729], + device='cuda:2'), covar=tensor([0.3677, 0.3824, 0.1793, 0.2846, 0.2862, 0.2827, 0.4322, 0.2201], + device='cuda:2'), in_proj_covar=tensor([0.0280, 0.0251, 0.0218, 0.0319, 0.0213, 0.0227, 0.0235, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 21:47:18,726 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 21:47:26,587 INFO [finetune.py:976] (2/7) Epoch 8, batch 5150, loss[loss=0.1943, simple_loss=0.2571, pruned_loss=0.06572, over 4690.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2583, pruned_loss=0.0657, over 951731.84 frames. ], batch size: 23, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:47:32,723 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45254.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:47:49,728 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0367, 1.5182, 1.3460, 1.6891, 1.4340, 1.6670, 1.2218, 3.1117], + device='cuda:2'), covar=tensor([0.0717, 0.0802, 0.0858, 0.1286, 0.0679, 0.0660, 0.0876, 0.0198], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:47:53,037 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-26 21:48:00,050 INFO [finetune.py:976] (2/7) Epoch 8, batch 5200, loss[loss=0.2156, simple_loss=0.29, pruned_loss=0.07064, over 4833.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2618, pruned_loss=0.06718, over 951505.63 frames. ], batch size: 49, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:48:44,628 INFO [optim.py:369] (2/7) 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 21:48:56,371 INFO [finetune.py:976] (2/7) Epoch 8, batch 5250, loss[loss=0.1787, simple_loss=0.2604, pruned_loss=0.04852, over 4819.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2629, pruned_loss=0.06719, over 951018.49 frames. ], batch size: 39, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:48:58,926 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45348.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:50:01,146 INFO [finetune.py:976] (2/7) Epoch 8, batch 5300, loss[loss=0.1966, simple_loss=0.2722, pruned_loss=0.06054, over 4846.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.263, pruned_loss=0.067, over 950057.45 frames. ], batch size: 44, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:50:12,157 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0310, 1.3515, 1.2014, 1.5459, 1.3767, 1.5669, 1.2085, 2.4602], + device='cuda:2'), covar=tensor([0.0629, 0.0821, 0.0815, 0.1255, 0.0675, 0.0464, 0.0749, 0.0204], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 21:50:21,698 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45409.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 21:50:23,501 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45412.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 21:50:55,375 INFO [optim.py:369] (2/7) 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 21:50:57,916 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 8, batch 5350, loss[loss=0.1609, simple_loss=0.2434, pruned_loss=0.03923, over 4329.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2633, pruned_loss=0.06675, over 950618.16 frames. ], batch size: 66, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:51:07,852 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9556, 3.8348, 2.7531, 4.5456, 3.9627, 3.9740, 1.4602, 3.8428], + device='cuda:2'), covar=tensor([0.1610, 0.1259, 0.3013, 0.1536, 0.2877, 0.1735, 0.5832, 0.2598], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0252, 0.0308, 0.0300, 0.0250, 0.0273, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 21:51:45,339 INFO [finetune.py:976] (2/7) Epoch 8, batch 5400, loss[loss=0.1818, simple_loss=0.244, pruned_loss=0.05986, over 4767.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2628, pruned_loss=0.06715, over 952036.54 frames. ], batch size: 28, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:51:58,131 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-04-26 21:52:11,221 INFO [optim.py:369] (2/7) 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 21:52:18,380 INFO [finetune.py:976] (2/7) Epoch 8, batch 5450, loss[loss=0.1916, simple_loss=0.2654, pruned_loss=0.05886, over 4903.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2596, pruned_loss=0.06578, over 952714.57 frames. ], batch size: 32, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:52:42,451 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-26 21:52:51,571 INFO [finetune.py:976] (2/7) Epoch 8, batch 5500, loss[loss=0.1485, simple_loss=0.2175, pruned_loss=0.03981, over 4778.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2566, pruned_loss=0.06437, over 952728.08 frames. ], batch size: 26, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:53:13,581 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 21:53:16,426 INFO [optim.py:369] (2/7) 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 21:53:20,501 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5412, 1.2552, 4.4581, 4.1847, 3.9328, 4.2563, 4.0905, 3.9370], + device='cuda:2'), covar=tensor([0.7415, 0.6483, 0.1037, 0.1721, 0.1075, 0.1744, 0.1387, 0.1503], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0305, 0.0405, 0.0410, 0.0346, 0.0403, 0.0314, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:53:24,570 INFO [finetune.py:976] (2/7) Epoch 8, batch 5550, loss[loss=0.1351, simple_loss=0.1908, pruned_loss=0.03972, over 4118.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2584, pruned_loss=0.06529, over 952200.95 frames. ], batch size: 17, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:53:32,084 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 21:53:45,614 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-26 21:54:01,921 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8747, 1.9591, 1.7242, 1.5572, 1.9979, 1.6215, 2.6739, 1.5760], + device='cuda:2'), covar=tensor([0.4320, 0.1883, 0.5013, 0.3420, 0.1936, 0.2592, 0.1323, 0.4551], + device='cuda:2'), in_proj_covar=tensor([0.0349, 0.0352, 0.0434, 0.0367, 0.0392, 0.0385, 0.0385, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:54:02,513 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5887, 2.6113, 2.8935, 3.1314, 2.9169, 2.5254, 1.9960, 2.6410], + device='cuda:2'), covar=tensor([0.0971, 0.0960, 0.0581, 0.0668, 0.0643, 0.0993, 0.0915, 0.0679], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0203, 0.0178, 0.0176, 0.0177, 0.0189, 0.0160, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:54:06,992 INFO [finetune.py:976] (2/7) Epoch 8, batch 5600, loss[loss=0.1864, simple_loss=0.2658, pruned_loss=0.0535, over 4864.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2611, pruned_loss=0.0662, over 952344.20 frames. ], batch size: 34, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:54:12,147 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45702.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:13,284 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45704.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 21:54:17,968 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45712.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 21:54:30,163 INFO [optim.py:369] (2/7) 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 21:54:32,579 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45737.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:36,956 INFO [finetune.py:976] (2/7) Epoch 8, batch 5650, loss[loss=0.2491, simple_loss=0.3176, pruned_loss=0.09035, over 4910.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2632, pruned_loss=0.06655, over 952844.70 frames. ], batch size: 37, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:54:46,842 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45760.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:48,655 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45763.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:54:54,998 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45774.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:55:07,627 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45785.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:55:18,506 INFO [finetune.py:976] (2/7) Epoch 8, batch 5700, loss[loss=0.1879, simple_loss=0.2275, pruned_loss=0.07418, over 4200.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2597, pruned_loss=0.0665, over 933767.19 frames. ], batch size: 18, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:55:59,018 INFO [finetune.py:976] (2/7) Epoch 9, batch 0, loss[loss=0.2463, simple_loss=0.3063, pruned_loss=0.09321, over 4822.00 frames. ], tot_loss[loss=0.2463, simple_loss=0.3063, pruned_loss=0.09321, over 4822.00 frames. ], batch size: 38, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:55:59,018 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 21:56:05,673 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4512, 3.0590, 0.9057, 1.7162, 1.9429, 2.2411, 1.8723, 1.0331], + device='cuda:2'), covar=tensor([0.1337, 0.1087, 0.2037, 0.1380, 0.0998, 0.1010, 0.1612, 0.1782], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0253, 0.0142, 0.0124, 0.0135, 0.0156, 0.0120, 0.0122], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:56:14,948 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 21:56:33,604 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45835.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:57:05,569 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45859.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:57:09,662 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4148, 1.2358, 1.6517, 1.5785, 1.2705, 1.1745, 1.3452, 0.9290], + device='cuda:2'), covar=tensor([0.0654, 0.0859, 0.0492, 0.0726, 0.0845, 0.1538, 0.0718, 0.0831], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0076, 0.0096, 0.0078, 0.0074], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:57:19,426 INFO [finetune.py:976] (2/7) Epoch 9, batch 50, loss[loss=0.2034, simple_loss=0.2672, pruned_loss=0.06982, over 4907.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.267, pruned_loss=0.06864, over 215713.93 frames. ], batch size: 38, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:57:22,855 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7310, 3.9998, 0.6672, 2.0349, 2.3927, 2.6415, 2.3643, 1.0243], + device='cuda:2'), covar=tensor([0.1287, 0.0839, 0.2362, 0.1333, 0.0934, 0.1047, 0.1309, 0.2121], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0252, 0.0141, 0.0124, 0.0135, 0.0155, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 21:57:27,716 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-04-26 21:57:43,636 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.04 vs. limit=5.0 +2023-04-26 21:57:51,877 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 100, loss[loss=0.1767, simple_loss=0.2392, pruned_loss=0.05705, over 4808.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2554, pruned_loss=0.06333, over 380889.33 frames. ], batch size: 25, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:58:01,519 INFO [optim.py:369] (2/7) 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 21:58:16,236 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9687, 1.9803, 1.8994, 1.6547, 2.1660, 1.6819, 2.7057, 1.7118], + device='cuda:2'), covar=tensor([0.3792, 0.1765, 0.4253, 0.3138, 0.1818, 0.2515, 0.1376, 0.4166], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0351, 0.0432, 0.0365, 0.0391, 0.0384, 0.0383, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 21:58:26,128 INFO [finetune.py:976] (2/7) Epoch 9, batch 150, loss[loss=0.1738, simple_loss=0.246, pruned_loss=0.05077, over 4828.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2516, pruned_loss=0.06215, over 508954.83 frames. ], batch size: 33, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:58:47,781 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46004.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 21:58:59,027 INFO [finetune.py:976] (2/7) Epoch 9, batch 200, loss[loss=0.1859, simple_loss=0.2358, pruned_loss=0.068, over 4818.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2492, pruned_loss=0.06158, over 606351.15 frames. ], batch size: 30, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:59:07,995 INFO [optim.py:369] (2/7) 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 21:59:09,448 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-04-26 21:59:19,780 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46052.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 21:59:23,444 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46058.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 21:59:29,166 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-26 21:59:32,404 INFO [finetune.py:976] (2/7) Epoch 9, batch 250, loss[loss=0.1899, simple_loss=0.2631, pruned_loss=0.05835, over 4803.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2555, pruned_loss=0.06416, over 684297.25 frames. ], batch size: 51, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 21:59:34,347 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 22:00:05,477 INFO [finetune.py:976] (2/7) Epoch 9, batch 300, loss[loss=0.2082, simple_loss=0.2789, pruned_loss=0.06874, over 4892.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2604, pruned_loss=0.06532, over 747308.48 frames. ], batch size: 35, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 22:00:08,159 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-26 22:00:10,894 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46130.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:00:12,128 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3204, 3.1010, 0.8897, 1.6943, 1.7025, 2.1854, 1.8048, 1.0175], + device='cuda:2'), covar=tensor([0.1512, 0.1058, 0.2089, 0.1481, 0.1176, 0.1115, 0.1606, 0.1799], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0253, 0.0142, 0.0124, 0.0136, 0.0155, 0.0120, 0.0122], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:00:12,630 INFO [optim.py:369] (2/7) 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 22:00:23,088 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1146, 0.8007, 0.9376, 0.7921, 1.2520, 1.0089, 0.8420, 1.0162], + device='cuda:2'), covar=tensor([0.2232, 0.1960, 0.2645, 0.2048, 0.1359, 0.1745, 0.2259, 0.2490], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0325, 0.0353, 0.0302, 0.0342, 0.0326, 0.0310, 0.0356], + device='cuda:2'), out_proj_covar=tensor([6.5951e-05, 6.9007e-05, 7.6286e-05, 6.2591e-05, 7.1827e-05, 7.0025e-05, + 6.6724e-05, 7.6350e-05], device='cuda:2') +2023-04-26 22:00:42,788 INFO [finetune.py:976] (2/7) Epoch 9, batch 350, loss[loss=0.2652, simple_loss=0.3184, pruned_loss=0.106, over 4833.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.264, pruned_loss=0.06616, over 794541.32 frames. ], batch size: 47, lr: 3.81e-03, grad_scale: 32.0 +2023-04-26 22:00:53,252 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4751, 1.3994, 1.8253, 1.7998, 1.4252, 1.1688, 1.5112, 0.9605], + device='cuda:2'), covar=tensor([0.0607, 0.0799, 0.0521, 0.0705, 0.0797, 0.1283, 0.0772, 0.1003], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0066, 0.0076, 0.0095, 0.0078, 0.0074], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:01:18,233 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7987, 1.7377, 1.6242, 1.4610, 1.9418, 1.6228, 2.4060, 1.4386], + device='cuda:2'), covar=tensor([0.4114, 0.1946, 0.4769, 0.3519, 0.1903, 0.2395, 0.1554, 0.5029], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0348, 0.0428, 0.0362, 0.0387, 0.0380, 0.0379, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:01:38,767 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46215.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:01:48,011 INFO [finetune.py:976] (2/7) Epoch 9, batch 400, loss[loss=0.236, simple_loss=0.2939, pruned_loss=0.08908, over 4890.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2651, pruned_loss=0.06695, over 827987.68 frames. ], batch size: 43, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:02:01,098 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 9, batch 450, loss[loss=0.2283, simple_loss=0.2856, pruned_loss=0.08556, over 4910.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2635, pruned_loss=0.06626, over 856155.81 frames. ], batch size: 46, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:02:29,459 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46276.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:03:00,648 INFO [finetune.py:976] (2/7) Epoch 9, batch 500, loss[loss=0.1792, simple_loss=0.2533, pruned_loss=0.05252, over 4735.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2604, pruned_loss=0.06528, over 877193.94 frames. ], batch size: 23, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:03:07,759 INFO [optim.py:369] (2/7) 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 22:03:10,766 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46337.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:03:13,175 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1059, 1.6554, 1.4528, 1.8193, 1.7440, 2.0616, 1.3898, 3.5798], + device='cuda:2'), covar=tensor([0.0669, 0.0789, 0.0777, 0.1188, 0.0626, 0.0469, 0.0787, 0.0135], + device='cuda:2'), in_proj_covar=tensor([0.0039, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 22:03:16,540 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 22:03:25,994 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46358.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:03:26,695 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-26 22:03:32,663 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7532, 1.5172, 4.6256, 4.3058, 4.0056, 4.4382, 4.2288, 4.0098], + device='cuda:2'), covar=tensor([0.7208, 0.5977, 0.0948, 0.1647, 0.1034, 0.1735, 0.1367, 0.1525], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0309, 0.0409, 0.0415, 0.0350, 0.0408, 0.0320, 0.0371], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:03:34,349 INFO [finetune.py:976] (2/7) Epoch 9, batch 550, loss[loss=0.2171, simple_loss=0.272, pruned_loss=0.08112, over 4924.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2566, pruned_loss=0.06404, over 894486.19 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:03:43,968 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0208, 2.4765, 1.2078, 1.3931, 1.9780, 1.2775, 2.8264, 1.6469], + device='cuda:2'), covar=tensor([0.0631, 0.0850, 0.0705, 0.1032, 0.0418, 0.0851, 0.0256, 0.0571], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0052, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 22:03:48,181 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.95 vs. limit=5.0 +2023-04-26 22:03:57,941 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 600, loss[loss=0.2045, simple_loss=0.2778, pruned_loss=0.06559, over 4863.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2592, pruned_loss=0.06524, over 907749.06 frames. ], batch size: 31, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:04:12,583 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46430.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:04:14,305 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 9, batch 650, loss[loss=0.1803, simple_loss=0.2595, pruned_loss=0.05053, over 4826.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2617, pruned_loss=0.06589, over 918093.92 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:04:44,728 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46478.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:05:10,663 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46515.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:05:14,904 INFO [finetune.py:976] (2/7) Epoch 9, batch 700, loss[loss=0.2162, simple_loss=0.2771, pruned_loss=0.07758, over 4752.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2647, pruned_loss=0.06677, over 926456.92 frames. ], batch size: 27, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:05:21,595 INFO [optim.py:369] (2/7) 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 22:05:21,740 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4037, 1.0416, 1.1904, 1.1174, 1.6080, 1.2764, 1.0539, 1.1477], + device='cuda:2'), covar=tensor([0.1515, 0.1249, 0.1548, 0.1503, 0.0663, 0.1277, 0.1644, 0.1687], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0325, 0.0354, 0.0302, 0.0342, 0.0325, 0.0310, 0.0357], + device='cuda:2'), out_proj_covar=tensor([6.6164e-05, 6.9013e-05, 7.6510e-05, 6.2532e-05, 7.1924e-05, 6.9859e-05, + 6.6639e-05, 7.6549e-05], device='cuda:2') +2023-04-26 22:05:39,612 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3886, 1.7221, 1.7029, 2.1630, 2.4285, 2.0120, 1.9494, 1.7211], + device='cuda:2'), covar=tensor([0.2512, 0.1885, 0.2308, 0.1906, 0.1684, 0.2396, 0.2704, 0.2376], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0325, 0.0354, 0.0302, 0.0342, 0.0325, 0.0310, 0.0357], + device='cuda:2'), out_proj_covar=tensor([6.6200e-05, 6.8978e-05, 7.6506e-05, 6.2536e-05, 7.1933e-05, 6.9866e-05, + 6.6612e-05, 7.6577e-05], device='cuda:2') +2023-04-26 22:05:43,203 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46563.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:05:53,547 INFO [finetune.py:976] (2/7) Epoch 9, batch 750, loss[loss=0.2382, simple_loss=0.2944, pruned_loss=0.09106, over 4776.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2662, pruned_loss=0.06732, over 930301.29 frames. ], batch size: 51, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:06:18,434 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46595.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:06:28,895 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46602.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:06:58,617 INFO [finetune.py:976] (2/7) Epoch 9, batch 800, loss[loss=0.2255, simple_loss=0.2819, pruned_loss=0.08459, over 4925.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2653, pruned_loss=0.06656, over 938637.15 frames. ], batch size: 38, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:07:09,959 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46632.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:07:10,488 INFO [optim.py:369] (2/7) 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 22:07:11,908 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6929, 1.1383, 1.6452, 2.1250, 1.7928, 1.6038, 1.6357, 1.7157], + device='cuda:2'), covar=tensor([0.6354, 0.9189, 0.8887, 0.8943, 0.8582, 1.0911, 1.1422, 0.9290], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0420, 0.0503, 0.0523, 0.0437, 0.0455, 0.0466, 0.0465], + device='cuda:2'), out_proj_covar=tensor([9.9311e-05, 1.0430e-04, 1.1355e-04, 1.2415e-04, 1.0616e-04, 1.1011e-04, + 1.1221e-04, 1.1264e-04], device='cuda:2') +2023-04-26 22:07:26,184 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46656.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 22:07:31,867 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46663.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:07:37,208 INFO [finetune.py:976] (2/7) Epoch 9, batch 850, loss[loss=0.1561, simple_loss=0.2243, pruned_loss=0.0439, over 4754.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2624, pruned_loss=0.0653, over 943067.29 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:07:45,220 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5692, 0.9533, 1.4878, 2.0202, 1.6836, 1.5228, 1.5253, 1.6015], + device='cuda:2'), covar=tensor([0.6599, 0.8669, 0.8798, 0.8822, 0.8127, 1.0114, 1.0535, 0.9074], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0421, 0.0503, 0.0523, 0.0437, 0.0455, 0.0466, 0.0465], + device='cuda:2'), out_proj_covar=tensor([9.9428e-05, 1.0450e-04, 1.1358e-04, 1.2419e-04, 1.0624e-04, 1.1012e-04, + 1.1228e-04, 1.1255e-04], device='cuda:2') +2023-04-26 22:08:10,356 INFO [finetune.py:976] (2/7) Epoch 9, batch 900, loss[loss=0.1673, simple_loss=0.2357, pruned_loss=0.04947, over 4791.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2598, pruned_loss=0.06465, over 946482.05 frames. ], batch size: 29, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:08:17,018 INFO [optim.py:369] (2/7) 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 22:08:34,775 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6651, 1.5141, 1.8497, 2.0411, 1.7137, 1.3015, 1.6508, 1.0097], + device='cuda:2'), covar=tensor([0.0715, 0.0820, 0.0785, 0.0885, 0.0791, 0.1151, 0.0860, 0.1029], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0067, 0.0076, 0.0096, 0.0078, 0.0074], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:08:41,122 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6191, 1.8029, 1.8679, 2.3516, 2.5367, 2.2360, 2.0454, 1.9054], + device='cuda:2'), covar=tensor([0.1739, 0.1892, 0.1839, 0.1738, 0.1296, 0.2031, 0.2500, 0.1882], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0325, 0.0354, 0.0301, 0.0340, 0.0325, 0.0309, 0.0355], + device='cuda:2'), out_proj_covar=tensor([6.6189e-05, 6.8978e-05, 7.6382e-05, 6.2247e-05, 7.1474e-05, 6.9741e-05, + 6.6494e-05, 7.6194e-05], device='cuda:2') +2023-04-26 22:08:43,449 INFO [finetune.py:976] (2/7) Epoch 9, batch 950, loss[loss=0.203, simple_loss=0.2621, pruned_loss=0.07194, over 4831.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.259, pruned_loss=0.06502, over 948170.69 frames. ], batch size: 40, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:09:08,144 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46810.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:09:16,375 INFO [finetune.py:976] (2/7) Epoch 9, batch 1000, loss[loss=0.1996, simple_loss=0.2716, pruned_loss=0.06384, over 4796.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2615, pruned_loss=0.06582, over 949976.37 frames. ], batch size: 29, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:09:22,998 INFO [optim.py:369] (2/7) 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 22:09:44,514 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3782, 1.7460, 1.6621, 2.2323, 2.3742, 2.0094, 1.8416, 1.7427], + device='cuda:2'), covar=tensor([0.1896, 0.1842, 0.2137, 0.1935, 0.1577, 0.1863, 0.2439, 0.2022], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0324, 0.0353, 0.0301, 0.0339, 0.0324, 0.0309, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.5862e-05, 6.8876e-05, 7.6199e-05, 6.2187e-05, 7.1352e-05, 6.9590e-05, + 6.6504e-05, 7.6056e-05], device='cuda:2') +2023-04-26 22:09:49,154 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46871.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:09:49,626 INFO [finetune.py:976] (2/7) Epoch 9, batch 1050, loss[loss=0.2042, simple_loss=0.268, pruned_loss=0.07019, over 4752.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2639, pruned_loss=0.06597, over 951351.51 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 64.0 +2023-04-26 22:10:22,798 INFO [finetune.py:976] (2/7) Epoch 9, batch 1100, loss[loss=0.2358, simple_loss=0.297, pruned_loss=0.08736, over 4807.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2656, pruned_loss=0.06722, over 951554.24 frames. ], batch size: 51, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:10:29,465 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:10:34,372 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 22:10:39,330 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-26 22:10:40,915 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46951.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 22:10:40,955 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9660, 1.6387, 1.8749, 2.2777, 2.2926, 1.8704, 1.5439, 2.0045], + device='cuda:2'), covar=tensor([0.0970, 0.1249, 0.0710, 0.0601, 0.0606, 0.0871, 0.0965, 0.0604], + device='cuda:2'), in_proj_covar=tensor([0.0201, 0.0207, 0.0184, 0.0179, 0.0181, 0.0194, 0.0163, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:10:45,717 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 1150, loss[loss=0.1865, simple_loss=0.2532, pruned_loss=0.05994, over 4757.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2656, pruned_loss=0.067, over 951635.69 frames. ], batch size: 27, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:11:01,459 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46980.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:11:41,175 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9670, 1.6839, 4.0776, 3.7877, 3.5633, 3.7828, 3.6756, 3.6055], + device='cuda:2'), covar=tensor([0.6378, 0.5207, 0.1022, 0.1679, 0.1064, 0.1711, 0.3355, 0.1420], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0311, 0.0410, 0.0416, 0.0352, 0.0409, 0.0321, 0.0374], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:11:43,958 INFO [finetune.py:976] (2/7) Epoch 9, batch 1200, loss[loss=0.1989, simple_loss=0.2574, pruned_loss=0.07023, over 4896.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2637, pruned_loss=0.06614, over 953113.71 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:12:03,717 INFO [optim.py:369] (2/7) 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 22:12:06,338 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2338, 1.6314, 2.0485, 2.4964, 2.0235, 1.6115, 1.3156, 1.9161], + device='cuda:2'), covar=tensor([0.3615, 0.3636, 0.1843, 0.2779, 0.3074, 0.2971, 0.4709, 0.2407], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0254, 0.0221, 0.0322, 0.0215, 0.0230, 0.0237, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 22:12:23,592 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-26 22:12:55,844 INFO [finetune.py:976] (2/7) Epoch 9, batch 1250, loss[loss=0.1899, simple_loss=0.2642, pruned_loss=0.05781, over 4824.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2619, pruned_loss=0.06581, over 953750.68 frames. ], batch size: 40, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:13:21,717 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8714, 1.2144, 3.2987, 3.0535, 2.9429, 3.2164, 3.2158, 2.9099], + device='cuda:2'), covar=tensor([0.7551, 0.5331, 0.1548, 0.2290, 0.1375, 0.1799, 0.1500, 0.1556], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0311, 0.0410, 0.0416, 0.0352, 0.0409, 0.0320, 0.0373], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:13:41,580 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4659, 2.8777, 0.9694, 1.6549, 2.2663, 1.5399, 4.2512, 2.2086], + device='cuda:2'), covar=tensor([0.0602, 0.0727, 0.0865, 0.1300, 0.0556, 0.0963, 0.0235, 0.0568], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0048, 0.0052, 0.0052, 0.0078, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 22:14:00,070 INFO [finetune.py:976] (2/7) Epoch 9, batch 1300, loss[loss=0.2172, simple_loss=0.2834, pruned_loss=0.07547, over 4834.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2583, pruned_loss=0.06459, over 954331.75 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:14:15,264 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47166.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:15:07,955 INFO [finetune.py:976] (2/7) Epoch 9, batch 1350, loss[loss=0.1902, simple_loss=0.2614, pruned_loss=0.05945, over 4932.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2599, pruned_loss=0.06535, over 953981.04 frames. ], batch size: 38, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:15:55,775 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 22:15:58,013 INFO [finetune.py:976] (2/7) Epoch 9, batch 1400, loss[loss=0.2013, simple_loss=0.275, pruned_loss=0.06375, over 4893.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2611, pruned_loss=0.06506, over 954945.31 frames. ], batch size: 43, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:16:06,775 INFO [optim.py:369] (2/7) 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 22:16:11,713 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47241.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:17,779 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47251.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:22,000 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 1450, loss[loss=0.2549, simple_loss=0.3044, pruned_loss=0.1027, over 4857.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.263, pruned_loss=0.06594, over 955507.03 frames. ], batch size: 31, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:16:49,492 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1694, 1.6359, 1.5606, 2.0264, 1.8028, 1.8361, 1.3880, 4.3636], + device='cuda:2'), covar=tensor([0.0689, 0.0943, 0.0896, 0.1304, 0.0742, 0.0643, 0.0921, 0.0129], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 22:16:50,058 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47299.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:51,898 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47302.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:54,266 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47306.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:16:58,562 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6288, 1.6472, 1.6248, 1.3503, 1.7624, 1.4384, 2.3025, 1.4569], + device='cuda:2'), covar=tensor([0.4167, 0.2047, 0.5491, 0.3062, 0.1820, 0.2766, 0.1678, 0.5301], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0351, 0.0431, 0.0364, 0.0390, 0.0387, 0.0383, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:17:03,913 INFO [finetune.py:976] (2/7) Epoch 9, batch 1500, loss[loss=0.1927, simple_loss=0.2765, pruned_loss=0.05446, over 4848.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2643, pruned_loss=0.06645, over 956648.83 frames. ], batch size: 44, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:17:12,626 INFO [optim.py:369] (2/7) 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 22:18:03,821 INFO [finetune.py:976] (2/7) Epoch 9, batch 1550, loss[loss=0.1818, simple_loss=0.254, pruned_loss=0.05481, over 4817.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2642, pruned_loss=0.06579, over 957369.52 frames. ], batch size: 38, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:18:13,654 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47379.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:18:48,888 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-26 22:18:49,437 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4317, 2.4558, 2.7503, 3.0794, 2.9363, 2.4251, 1.9725, 2.5605], + device='cuda:2'), covar=tensor([0.1070, 0.1012, 0.0618, 0.0720, 0.0602, 0.0959, 0.0989, 0.0676], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0205, 0.0182, 0.0177, 0.0178, 0.0190, 0.0161, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:19:10,271 INFO [finetune.py:976] (2/7) Epoch 9, batch 1600, loss[loss=0.1757, simple_loss=0.2317, pruned_loss=0.05983, over 4824.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2603, pruned_loss=0.0646, over 956165.96 frames. ], batch size: 25, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:19:23,850 INFO [optim.py:369] (2/7) 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 22:19:34,492 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47440.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:19:46,071 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6851, 1.9893, 1.8909, 2.0463, 1.8571, 2.0168, 1.9576, 1.8884], + device='cuda:2'), covar=tensor([0.6166, 0.9191, 0.7646, 0.6798, 0.8140, 1.1020, 1.0039, 0.8200], + device='cuda:2'), in_proj_covar=tensor([0.0325, 0.0386, 0.0319, 0.0328, 0.0344, 0.0406, 0.0366, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 22:20:08,741 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 1650, loss[loss=0.153, simple_loss=0.2107, pruned_loss=0.04768, over 4675.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2564, pruned_loss=0.063, over 955241.63 frames. ], batch size: 23, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:20:46,553 INFO [scaling.py:679] (2/7) 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] (2/7) 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] (2/7) Epoch 9, batch 1700, loss[loss=0.1908, simple_loss=0.2447, pruned_loss=0.06844, over 4768.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.255, pruned_loss=0.06324, over 955895.22 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:21:12,554 INFO [optim.py:369] (2/7) 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 22:21:14,956 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1361, 0.8351, 0.9269, 0.8268, 1.2722, 0.9929, 0.8757, 0.9758], + device='cuda:2'), covar=tensor([0.1513, 0.1422, 0.1987, 0.1397, 0.0968, 0.1274, 0.1726, 0.1949], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0328, 0.0358, 0.0303, 0.0340, 0.0328, 0.0311, 0.0358], + device='cuda:2'), 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:2') +2023-04-26 22:21:38,481 INFO [finetune.py:976] (2/7) Epoch 9, batch 1750, loss[loss=0.2049, simple_loss=0.2763, pruned_loss=0.06678, over 4746.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2581, pruned_loss=0.06469, over 955206.13 frames. ], batch size: 28, lr: 3.80e-03, grad_scale: 32.0 +2023-04-26 22:21:50,032 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1024, 2.5138, 1.0708, 1.3814, 1.9384, 1.3830, 3.5995, 2.0315], + device='cuda:2'), covar=tensor([0.0693, 0.0909, 0.0892, 0.1372, 0.0580, 0.0974, 0.0225, 0.0577], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0068, 0.0050, 0.0047, 0.0052, 0.0053, 0.0079, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 22:21:54,722 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47597.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:22:11,819 INFO [finetune.py:976] (2/7) Epoch 9, batch 1800, loss[loss=0.1722, simple_loss=0.242, pruned_loss=0.05122, over 4867.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2612, pruned_loss=0.06543, over 955484.96 frames. ], batch size: 44, lr: 3.79e-03, grad_scale: 32.0 +2023-04-26 22:22:16,225 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5084, 1.7391, 0.7830, 1.2509, 1.8445, 1.3684, 1.3324, 1.4388], + device='cuda:2'), covar=tensor([0.0565, 0.0395, 0.0409, 0.0622, 0.0310, 0.0600, 0.0578, 0.0606], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 22:22:19,164 INFO [optim.py:369] (2/7) 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 22:22:37,659 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-26 22:22:45,310 INFO [finetune.py:976] (2/7) Epoch 9, batch 1850, loss[loss=0.2288, simple_loss=0.2888, pruned_loss=0.08446, over 4729.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2632, pruned_loss=0.06629, over 954218.35 frames. ], batch size: 54, lr: 3.79e-03, grad_scale: 32.0 +2023-04-26 22:22:49,156 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8023, 1.3348, 1.3468, 1.5802, 2.0315, 1.6424, 1.3422, 1.2560], + device='cuda:2'), covar=tensor([0.1658, 0.1746, 0.1953, 0.1412, 0.0814, 0.1562, 0.2319, 0.2109], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0328, 0.0358, 0.0303, 0.0341, 0.0329, 0.0312, 0.0359], + device='cuda:2'), out_proj_covar=tensor([6.6486e-05, 6.9716e-05, 7.7308e-05, 6.2727e-05, 7.1550e-05, 7.0739e-05, + 6.7069e-05, 7.7107e-05], device='cuda:2') +2023-04-26 22:23:44,013 INFO [finetune.py:976] (2/7) Epoch 9, batch 1900, loss[loss=0.2357, simple_loss=0.2859, pruned_loss=0.09271, over 4903.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.264, pruned_loss=0.06582, over 954520.87 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:24:02,111 INFO [optim.py:369] (2/7) 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 22:24:02,198 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47735.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:04,089 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47738.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:33,493 INFO [finetune.py:976] (2/7) Epoch 9, batch 1950, loss[loss=0.1902, simple_loss=0.2536, pruned_loss=0.06337, over 4907.00 frames. ], tot_loss[loss=0.195, simple_loss=0.261, pruned_loss=0.06452, over 954958.73 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:24:43,265 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47788.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:50,421 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47799.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:24:57,030 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47809.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:25:10,595 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1162, 1.5062, 1.9469, 2.0963, 1.9084, 1.5291, 1.0250, 1.6100], + device='cuda:2'), covar=tensor([0.3938, 0.4054, 0.1932, 0.2922, 0.3285, 0.3152, 0.5122, 0.2810], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0253, 0.0221, 0.0321, 0.0215, 0.0230, 0.0237, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 22:25:11,659 INFO [finetune.py:976] (2/7) Epoch 9, batch 2000, loss[loss=0.2674, simple_loss=0.3053, pruned_loss=0.1148, over 4236.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2583, pruned_loss=0.06382, over 954096.29 frames. ], batch size: 65, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:25:30,750 INFO [optim.py:369] (2/7) 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 22:25:44,893 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:26:02,239 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5447, 1.1257, 1.3316, 1.2221, 1.6855, 1.3774, 1.1028, 1.2970], + device='cuda:2'), covar=tensor([0.1524, 0.1334, 0.1875, 0.1461, 0.0830, 0.1451, 0.1752, 0.1808], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0325, 0.0353, 0.0299, 0.0338, 0.0326, 0.0307, 0.0355], + device='cuda:2'), out_proj_covar=tensor([6.5495e-05, 6.9011e-05, 7.6274e-05, 6.1887e-05, 7.0950e-05, 6.9982e-05, + 6.6090e-05, 7.6165e-05], device='cuda:2') +2023-04-26 22:26:17,406 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47870.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 22:26:24,628 INFO [finetune.py:976] (2/7) Epoch 9, batch 2050, loss[loss=0.201, simple_loss=0.2447, pruned_loss=0.07867, over 4732.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2541, pruned_loss=0.06245, over 955785.07 frames. ], batch size: 59, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:26:57,932 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47897.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:27:30,785 INFO [finetune.py:976] (2/7) Epoch 9, batch 2100, loss[loss=0.2651, simple_loss=0.3128, pruned_loss=0.1087, over 4016.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2527, pruned_loss=0.06174, over 956289.07 frames. ], batch size: 65, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:27:39,268 INFO [optim.py:369] (2/7) 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 22:27:45,502 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 2150, loss[loss=0.2613, simple_loss=0.3227, pruned_loss=0.09995, over 4866.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2559, pruned_loss=0.06326, over 953684.27 frames. ], batch size: 34, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:29:17,268 INFO [finetune.py:976] (2/7) Epoch 9, batch 2200, loss[loss=0.1859, simple_loss=0.2586, pruned_loss=0.05661, over 4823.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2596, pruned_loss=0.064, over 955848.89 frames. ], batch size: 33, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:29:20,760 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3366, 1.6024, 1.3955, 1.5144, 1.3732, 1.2862, 1.3918, 1.1200], + device='cuda:2'), covar=tensor([0.1777, 0.1305, 0.1035, 0.1280, 0.3367, 0.1311, 0.1822, 0.2267], + device='cuda:2'), in_proj_covar=tensor([0.0299, 0.0319, 0.0230, 0.0293, 0.0317, 0.0273, 0.0261, 0.0284], + device='cuda:2'), out_proj_covar=tensor([1.2100e-04, 1.2868e-04, 9.2479e-05, 1.1715e-04, 1.3034e-04, 1.1018e-04, + 1.0669e-04, 1.1403e-04], device='cuda:2') +2023-04-26 22:29:32,259 INFO [optim.py:369] (2/7) 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 22:29:32,363 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48035.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:30:08,349 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-26 22:30:13,116 INFO [finetune.py:976] (2/7) Epoch 9, batch 2250, loss[loss=0.2251, simple_loss=0.2913, pruned_loss=0.0794, over 4765.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2614, pruned_loss=0.06445, over 954597.59 frames. ], batch size: 28, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:30:18,385 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48072.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:30:32,862 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48083.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:30:45,746 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48094.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:31:13,253 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3172, 2.8084, 0.9929, 1.3889, 2.0201, 1.2740, 3.7998, 1.8663], + device='cuda:2'), covar=tensor([0.0615, 0.0645, 0.0844, 0.1261, 0.0527, 0.1004, 0.0253, 0.0623], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0067, 0.0050, 0.0047, 0.0052, 0.0052, 0.0078, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-26 22:31:21,271 INFO [finetune.py:976] (2/7) Epoch 9, batch 2300, loss[loss=0.1834, simple_loss=0.253, pruned_loss=0.0569, over 4715.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2609, pruned_loss=0.06401, over 953378.03 frames. ], batch size: 59, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:31:35,707 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:31:50,257 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48144.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:32:14,435 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48165.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:32:25,261 INFO [finetune.py:976] (2/7) Epoch 9, batch 2350, loss[loss=0.1969, simple_loss=0.2614, pruned_loss=0.06613, over 4907.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.259, pruned_loss=0.06356, over 955133.44 frames. ], batch size: 43, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:33:19,207 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48212.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:33:30,439 INFO [finetune.py:976] (2/7) Epoch 9, batch 2400, loss[loss=0.2057, simple_loss=0.2649, pruned_loss=0.07323, over 4847.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2556, pruned_loss=0.06247, over 956100.05 frames. ], batch size: 44, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:33:45,096 INFO [optim.py:369] (2/7) 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 22:34:24,262 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5297, 3.4591, 2.4944, 2.7948, 2.4123, 2.6964, 2.7188, 2.3518], + device='cuda:2'), covar=tensor([0.2370, 0.1237, 0.1082, 0.1592, 0.3245, 0.1457, 0.2333, 0.3180], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0317, 0.0228, 0.0290, 0.0313, 0.0271, 0.0259, 0.0281], + device='cuda:2'), out_proj_covar=tensor([1.2011e-04, 1.2767e-04, 9.1951e-05, 1.1633e-04, 1.2876e-04, 1.0947e-04, + 1.0594e-04, 1.1294e-04], device='cuda:2') +2023-04-26 22:34:26,769 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.60 vs. limit=5.0 +2023-04-26 22:34:34,223 INFO [finetune.py:976] (2/7) Epoch 9, batch 2450, loss[loss=0.216, simple_loss=0.2718, pruned_loss=0.08013, over 4770.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2536, pruned_loss=0.06253, over 954727.35 frames. ], batch size: 26, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:34:34,997 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48273.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:35:25,356 INFO [finetune.py:976] (2/7) Epoch 9, batch 2500, loss[loss=0.2068, simple_loss=0.2779, pruned_loss=0.06784, over 4911.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2558, pruned_loss=0.06326, over 954013.61 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:35:26,642 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48324.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:35:34,244 INFO [optim.py:369] (2/7) 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 22:35:50,864 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-26 22:35:52,468 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3830, 3.2769, 2.5333, 3.8526, 3.3646, 3.3614, 1.3324, 3.3393], + device='cuda:2'), covar=tensor([0.1634, 0.1455, 0.3235, 0.2150, 0.3371, 0.1804, 0.5875, 0.2431], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0218, 0.0252, 0.0307, 0.0300, 0.0251, 0.0271, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 22:36:14,628 INFO [finetune.py:976] (2/7) Epoch 9, batch 2550, loss[loss=0.1755, simple_loss=0.2539, pruned_loss=0.04848, over 4901.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.259, pruned_loss=0.06336, over 955473.13 frames. ], batch size: 35, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:36:32,430 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 22:36:34,703 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5513, 2.0072, 1.6494, 1.8465, 1.4554, 1.6202, 1.5399, 1.3866], + device='cuda:2'), covar=tensor([0.1863, 0.1222, 0.0964, 0.1257, 0.3271, 0.1164, 0.1854, 0.2354], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0315, 0.0227, 0.0288, 0.0311, 0.0269, 0.0257, 0.0279], + device='cuda:2'), out_proj_covar=tensor([1.1891e-04, 1.2673e-04, 9.1344e-05, 1.1525e-04, 1.2796e-04, 1.0857e-04, + 1.0516e-04, 1.1199e-04], device='cuda:2') +2023-04-26 22:36:34,719 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48385.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:36:53,741 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48394.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:36:55,539 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48397.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:06,247 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3720, 1.1925, 1.6449, 1.5223, 1.2580, 1.0866, 1.2482, 0.7626], + device='cuda:2'), covar=tensor([0.0584, 0.0789, 0.0484, 0.0675, 0.0845, 0.1335, 0.0694, 0.0815], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0071, 0.0066, 0.0076, 0.0095, 0.0078, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:37:18,655 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9361, 1.5646, 2.0131, 2.3205, 2.0431, 1.9095, 1.9804, 1.9804], + device='cuda:2'), covar=tensor([0.5802, 0.7461, 0.8156, 0.8271, 0.7382, 1.0237, 1.0054, 0.8852], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0419, 0.0501, 0.0521, 0.0437, 0.0456, 0.0468, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9282e-05, 1.0397e-04, 1.1327e-04, 1.2391e-04, 1.0624e-04, 1.1024e-04, + 1.1251e-04, 1.1266e-04], device='cuda:2') +2023-04-26 22:37:18,801 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-04-26 22:37:20,469 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7012, 2.2630, 1.6498, 1.4957, 1.2411, 1.2827, 1.8262, 1.1978], + device='cuda:2'), covar=tensor([0.1843, 0.1518, 0.1684, 0.2177, 0.2648, 0.2127, 0.1150, 0.2269], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0216, 0.0171, 0.0205, 0.0205, 0.0183, 0.0161, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 22:37:28,005 INFO [finetune.py:976] (2/7) Epoch 9, batch 2600, loss[loss=0.1951, simple_loss=0.2731, pruned_loss=0.05851, over 4766.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2617, pruned_loss=0.06466, over 953894.67 frames. ], batch size: 28, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:37:28,235 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-26 22:37:28,726 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48423.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:31,744 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:37:49,751 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 22:37:51,881 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6224, 1.5122, 0.7279, 1.2698, 1.6521, 1.4692, 1.3619, 1.4283], + device='cuda:2'), covar=tensor([0.0567, 0.0428, 0.0423, 0.0616, 0.0309, 0.0598, 0.0569, 0.0644], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 22:37:52,941 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48442.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:37:54,158 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48444.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:38:03,691 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.43 vs. limit=5.0 +2023-04-26 22:38:13,996 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48458.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:38:24,212 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48465.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:38:34,661 INFO [finetune.py:976] (2/7) Epoch 9, batch 2650, loss[loss=0.174, simple_loss=0.248, pruned_loss=0.04995, over 4879.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2643, pruned_loss=0.06618, over 953897.57 frames. ], batch size: 32, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:38:46,965 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) 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] (2/7) Epoch 9, batch 2700, loss[loss=0.1683, simple_loss=0.2399, pruned_loss=0.04837, over 4902.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2623, pruned_loss=0.06512, over 954376.11 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:39:40,872 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48568.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 22:40:28,390 INFO [finetune.py:976] (2/7) Epoch 9, batch 2750, loss[loss=0.1994, simple_loss=0.2632, pruned_loss=0.06777, over 4751.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2596, pruned_loss=0.06439, over 955680.22 frames. ], batch size: 27, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:41:24,376 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2792, 1.3550, 1.5225, 1.6344, 1.6388, 1.2441, 0.9890, 1.4821], + device='cuda:2'), covar=tensor([0.0921, 0.1369, 0.0884, 0.0639, 0.0682, 0.0955, 0.0886, 0.0691], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0206, 0.0184, 0.0178, 0.0180, 0.0191, 0.0161, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:41:33,773 INFO [finetune.py:976] (2/7) Epoch 9, batch 2800, loss[loss=0.1533, simple_loss=0.2186, pruned_loss=0.04397, over 4811.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2555, pruned_loss=0.06305, over 952726.27 frames. ], batch size: 25, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:41:34,493 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0289, 1.0037, 1.2841, 1.1532, 1.0126, 0.9135, 1.0002, 0.5788], + device='cuda:2'), covar=tensor([0.0569, 0.0577, 0.0527, 0.0596, 0.0757, 0.1254, 0.0473, 0.0855], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0072, 0.0066, 0.0076, 0.0095, 0.0078, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:41:46,303 INFO [optim.py:369] (2/7) 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 22:41:56,226 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-04-26 22:42:08,358 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8245, 2.3460, 1.9694, 2.1642, 1.6126, 1.9579, 1.8836, 1.5247], + device='cuda:2'), covar=tensor([0.1962, 0.1149, 0.0819, 0.1123, 0.3186, 0.1111, 0.2171, 0.2536], + device='cuda:2'), in_proj_covar=tensor([0.0298, 0.0319, 0.0229, 0.0290, 0.0315, 0.0273, 0.0259, 0.0282], + device='cuda:2'), out_proj_covar=tensor([1.2054e-04, 1.2847e-04, 9.2210e-05, 1.1629e-04, 1.2935e-04, 1.0995e-04, + 1.0602e-04, 1.1326e-04], device='cuda:2') +2023-04-26 22:42:38,717 INFO [finetune.py:976] (2/7) Epoch 9, batch 2850, loss[loss=0.2016, simple_loss=0.2591, pruned_loss=0.07204, over 4827.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2535, pruned_loss=0.06224, over 952334.09 frames. ], batch size: 33, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:42:49,437 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2312, 1.4093, 3.8434, 3.5617, 3.4002, 3.6803, 3.6928, 3.4024], + device='cuda:2'), covar=tensor([0.7115, 0.5358, 0.1193, 0.1874, 0.1318, 0.2017, 0.1339, 0.1444], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0311, 0.0411, 0.0414, 0.0354, 0.0408, 0.0318, 0.0373], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:42:49,439 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48680.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:43:01,343 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4973, 1.4301, 0.7419, 1.2246, 1.5208, 1.3736, 1.2961, 1.3177], + device='cuda:2'), covar=tensor([0.0558, 0.0420, 0.0428, 0.0612, 0.0333, 0.0573, 0.0531, 0.0613], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 22:43:44,377 INFO [finetune.py:976] (2/7) Epoch 9, batch 2900, loss[loss=0.1859, simple_loss=0.2483, pruned_loss=0.06176, over 4742.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.257, pruned_loss=0.06344, over 952230.80 frames. ], batch size: 26, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:43:48,142 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48728.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:43:52,841 INFO [optim.py:369] (2/7) 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 22:43:54,153 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9577, 4.0379, 2.8214, 4.5905, 4.0553, 4.0363, 1.9415, 3.8432], + device='cuda:2'), covar=tensor([0.1720, 0.0938, 0.2921, 0.1381, 0.2579, 0.1563, 0.5306, 0.2615], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0217, 0.0251, 0.0305, 0.0299, 0.0250, 0.0270, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 22:44:04,269 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48753.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:44:17,828 INFO [finetune.py:976] (2/7) Epoch 9, batch 2950, loss[loss=0.1603, simple_loss=0.229, pruned_loss=0.04576, over 4819.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2597, pruned_loss=0.06404, over 952341.50 frames. ], batch size: 25, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:44:19,208 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8100, 1.7111, 2.0497, 2.0133, 1.8363, 1.7544, 1.8698, 1.9415], + device='cuda:2'), covar=tensor([0.9101, 1.2299, 1.4578, 1.4860, 1.1412, 1.6188, 1.6497, 1.4904], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0423, 0.0505, 0.0524, 0.0439, 0.0459, 0.0471, 0.0468], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:44:20,345 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48776.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:44:22,221 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48779.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:44:25,282 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3287, 1.2527, 4.0153, 3.7300, 3.5374, 3.8062, 3.7859, 3.5650], + device='cuda:2'), covar=tensor([0.7848, 0.6256, 0.1141, 0.1835, 0.1273, 0.1668, 0.1668, 0.1672], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0312, 0.0412, 0.0416, 0.0355, 0.0410, 0.0319, 0.0375], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 22:44:39,683 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4373, 1.2745, 1.6963, 1.6642, 1.3464, 1.0757, 1.4447, 0.9662], + device='cuda:2'), covar=tensor([0.0661, 0.0801, 0.0477, 0.0754, 0.0876, 0.1269, 0.0623, 0.0779], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0071, 0.0066, 0.0076, 0.0095, 0.0078, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:45:10,534 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48821.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:45:11,035 INFO [finetune.py:976] (2/7) Epoch 9, batch 3000, loss[loss=0.1987, simple_loss=0.2715, pruned_loss=0.06291, over 4788.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2624, pruned_loss=0.06457, over 953556.42 frames. ], batch size: 29, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:45:11,035 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 22:45:27,495 INFO [finetune.py:1010] (2/7) Epoch 9, validation: loss=0.1543, simple_loss=0.2267, pruned_loss=0.04097, over 2265189.00 frames. +2023-04-26 22:45:27,496 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 22:45:46,035 INFO [optim.py:369] (2/7) 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 22:45:47,989 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48838.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:46:09,312 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7333, 1.3479, 1.3221, 1.6326, 1.9642, 1.5841, 1.3247, 1.2497], + device='cuda:2'), covar=tensor([0.1534, 0.1407, 0.1826, 0.1233, 0.0858, 0.1581, 0.2293, 0.2130], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0323, 0.0354, 0.0297, 0.0336, 0.0324, 0.0307, 0.0354], + device='cuda:2'), out_proj_covar=tensor([6.4993e-05, 6.8698e-05, 7.6374e-05, 6.1399e-05, 7.0451e-05, 6.9591e-05, + 6.6128e-05, 7.5992e-05], device='cuda:2') +2023-04-26 22:46:28,630 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3150, 1.5858, 1.6808, 1.8019, 1.6586, 1.7344, 1.7740, 1.7447], + device='cuda:2'), covar=tensor([0.5112, 0.7665, 0.6581, 0.6054, 0.7400, 1.0472, 0.7397, 0.7139], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0384, 0.0317, 0.0326, 0.0343, 0.0405, 0.0364, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 22:46:30,380 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48868.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:46:31,004 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5984, 2.1315, 1.5529, 1.3119, 1.2162, 1.2210, 1.5494, 1.1694], + device='cuda:2'), covar=tensor([0.1839, 0.1300, 0.1649, 0.1962, 0.2513, 0.2093, 0.1109, 0.2190], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0216, 0.0171, 0.0204, 0.0204, 0.0183, 0.0161, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 22:46:32,696 INFO [finetune.py:976] (2/7) Epoch 9, batch 3050, loss[loss=0.191, simple_loss=0.2744, pruned_loss=0.0538, over 4902.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2641, pruned_loss=0.06521, over 954402.07 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:46:51,111 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48882.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:47:12,128 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48899.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:47:22,975 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0868, 1.5847, 2.0319, 2.3002, 1.9357, 1.5149, 1.0935, 1.6515], + device='cuda:2'), covar=tensor([0.3817, 0.3947, 0.1912, 0.3028, 0.2992, 0.3034, 0.5071, 0.2879], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0250, 0.0219, 0.0319, 0.0213, 0.0228, 0.0233, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 22:47:28,833 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48916.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 22:47:32,441 INFO [finetune.py:976] (2/7) Epoch 9, batch 3100, loss[loss=0.2172, simple_loss=0.276, pruned_loss=0.07919, over 4901.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2623, pruned_loss=0.06494, over 955302.47 frames. ], batch size: 32, lr: 3.79e-03, grad_scale: 16.0 +2023-04-26 22:47:42,242 INFO [optim.py:369] (2/7) 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 22:47:55,819 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 22:48:05,673 INFO [finetune.py:976] (2/7) Epoch 9, batch 3150, loss[loss=0.1642, simple_loss=0.2248, pruned_loss=0.05179, over 4741.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2601, pruned_loss=0.06475, over 955303.78 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:48:11,544 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48980.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:48:14,631 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 22:48:27,401 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-26 22:48:38,602 INFO [finetune.py:976] (2/7) Epoch 9, batch 3200, loss[loss=0.1669, simple_loss=0.2326, pruned_loss=0.05065, over 4779.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2552, pruned_loss=0.06254, over 954882.96 frames. ], batch size: 26, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:48:42,797 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:48:49,949 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-26 22:48:55,117 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8667, 2.1244, 1.1245, 1.5605, 2.4620, 1.7243, 1.6627, 1.8022], + device='cuda:2'), covar=tensor([0.0510, 0.0370, 0.0328, 0.0570, 0.0223, 0.0551, 0.0540, 0.0576], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0024, 0.0030, 0.0021, 0.0030, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 22:48:59,924 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49053.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:49:10,365 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-26 22:49:12,011 INFO [finetune.py:976] (2/7) Epoch 9, batch 3250, loss[loss=0.2204, simple_loss=0.2866, pruned_loss=0.07715, over 4901.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2555, pruned_loss=0.06281, over 953570.58 frames. ], batch size: 43, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:49:16,884 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49079.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:49:32,077 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49101.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:49:45,282 INFO [finetune.py:976] (2/7) Epoch 9, batch 3300, loss[loss=0.1973, simple_loss=0.264, pruned_loss=0.06525, over 4908.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2601, pruned_loss=0.06436, over 952943.76 frames. ], batch size: 37, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:49:48,898 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:50:18,718 INFO [finetune.py:976] (2/7) Epoch 9, batch 3350, loss[loss=0.2058, simple_loss=0.2796, pruned_loss=0.06595, over 4842.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2618, pruned_loss=0.06493, over 953353.85 frames. ], batch size: 49, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:50:21,873 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49177.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:50:46,871 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49194.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:50:59,412 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3758, 3.3944, 2.5265, 3.9872, 3.3988, 3.4321, 1.5928, 3.3853], + device='cuda:2'), covar=tensor([0.1938, 0.1257, 0.3855, 0.1948, 0.2315, 0.1951, 0.5278, 0.2382], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0251, 0.0306, 0.0302, 0.0252, 0.0271, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 22:51:20,617 INFO [finetune.py:976] (2/7) Epoch 9, batch 3400, loss[loss=0.1447, simple_loss=0.2087, pruned_loss=0.04032, over 4744.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2624, pruned_loss=0.06518, over 952848.30 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:51:38,491 INFO [optim.py:369] (2/7) 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 22:51:39,201 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49236.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:52:09,280 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.65 vs. limit=5.0 +2023-04-26 22:52:13,320 INFO [finetune.py:976] (2/7) Epoch 9, batch 3450, loss[loss=0.2387, simple_loss=0.2894, pruned_loss=0.09396, over 4724.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2613, pruned_loss=0.06392, over 952531.51 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:52:23,662 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6802, 2.4066, 1.7112, 1.7777, 1.2514, 1.2968, 1.7043, 1.2397], + device='cuda:2'), covar=tensor([0.1680, 0.1275, 0.1561, 0.1723, 0.2605, 0.2053, 0.1130, 0.2126], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0216, 0.0170, 0.0204, 0.0204, 0.0183, 0.0160, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 22:52:30,777 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49297.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:53:02,578 INFO [finetune.py:976] (2/7) Epoch 9, batch 3500, loss[loss=0.2001, simple_loss=0.2672, pruned_loss=0.06652, over 4763.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2578, pruned_loss=0.06241, over 954292.67 frames. ], batch size: 54, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:53:15,829 INFO [optim.py:369] (2/7) 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 22:53:37,248 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4703, 1.7051, 1.8032, 1.9778, 1.8249, 2.0252, 1.9338, 1.8302], + device='cuda:2'), covar=tensor([0.4483, 0.6191, 0.5211, 0.4389, 0.6068, 0.7949, 0.5854, 0.5786], + device='cuda:2'), in_proj_covar=tensor([0.0321, 0.0382, 0.0313, 0.0324, 0.0339, 0.0401, 0.0361, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 22:54:01,001 INFO [finetune.py:976] (2/7) Epoch 9, batch 3550, loss[loss=0.2021, simple_loss=0.2537, pruned_loss=0.07526, over 4822.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2562, pruned_loss=0.06251, over 955296.77 frames. ], batch size: 40, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:54:42,421 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.83 vs. limit=5.0 +2023-04-26 22:55:07,551 INFO [finetune.py:976] (2/7) Epoch 9, batch 3600, loss[loss=0.1918, simple_loss=0.2596, pruned_loss=0.06198, over 4753.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2546, pruned_loss=0.06231, over 954458.80 frames. ], batch size: 27, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:55:07,675 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49422.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:55:17,986 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 22:55:24,162 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49432.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:55:25,885 INFO [optim.py:369] (2/7) 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 22:55:28,456 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8931, 2.3327, 1.8546, 2.1665, 1.6077, 1.7808, 1.9562, 1.4168], + device='cuda:2'), covar=tensor([0.2122, 0.1347, 0.0946, 0.1278, 0.3385, 0.1355, 0.1972, 0.3028], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0317, 0.0228, 0.0289, 0.0314, 0.0271, 0.0258, 0.0282], + device='cuda:2'), out_proj_covar=tensor([1.2008e-04, 1.2774e-04, 9.1549e-05, 1.1590e-04, 1.2882e-04, 1.0944e-04, + 1.0557e-04, 1.1305e-04], device='cuda:2') +2023-04-26 22:55:38,419 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5035, 4.0538, 0.7123, 1.9254, 1.7610, 2.6600, 2.2191, 0.8509], + device='cuda:2'), covar=tensor([0.1974, 0.1778, 0.2904, 0.2076, 0.1675, 0.1467, 0.1975, 0.2647], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0248, 0.0140, 0.0122, 0.0135, 0.0152, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:55:50,736 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-26 22:56:09,752 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0442, 1.3266, 1.1410, 1.5747, 1.3598, 1.3989, 1.2384, 2.3905], + device='cuda:2'), covar=tensor([0.0664, 0.0837, 0.0905, 0.1238, 0.0715, 0.0559, 0.0805, 0.0240], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 22:56:18,974 INFO [finetune.py:976] (2/7) Epoch 9, batch 3650, loss[loss=0.2005, simple_loss=0.2628, pruned_loss=0.06915, over 4878.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2583, pruned_loss=0.06428, over 951725.91 frames. ], batch size: 31, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:56:22,164 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49477.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:56:30,064 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49483.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:56:36,698 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49493.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 22:56:37,260 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49494.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:56:44,470 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-26 22:56:56,124 INFO [finetune.py:976] (2/7) Epoch 9, batch 3700, loss[loss=0.1491, simple_loss=0.2314, pruned_loss=0.03336, over 4894.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2623, pruned_loss=0.06585, over 951367.21 frames. ], batch size: 32, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:56:58,031 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49525.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:57:09,404 INFO [optim.py:369] (2/7) 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 22:57:13,690 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49542.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:57:55,118 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-04-26 22:57:56,799 INFO [finetune.py:976] (2/7) Epoch 9, batch 3750, loss[loss=0.2126, simple_loss=0.2745, pruned_loss=0.07532, over 4920.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2641, pruned_loss=0.06659, over 951418.26 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:58:15,038 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-26 22:58:23,846 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49592.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 22:58:44,875 INFO [finetune.py:976] (2/7) Epoch 9, batch 3800, loss[loss=0.2035, simple_loss=0.2768, pruned_loss=0.06513, over 4887.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2648, pruned_loss=0.0663, over 951916.82 frames. ], batch size: 35, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:58:52,834 INFO [optim.py:369] (2/7) 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 22:59:17,906 INFO [finetune.py:976] (2/7) Epoch 9, batch 3850, loss[loss=0.2626, simple_loss=0.3014, pruned_loss=0.1119, over 4805.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2629, pruned_loss=0.06547, over 952565.94 frames. ], batch size: 51, lr: 3.78e-03, grad_scale: 16.0 +2023-04-26 22:59:43,353 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4620, 3.4074, 0.8898, 1.8864, 1.9209, 2.4754, 1.9990, 1.0022], + device='cuda:2'), covar=tensor([0.1398, 0.0931, 0.2016, 0.1302, 0.1123, 0.1013, 0.1511, 0.1919], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0248, 0.0140, 0.0121, 0.0135, 0.0152, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 22:59:49,729 INFO [finetune.py:976] (2/7) Epoch 9, batch 3900, loss[loss=0.2049, simple_loss=0.2697, pruned_loss=0.07006, over 4909.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2599, pruned_loss=0.06434, over 952845.01 frames. ], batch size: 35, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 22:59:55,899 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-26 22:59:56,985 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5893, 3.4822, 2.6040, 4.2350, 3.6886, 3.6884, 1.4503, 3.5917], + device='cuda:2'), covar=tensor([0.1957, 0.1442, 0.3387, 0.2222, 0.3020, 0.2271, 0.6883, 0.2890], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0218, 0.0252, 0.0307, 0.0303, 0.0253, 0.0273, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 22:59:58,108 INFO [optim.py:369] (2/7) 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 23:00:18,050 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 3950, loss[loss=0.1988, simple_loss=0.2576, pruned_loss=0.06998, over 4828.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2559, pruned_loss=0.0628, over 954579.18 frames. ], batch size: 30, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:00:27,205 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49778.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:00:33,308 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49788.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 23:00:55,287 INFO [finetune.py:976] (2/7) Epoch 9, batch 4000, loss[loss=0.1492, simple_loss=0.2242, pruned_loss=0.03704, over 4764.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2543, pruned_loss=0.0622, over 955742.43 frames. ], batch size: 26, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:00:59,439 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49827.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:01:05,129 INFO [optim.py:369] (2/7) 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 23:01:44,301 INFO [finetune.py:976] (2/7) Epoch 9, batch 4050, loss[loss=0.1549, simple_loss=0.2202, pruned_loss=0.04481, over 4777.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2574, pruned_loss=0.06336, over 953655.68 frames. ], batch size: 26, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:02:14,668 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49892.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:02:28,886 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-26 23:02:41,625 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.03 vs. limit=2.0 +2023-04-26 23:02:49,741 INFO [finetune.py:976] (2/7) Epoch 9, batch 4100, loss[loss=0.1525, simple_loss=0.2157, pruned_loss=0.04469, over 4725.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2614, pruned_loss=0.06454, over 952120.60 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:03:10,345 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49940.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:03:53,587 INFO [finetune.py:976] (2/7) Epoch 9, batch 4150, loss[loss=0.1542, simple_loss=0.2321, pruned_loss=0.03813, over 4769.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2631, pruned_loss=0.06517, over 952816.58 frames. ], batch size: 28, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:04:33,305 INFO [finetune.py:976] (2/7) Epoch 9, batch 4200, loss[loss=0.158, simple_loss=0.22, pruned_loss=0.04798, over 4092.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2625, pruned_loss=0.06447, over 953371.34 frames. ], batch size: 18, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:04:41,638 INFO [optim.py:369] (2/7) 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 23:04:50,503 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7124, 1.8607, 0.9035, 1.4698, 2.2347, 1.6250, 1.6208, 1.6143], + device='cuda:2'), covar=tensor([0.0521, 0.0357, 0.0328, 0.0526, 0.0244, 0.0494, 0.0467, 0.0566], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 23:05:05,680 INFO [finetune.py:976] (2/7) Epoch 9, batch 4250, loss[loss=0.1923, simple_loss=0.2515, pruned_loss=0.06657, over 4934.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2603, pruned_loss=0.06352, over 955125.44 frames. ], batch size: 38, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:05:09,943 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50078.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:16,029 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50088.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:05:21,927 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7815, 1.0901, 1.2700, 1.3646, 1.7737, 1.4993, 1.2623, 1.2259], + device='cuda:2'), covar=tensor([0.1522, 0.1965, 0.2324, 0.1596, 0.1092, 0.1913, 0.2304, 0.2437], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0321, 0.0351, 0.0296, 0.0333, 0.0320, 0.0306, 0.0355], + device='cuda:2'), out_proj_covar=tensor([6.4489e-05, 6.8186e-05, 7.5788e-05, 6.1124e-05, 6.9638e-05, 6.8722e-05, + 6.5709e-05, 7.6198e-05], device='cuda:2') +2023-04-26 23:05:37,583 INFO [finetune.py:976] (2/7) Epoch 9, batch 4300, loss[loss=0.1985, simple_loss=0.2571, pruned_loss=0.0699, over 4865.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2579, pruned_loss=0.0634, over 954163.67 frames. ], batch size: 49, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:05:37,653 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50122.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:40,097 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50126.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:41,355 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50128.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:05:41,985 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8316, 2.0830, 1.9446, 2.1162, 1.8839, 2.1049, 2.1177, 2.0116], + device='cuda:2'), covar=tensor([0.5058, 0.8000, 0.7081, 0.6144, 0.7721, 0.9929, 0.8228, 0.7135], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0381, 0.0314, 0.0325, 0.0339, 0.0401, 0.0361, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:05:46,442 INFO [optim.py:369] (2/7) 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 23:05:47,120 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50136.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:06:04,212 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2048, 2.9433, 2.5074, 2.5816, 2.2669, 2.4561, 2.5979, 2.0956], + device='cuda:2'), covar=tensor([0.2702, 0.2086, 0.1189, 0.1996, 0.3153, 0.1835, 0.2384, 0.2871], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0317, 0.0227, 0.0289, 0.0312, 0.0271, 0.0258, 0.0280], + device='cuda:2'), out_proj_covar=tensor([1.1974e-04, 1.2773e-04, 9.1471e-05, 1.1562e-04, 1.2794e-04, 1.0918e-04, + 1.0556e-04, 1.1256e-04], device='cuda:2') +2023-04-26 23:06:07,915 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4887, 1.8272, 1.8954, 2.0422, 1.9195, 2.0487, 1.9797, 1.9652], + device='cuda:2'), covar=tensor([0.5148, 0.7360, 0.6163, 0.5986, 0.6685, 0.9111, 0.6696, 0.6153], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0381, 0.0313, 0.0325, 0.0339, 0.0401, 0.0361, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:06:10,816 INFO [finetune.py:976] (2/7) Epoch 9, batch 4350, loss[loss=0.2028, simple_loss=0.2613, pruned_loss=0.07215, over 4884.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2562, pruned_loss=0.06334, over 953286.17 frames. ], batch size: 32, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:06:15,251 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-04-26 23:06:22,150 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50189.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:06:27,931 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50197.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:06:44,562 INFO [finetune.py:976] (2/7) Epoch 9, batch 4400, loss[loss=0.2268, simple_loss=0.286, pruned_loss=0.08383, over 4826.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.259, pruned_loss=0.06489, over 954733.04 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-04-26 23:06:52,519 INFO [optim.py:369] (2/7) 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 23:07:13,997 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50258.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:07:27,674 INFO [finetune.py:976] (2/7) Epoch 9, batch 4450, loss[loss=0.1557, simple_loss=0.2185, pruned_loss=0.04645, over 3910.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2614, pruned_loss=0.06532, over 956026.09 frames. ], batch size: 17, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:07:35,661 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 23:08:33,136 INFO [finetune.py:976] (2/7) Epoch 9, batch 4500, loss[loss=0.1731, simple_loss=0.2548, pruned_loss=0.04563, over 4892.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2624, pruned_loss=0.06556, over 954618.36 frames. ], batch size: 35, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:08:46,790 INFO [optim.py:369] (2/7) 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 23:08:49,547 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-26 23:08:55,928 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2175, 1.4851, 1.6597, 1.7822, 1.6830, 1.7855, 1.7883, 1.7034], + device='cuda:2'), covar=tensor([0.5477, 0.7377, 0.6064, 0.5604, 0.6755, 0.9717, 0.6624, 0.6109], + device='cuda:2'), in_proj_covar=tensor([0.0322, 0.0381, 0.0314, 0.0324, 0.0338, 0.0400, 0.0360, 0.0322], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:09:12,331 INFO [finetune.py:976] (2/7) Epoch 9, batch 4550, loss[loss=0.1905, simple_loss=0.2645, pruned_loss=0.05827, over 4816.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2643, pruned_loss=0.06656, over 955241.25 frames. ], batch size: 30, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:09:56,481 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4241, 1.0507, 0.3077, 1.1466, 1.1560, 1.2787, 1.1982, 1.2246], + device='cuda:2'), covar=tensor([0.0568, 0.0463, 0.0502, 0.0647, 0.0335, 0.0619, 0.0597, 0.0661], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0026, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 23:10:20,888 INFO [finetune.py:976] (2/7) Epoch 9, batch 4600, loss[loss=0.157, simple_loss=0.2352, pruned_loss=0.03943, over 4900.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.263, pruned_loss=0.06536, over 955000.88 frames. ], batch size: 37, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:10:20,979 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50422.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:10:39,535 INFO [optim.py:369] (2/7) 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 23:11:09,109 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2049, 1.6312, 2.0702, 2.6484, 2.0398, 1.6174, 1.2904, 1.8645], + device='cuda:2'), covar=tensor([0.3775, 0.3746, 0.1815, 0.2538, 0.3543, 0.3069, 0.5008, 0.2735], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0251, 0.0220, 0.0318, 0.0214, 0.0229, 0.0233, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 23:11:14,981 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 4650, loss[loss=0.2115, simple_loss=0.2661, pruned_loss=0.07845, over 4729.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2597, pruned_loss=0.0643, over 954510.52 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:11:23,417 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50484.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:11:46,336 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2599, 1.5937, 1.4598, 1.8097, 1.7093, 1.9702, 1.4263, 3.6628], + device='cuda:2'), covar=tensor([0.0630, 0.0753, 0.0777, 0.1169, 0.0642, 0.0545, 0.0725, 0.0133], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 23:11:49,303 INFO [finetune.py:976] (2/7) Epoch 9, batch 4700, loss[loss=0.163, simple_loss=0.2237, pruned_loss=0.05114, over 4725.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2557, pruned_loss=0.06294, over 955575.48 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:11:57,183 INFO [optim.py:369] (2/7) 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 23:12:04,107 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50546.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:12:08,325 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50553.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 23:12:22,117 INFO [finetune.py:976] (2/7) Epoch 9, batch 4750, loss[loss=0.1961, simple_loss=0.2632, pruned_loss=0.06449, over 4935.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2523, pruned_loss=0.06141, over 954811.82 frames. ], batch size: 33, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:12:22,238 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8676, 2.2912, 2.0098, 2.1444, 1.6693, 1.9143, 1.9799, 1.4924], + device='cuda:2'), covar=tensor([0.2267, 0.1188, 0.0759, 0.1326, 0.3350, 0.1090, 0.1873, 0.2664], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0317, 0.0227, 0.0288, 0.0315, 0.0271, 0.0258, 0.0280], + device='cuda:2'), out_proj_covar=tensor([1.2002e-04, 1.2773e-04, 9.1356e-05, 1.1558e-04, 1.2895e-04, 1.0932e-04, + 1.0527e-04, 1.1227e-04], device='cuda:2') +2023-04-26 23:12:45,648 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-26 23:12:54,832 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50607.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:13:16,244 INFO [finetune.py:976] (2/7) Epoch 9, batch 4800, loss[loss=0.1704, simple_loss=0.2546, pruned_loss=0.04303, over 4891.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.256, pruned_loss=0.06274, over 952812.30 frames. ], batch size: 46, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:13:30,154 INFO [optim.py:369] (2/7) 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 23:13:44,327 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-26 23:13:55,576 INFO [finetune.py:976] (2/7) Epoch 9, batch 4850, loss[loss=0.1585, simple_loss=0.2237, pruned_loss=0.04667, over 4691.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2593, pruned_loss=0.064, over 950928.59 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:14:16,528 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-26 23:14:28,052 INFO [finetune.py:976] (2/7) Epoch 9, batch 4900, loss[loss=0.207, simple_loss=0.2738, pruned_loss=0.07008, over 4835.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2623, pruned_loss=0.06541, over 951750.06 frames. ], batch size: 30, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:14:36,898 INFO [optim.py:369] (2/7) 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 23:15:14,089 INFO [finetune.py:976] (2/7) Epoch 9, batch 4950, loss[loss=0.1926, simple_loss=0.2601, pruned_loss=0.06254, over 4859.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2636, pruned_loss=0.06568, over 952447.81 frames. ], batch size: 34, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:15:33,286 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50782.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:15:34,461 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50784.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:15:42,665 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1446, 2.7668, 2.2668, 2.4039, 2.0146, 2.2632, 2.5458, 1.6439], + device='cuda:2'), covar=tensor([0.2473, 0.1241, 0.0881, 0.1643, 0.3478, 0.1350, 0.2120, 0.3270], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0316, 0.0227, 0.0289, 0.0316, 0.0272, 0.0258, 0.0280], + device='cuda:2'), out_proj_covar=tensor([1.2012e-04, 1.2743e-04, 9.1247e-05, 1.1574e-04, 1.2932e-04, 1.0959e-04, + 1.0535e-04, 1.1216e-04], device='cuda:2') +2023-04-26 23:15:44,633 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-26 23:16:02,477 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50808.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:16:11,330 INFO [finetune.py:976] (2/7) Epoch 9, batch 5000, loss[loss=0.2215, simple_loss=0.2524, pruned_loss=0.09527, over 4367.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2604, pruned_loss=0.06483, over 952531.00 frames. ], batch size: 19, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:16:19,880 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 23:16:26,683 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50843.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:16:32,730 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50853.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:16:42,405 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50869.0, num_to_drop=1, layers_to_drop={3} +2023-04-26 23:16:44,512 INFO [finetune.py:976] (2/7) Epoch 9, batch 5050, loss[loss=0.1573, simple_loss=0.221, pruned_loss=0.04686, over 4315.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2575, pruned_loss=0.06391, over 951756.10 frames. ], batch size: 65, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:17:04,639 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50902.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:17:17,316 INFO [finetune.py:976] (2/7) Epoch 9, batch 5100, loss[loss=0.1899, simple_loss=0.2548, pruned_loss=0.06255, over 4856.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2541, pruned_loss=0.0626, over 951368.54 frames. ], batch size: 31, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:17:22,775 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3836, 1.2740, 1.6150, 1.5392, 1.2862, 1.1305, 1.2451, 0.7633], + device='cuda:2'), covar=tensor([0.0599, 0.0716, 0.0474, 0.0564, 0.0743, 0.1192, 0.0665, 0.0818], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0072, 0.0071, 0.0067, 0.0076, 0.0096, 0.0078, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 23:17:26,166 INFO [optim.py:369] (2/7) 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 23:17:34,469 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50946.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:17:50,269 INFO [finetune.py:976] (2/7) Epoch 9, batch 5150, loss[loss=0.2127, simple_loss=0.2801, pruned_loss=0.07265, over 4790.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2553, pruned_loss=0.06357, over 949044.52 frames. ], batch size: 41, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:17:58,869 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.69 vs. limit=2.0 +2023-04-26 23:18:25,469 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51007.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:18:40,597 INFO [finetune.py:976] (2/7) Epoch 9, batch 5200, loss[loss=0.1974, simple_loss=0.2735, pruned_loss=0.06069, over 4829.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2598, pruned_loss=0.0647, over 948976.18 frames. ], batch size: 33, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:18:49,051 INFO [optim.py:369] (2/7) 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 23:19:08,577 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 5250, loss[loss=0.2057, simple_loss=0.2733, pruned_loss=0.0691, over 4829.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2627, pruned_loss=0.06534, over 952059.69 frames. ], batch size: 30, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:19:21,892 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.41 vs. limit=5.0 +2023-04-26 23:19:47,750 INFO [finetune.py:976] (2/7) Epoch 9, batch 5300, loss[loss=0.2192, simple_loss=0.2978, pruned_loss=0.07025, over 4894.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.2642, pruned_loss=0.06638, over 949868.00 frames. ], batch size: 43, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:19:48,482 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51123.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:19:56,070 INFO [optim.py:369] (2/7) 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 23:19:57,977 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51138.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:20:00,454 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0603, 2.6301, 2.1650, 2.4334, 1.6530, 2.1261, 2.2144, 1.7149], + device='cuda:2'), covar=tensor([0.2147, 0.1024, 0.0699, 0.1034, 0.3395, 0.1207, 0.1793, 0.2651], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0316, 0.0226, 0.0288, 0.0316, 0.0272, 0.0257, 0.0278], + device='cuda:2'), out_proj_covar=tensor([1.1970e-04, 1.2744e-04, 9.0832e-05, 1.1566e-04, 1.2936e-04, 1.0961e-04, + 1.0509e-04, 1.1170e-04], device='cuda:2') +2023-04-26 23:20:15,867 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51164.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:20:20,663 INFO [finetune.py:976] (2/7) Epoch 9, batch 5350, loss[loss=0.151, simple_loss=0.2245, pruned_loss=0.03871, over 4774.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2644, pruned_loss=0.06586, over 952605.01 frames. ], batch size: 29, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:20:20,768 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51172.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:20:23,248 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-26 23:20:25,740 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-26 23:20:46,712 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51202.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:21:10,256 INFO [finetune.py:976] (2/7) Epoch 9, batch 5400, loss[loss=0.221, simple_loss=0.2622, pruned_loss=0.08992, over 4288.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2601, pruned_loss=0.06408, over 952358.55 frames. ], batch size: 65, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:21:21,602 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51233.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:21:22,681 INFO [optim.py:369] (2/7) 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 23:21:29,801 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9667, 2.3292, 2.0507, 2.2671, 1.8206, 1.9667, 2.0115, 1.6457], + device='cuda:2'), covar=tensor([0.1784, 0.1232, 0.0850, 0.1036, 0.2614, 0.1178, 0.1765, 0.2328], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0314, 0.0225, 0.0286, 0.0313, 0.0271, 0.0256, 0.0278], + device='cuda:2'), out_proj_covar=tensor([1.1894e-04, 1.2675e-04, 9.0331e-05, 1.1493e-04, 1.2834e-04, 1.0906e-04, + 1.0469e-04, 1.1154e-04], device='cuda:2') +2023-04-26 23:21:43,079 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1148, 2.5079, 1.0226, 1.5549, 1.5298, 1.9250, 1.6222, 0.8725], + device='cuda:2'), covar=tensor([0.1565, 0.1046, 0.1667, 0.1273, 0.1171, 0.0876, 0.1532, 0.1635], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0252, 0.0142, 0.0123, 0.0137, 0.0156, 0.0119, 0.0124], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 23:21:44,190 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51250.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:22:06,405 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.05 vs. limit=2.0 +2023-04-26 23:22:14,586 INFO [finetune.py:976] (2/7) Epoch 9, batch 5450, loss[loss=0.1857, simple_loss=0.248, pruned_loss=0.0617, over 4908.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2562, pruned_loss=0.06237, over 952552.98 frames. ], batch size: 36, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:22:47,579 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51302.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:23:19,200 INFO [finetune.py:976] (2/7) Epoch 9, batch 5500, loss[loss=0.1496, simple_loss=0.2241, pruned_loss=0.0375, over 4913.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2529, pruned_loss=0.06122, over 952967.37 frames. ], batch size: 36, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:23:32,910 INFO [optim.py:369] (2/7) 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 23:24:09,341 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6069, 3.5087, 2.6327, 4.1922, 3.6500, 3.4980, 1.7257, 3.5392], + device='cuda:2'), covar=tensor([0.1704, 0.1344, 0.3712, 0.1696, 0.3300, 0.2055, 0.5528, 0.2851], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0215, 0.0248, 0.0300, 0.0298, 0.0249, 0.0268, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:24:14,746 INFO [finetune.py:976] (2/7) Epoch 9, batch 5550, loss[loss=0.1272, simple_loss=0.2018, pruned_loss=0.0263, over 4786.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2545, pruned_loss=0.06192, over 951466.63 frames. ], batch size: 26, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:24:20,977 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51382.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:24:26,985 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51392.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:24:42,940 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 9, batch 5600, loss[loss=0.1914, simple_loss=0.2627, pruned_loss=0.06007, over 4736.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2607, pruned_loss=0.06454, over 952865.03 frames. ], batch size: 27, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:24:52,678 INFO [optim.py:369] (2/7) 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 23:24:54,552 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51438.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:24:57,452 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51443.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:03,229 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51453.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:10,021 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51464.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:25:15,160 INFO [finetune.py:976] (2/7) Epoch 9, batch 5650, loss[loss=0.1864, simple_loss=0.2658, pruned_loss=0.0535, over 4817.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2643, pruned_loss=0.06565, over 953120.01 frames. ], batch size: 40, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:25:23,729 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51486.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:29,051 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51495.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:25:37,342 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6171, 4.6074, 1.2893, 2.8597, 3.0245, 3.3458, 2.9310, 1.3150], + device='cuda:2'), covar=tensor([0.0968, 0.0988, 0.1845, 0.1027, 0.0816, 0.0870, 0.1178, 0.1899], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0250, 0.0141, 0.0122, 0.0136, 0.0154, 0.0119, 0.0123], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 23:25:39,094 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51512.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:25:45,276 INFO [finetune.py:976] (2/7) Epoch 9, batch 5700, loss[loss=0.1458, simple_loss=0.2075, pruned_loss=0.04205, over 4004.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2601, pruned_loss=0.06517, over 933952.55 frames. ], batch size: 17, lr: 3.77e-03, grad_scale: 32.0 +2023-04-26 23:25:48,917 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51528.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:25:53,110 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 0, loss[loss=0.2078, simple_loss=0.2731, pruned_loss=0.07129, over 4812.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2731, pruned_loss=0.07129, over 4812.00 frames. ], batch size: 39, lr: 3.76e-03, grad_scale: 32.0 +2023-04-26 23:26:16,105 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-26 23:26:31,820 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6355MB +2023-04-26 23:26:37,659 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51556.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:27:05,706 INFO [finetune.py:976] (2/7) Epoch 10, batch 50, loss[loss=0.1927, simple_loss=0.262, pruned_loss=0.06172, over 4896.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2629, pruned_loss=0.06437, over 216337.14 frames. ], batch size: 37, lr: 3.76e-03, grad_scale: 32.0 +2023-04-26 23:27:05,831 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0390, 0.9830, 1.2687, 1.1706, 1.0069, 0.9056, 0.8636, 0.5169], + device='cuda:2'), covar=tensor([0.0607, 0.0737, 0.0498, 0.0602, 0.0783, 0.1411, 0.0533, 0.0905], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0070, 0.0067, 0.0076, 0.0095, 0.0078, 0.0074], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 23:27:08,565 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51602.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:27:13,943 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9416, 1.5894, 1.8401, 2.1583, 2.2113, 1.6988, 1.4373, 1.8527], + device='cuda:2'), covar=tensor([0.0811, 0.1163, 0.0713, 0.0606, 0.0600, 0.0899, 0.0892, 0.0610], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0203, 0.0181, 0.0174, 0.0176, 0.0186, 0.0159, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:27:31,938 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 100, loss[loss=0.2033, simple_loss=0.2591, pruned_loss=0.07377, over 4783.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2568, pruned_loss=0.06233, over 380563.45 frames. ], batch size: 29, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:27:47,278 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51650.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:28:38,563 INFO [finetune.py:976] (2/7) Epoch 10, batch 150, loss[loss=0.1748, simple_loss=0.2419, pruned_loss=0.0538, over 4813.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2524, pruned_loss=0.061, over 508398.88 frames. ], batch size: 39, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:29:04,454 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51718.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:20,969 INFO [optim.py:369] (2/7) 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 23:29:22,229 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5267, 1.1461, 4.3353, 4.0524, 3.7223, 3.9662, 3.9317, 3.8098], + device='cuda:2'), covar=tensor([0.7047, 0.6286, 0.0908, 0.1685, 0.1183, 0.1916, 0.1966, 0.1521], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0307, 0.0406, 0.0410, 0.0349, 0.0405, 0.0315, 0.0371], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:29:22,231 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51738.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:29,222 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51747.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:29,797 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51748.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:30,344 INFO [finetune.py:976] (2/7) Epoch 10, batch 200, loss[loss=0.1665, simple_loss=0.2281, pruned_loss=0.05248, over 4788.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2511, pruned_loss=0.06095, over 608912.89 frames. ], batch size: 29, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:29:42,681 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51766.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:29:44,621 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5709, 1.1454, 1.6469, 1.9943, 1.7548, 1.5811, 1.6111, 1.5910], + device='cuda:2'), covar=tensor([0.4684, 0.6254, 0.5770, 0.6655, 0.5635, 0.7561, 0.7333, 0.7358], + device='cuda:2'), in_proj_covar=tensor([0.0405, 0.0414, 0.0498, 0.0517, 0.0435, 0.0452, 0.0464, 0.0462], + device='cuda:2'), out_proj_covar=tensor([9.8725e-05, 1.0249e-04, 1.1241e-04, 1.2294e-04, 1.0552e-04, 1.0937e-04, + 1.1148e-04, 1.1147e-04], device='cuda:2') +2023-04-26 23:30:04,062 INFO [finetune.py:976] (2/7) Epoch 10, batch 250, loss[loss=0.1987, simple_loss=0.2607, pruned_loss=0.06835, over 4827.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2576, pruned_loss=0.06367, over 683497.74 frames. ], batch size: 30, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:30:04,215 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5935, 1.8104, 1.8052, 1.9572, 1.8427, 2.0208, 1.9465, 1.8853], + device='cuda:2'), covar=tensor([0.5127, 0.7390, 0.6283, 0.5766, 0.7320, 1.0130, 0.7165, 0.6851], + device='cuda:2'), in_proj_covar=tensor([0.0326, 0.0384, 0.0317, 0.0327, 0.0340, 0.0403, 0.0361, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:30:11,190 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51808.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:30:16,538 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4313, 1.3262, 1.7358, 1.6317, 1.3205, 1.1675, 1.3748, 0.8207], + device='cuda:2'), covar=tensor([0.0716, 0.0727, 0.0490, 0.0797, 0.0818, 0.1357, 0.0763, 0.0863], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0073, 0.0071, 0.0067, 0.0076, 0.0095, 0.0078, 0.0074], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 23:30:23,241 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9108, 1.1714, 5.2683, 4.9198, 4.5026, 4.9957, 4.5900, 4.6404], + device='cuda:2'), covar=tensor([0.7180, 0.7123, 0.0921, 0.1740, 0.1116, 0.1540, 0.1122, 0.1371], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0308, 0.0408, 0.0411, 0.0351, 0.0407, 0.0317, 0.0372], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:30:23,894 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51828.0, num_to_drop=1, layers_to_drop={1} +2023-04-26 23:30:26,329 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5832, 1.3956, 4.3249, 4.0092, 3.7657, 4.0560, 4.0247, 3.8098], + device='cuda:2'), covar=tensor([0.6860, 0.6106, 0.1042, 0.1873, 0.1165, 0.1708, 0.1415, 0.1543], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0308, 0.0408, 0.0411, 0.0351, 0.0407, 0.0317, 0.0372], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:30:28,669 INFO [optim.py:369] (2/7) 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 23:30:30,639 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8655, 2.6073, 1.8085, 1.8721, 1.3827, 1.3377, 1.9777, 1.3090], + device='cuda:2'), covar=tensor([0.1756, 0.1687, 0.1678, 0.2091, 0.2711, 0.2220, 0.1184, 0.2254], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0217, 0.0172, 0.0206, 0.0206, 0.0185, 0.0161, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 23:30:37,590 INFO [finetune.py:976] (2/7) Epoch 10, batch 300, loss[loss=0.2054, simple_loss=0.2757, pruned_loss=0.06759, over 4764.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2604, pruned_loss=0.06401, over 744030.61 frames. ], batch size: 29, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:30:38,886 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51851.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:30:56,475 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51876.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:31:01,947 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6272, 1.5041, 0.8320, 1.2554, 1.8647, 1.5023, 1.3822, 1.4256], + device='cuda:2'), covar=tensor([0.0527, 0.0413, 0.0387, 0.0608, 0.0295, 0.0559, 0.0525, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 23:31:10,831 INFO [finetune.py:976] (2/7) Epoch 10, batch 350, loss[loss=0.231, simple_loss=0.2911, pruned_loss=0.08543, over 4898.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.263, pruned_loss=0.06527, over 791709.60 frames. ], batch size: 36, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:31:21,527 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1844, 1.5193, 1.3687, 1.7983, 1.6693, 1.8114, 1.4364, 3.1220], + device='cuda:2'), covar=tensor([0.0664, 0.0821, 0.0855, 0.1224, 0.0657, 0.0465, 0.0778, 0.0182], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 23:31:41,424 INFO [optim.py:369] (2/7) 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 23:32:00,717 INFO [finetune.py:976] (2/7) Epoch 10, batch 400, loss[loss=0.1912, simple_loss=0.2626, pruned_loss=0.05994, over 4926.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2624, pruned_loss=0.06426, over 828721.50 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:32:03,931 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-26 23:32:51,604 INFO [finetune.py:976] (2/7) Epoch 10, batch 450, loss[loss=0.2062, simple_loss=0.2694, pruned_loss=0.07148, over 4919.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2615, pruned_loss=0.06394, over 857448.08 frames. ], batch size: 37, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:33:45,360 INFO [optim.py:369] (2/7) 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 23:33:46,695 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52040.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:33:56,493 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.51 vs. limit=5.0 +2023-04-26 23:33:58,081 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 500, loss[loss=0.1665, simple_loss=0.2355, pruned_loss=0.04877, over 4762.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2584, pruned_loss=0.06321, over 880574.29 frames. ], batch size: 27, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:34:51,735 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52086.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:03,597 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52096.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:09,986 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6090, 1.4344, 4.4642, 4.2132, 3.9111, 4.2866, 4.1340, 4.0066], + device='cuda:2'), covar=tensor([0.6976, 0.6127, 0.0979, 0.1521, 0.0984, 0.1679, 0.1067, 0.1204], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0307, 0.0406, 0.0410, 0.0349, 0.0404, 0.0314, 0.0370], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:35:11,182 INFO [finetune.py:976] (2/7) Epoch 10, batch 550, loss[loss=0.1897, simple_loss=0.2595, pruned_loss=0.05995, over 4852.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2558, pruned_loss=0.06286, over 898861.17 frames. ], batch size: 49, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:35:12,520 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52101.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:13,633 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52103.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:35:22,464 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7885, 3.5170, 2.8082, 3.1093, 2.3904, 2.3690, 2.9364, 2.3465], + device='cuda:2'), covar=tensor([0.1469, 0.1315, 0.1328, 0.1319, 0.1964, 0.1861, 0.0801, 0.1694], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0216, 0.0170, 0.0205, 0.0205, 0.0184, 0.0161, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 23:35:32,446 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0518, 3.8124, 2.8266, 4.5948, 4.0219, 3.9550, 1.7279, 3.9801], + device='cuda:2'), covar=tensor([0.1542, 0.1379, 0.3289, 0.1422, 0.2930, 0.1957, 0.5861, 0.2265], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0216, 0.0250, 0.0304, 0.0301, 0.0251, 0.0270, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:35:34,835 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6230, 1.7138, 1.5803, 1.3383, 1.8325, 1.5200, 2.2020, 1.3595], + device='cuda:2'), covar=tensor([0.4128, 0.1865, 0.5106, 0.2998, 0.1664, 0.2514, 0.1623, 0.5263], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0350, 0.0428, 0.0363, 0.0390, 0.0385, 0.0381, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:35:58,006 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5726, 1.7959, 1.3667, 1.0413, 1.1997, 1.1614, 1.3042, 1.0955], + device='cuda:2'), covar=tensor([0.1695, 0.1384, 0.1621, 0.1875, 0.2535, 0.2055, 0.1198, 0.2160], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0216, 0.0170, 0.0205, 0.0205, 0.0184, 0.0161, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-26 23:36:05,085 INFO [optim.py:369] (2/7) 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 23:36:18,684 INFO [finetune.py:976] (2/7) Epoch 10, batch 600, loss[loss=0.2427, simple_loss=0.308, pruned_loss=0.08875, over 4860.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.256, pruned_loss=0.06325, over 912255.05 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:36:19,980 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52151.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:36:49,119 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-26 23:37:16,284 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52193.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:37:25,226 INFO [finetune.py:976] (2/7) Epoch 10, batch 650, loss[loss=0.202, simple_loss=0.2839, pruned_loss=0.06003, over 4813.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2599, pruned_loss=0.06451, over 921162.18 frames. ], batch size: 45, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:37:25,287 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52199.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:38:17,625 INFO [optim.py:369] (2/7) 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 23:38:18,948 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2475, 1.9417, 1.7567, 2.1302, 2.0646, 2.0573, 1.7396, 4.5468], + device='cuda:2'), covar=tensor([0.0609, 0.0724, 0.0725, 0.1130, 0.0602, 0.0541, 0.0730, 0.0098], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 23:38:30,841 INFO [finetune.py:976] (2/7) Epoch 10, batch 700, loss[loss=0.2582, simple_loss=0.3111, pruned_loss=0.1027, over 4231.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2613, pruned_loss=0.06481, over 929938.79 frames. ], batch size: 65, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:38:38,110 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-26 23:38:39,793 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52254.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:39:43,970 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0500, 1.5597, 1.4030, 1.9120, 2.1514, 1.7977, 1.7905, 1.5160], + device='cuda:2'), covar=tensor([0.2003, 0.1730, 0.2038, 0.1588, 0.1382, 0.1966, 0.2183, 0.1907], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0318, 0.0351, 0.0294, 0.0332, 0.0317, 0.0302, 0.0353], + device='cuda:2'), out_proj_covar=tensor([6.4408e-05, 6.7312e-05, 7.5737e-05, 6.0495e-05, 6.9582e-05, 6.8011e-05, + 6.4832e-05, 7.5754e-05], device='cuda:2') +2023-04-26 23:39:44,474 INFO [finetune.py:976] (2/7) Epoch 10, batch 750, loss[loss=0.1751, simple_loss=0.2443, pruned_loss=0.05292, over 4736.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2628, pruned_loss=0.06514, over 935640.25 frames. ], batch size: 59, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:39:56,798 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52312.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:40:07,446 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-26 23:40:11,924 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-04-26 23:40:13,561 INFO [optim.py:369] (2/7) 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 23:40:22,481 INFO [finetune.py:976] (2/7) Epoch 10, batch 800, loss[loss=0.1794, simple_loss=0.2428, pruned_loss=0.05804, over 4815.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2621, pruned_loss=0.06426, over 938918.88 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:40:30,280 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-26 23:40:37,968 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52373.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:40:46,709 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6880, 1.2989, 1.7506, 2.1967, 1.8576, 1.6233, 1.6718, 1.7275], + device='cuda:2'), covar=tensor([0.5698, 0.8140, 0.7961, 0.7528, 0.6795, 0.9971, 0.9790, 0.8618], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0419, 0.0502, 0.0521, 0.0439, 0.0460, 0.0469, 0.0469], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:40:50,104 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3004, 1.1267, 3.8479, 3.6076, 3.4485, 3.6754, 3.7015, 3.4506], + device='cuda:2'), covar=tensor([0.7025, 0.6262, 0.1245, 0.1888, 0.1195, 0.1739, 0.1455, 0.1467], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0308, 0.0407, 0.0410, 0.0350, 0.0406, 0.0315, 0.0372], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:40:54,333 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52396.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:40:55,822 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-26 23:40:56,135 INFO [finetune.py:976] (2/7) Epoch 10, batch 850, loss[loss=0.1377, simple_loss=0.2074, pruned_loss=0.03399, over 4795.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2602, pruned_loss=0.0636, over 943476.02 frames. ], batch size: 51, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:40:58,773 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52403.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:41:30,883 INFO [optim.py:369] (2/7) 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 23:41:44,733 INFO [finetune.py:976] (2/7) Epoch 10, batch 900, loss[loss=0.2001, simple_loss=0.2589, pruned_loss=0.07062, over 4285.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2567, pruned_loss=0.06262, over 944384.69 frames. ], batch size: 65, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:41:51,262 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52451.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:42:46,851 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3031, 1.7462, 1.4850, 2.0217, 1.8394, 2.1938, 1.5729, 4.3274], + device='cuda:2'), covar=tensor([0.0600, 0.0750, 0.0801, 0.1193, 0.0611, 0.0513, 0.0771, 0.0119], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 23:42:49,241 INFO [finetune.py:976] (2/7) Epoch 10, batch 950, loss[loss=0.2358, simple_loss=0.2925, pruned_loss=0.08955, over 4869.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2553, pruned_loss=0.06295, over 946640.71 frames. ], batch size: 34, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:43:41,297 INFO [optim.py:369] (2/7) 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 23:43:55,593 INFO [finetune.py:976] (2/7) Epoch 10, batch 1000, loss[loss=0.2237, simple_loss=0.293, pruned_loss=0.07717, over 4150.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2557, pruned_loss=0.06253, over 947415.66 frames. ], batch size: 65, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:44:01,403 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52549.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:44:05,061 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 1050, loss[loss=0.1876, simple_loss=0.2649, pruned_loss=0.05522, over 4821.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2582, pruned_loss=0.06265, over 948335.42 frames. ], batch size: 39, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:45:29,140 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52616.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:45:52,966 INFO [optim.py:369] (2/7) 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 23:45:54,270 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-26 23:46:13,537 INFO [finetune.py:976] (2/7) Epoch 10, batch 1100, loss[loss=0.1855, simple_loss=0.2586, pruned_loss=0.05621, over 4850.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2588, pruned_loss=0.06226, over 950031.41 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 16.0 +2023-04-26 23:46:15,952 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5035, 1.3656, 4.5012, 4.2295, 3.9948, 4.2397, 4.1547, 3.9273], + device='cuda:2'), covar=tensor([0.6974, 0.5759, 0.0975, 0.1562, 0.1016, 0.1367, 0.1190, 0.1632], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0306, 0.0406, 0.0410, 0.0348, 0.0404, 0.0314, 0.0371], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:46:36,536 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52668.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:46:39,180 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-26 23:46:50,589 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8699, 1.5727, 2.0516, 2.3323, 1.9875, 1.7966, 1.8885, 1.9281], + device='cuda:2'), covar=tensor([0.6933, 0.9265, 1.0403, 0.8909, 0.8121, 1.2532, 1.2668, 1.0730], + device='cuda:2'), in_proj_covar=tensor([0.0412, 0.0417, 0.0502, 0.0520, 0.0439, 0.0459, 0.0469, 0.0468], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:46:56,148 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52696.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:46:57,843 INFO [finetune.py:976] (2/7) Epoch 10, batch 1150, loss[loss=0.1547, simple_loss=0.2297, pruned_loss=0.03981, over 4920.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2595, pruned_loss=0.06201, over 951381.22 frames. ], batch size: 38, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:47:21,724 INFO [optim.py:369] (2/7) 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 23:47:33,672 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52744.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:47:43,258 INFO [finetune.py:976] (2/7) Epoch 10, batch 1200, loss[loss=0.2007, simple_loss=0.2672, pruned_loss=0.06708, over 4814.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2589, pruned_loss=0.06226, over 953551.87 frames. ], batch size: 41, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:47:45,183 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52752.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:48:48,806 INFO [finetune.py:976] (2/7) Epoch 10, batch 1250, loss[loss=0.1572, simple_loss=0.2238, pruned_loss=0.0453, over 4859.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2565, pruned_loss=0.06144, over 954038.12 frames. ], batch size: 31, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:49:09,956 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52813.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:49:34,630 INFO [optim.py:369] (2/7) 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 23:49:48,761 INFO [finetune.py:976] (2/7) Epoch 10, batch 1300, loss[loss=0.148, simple_loss=0.2178, pruned_loss=0.03912, over 4915.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2529, pruned_loss=0.06017, over 953121.79 frames. ], batch size: 36, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:49:48,847 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52849.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:50:11,025 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8249, 2.5539, 2.8036, 3.4507, 3.0499, 2.7316, 2.2489, 2.9463], + device='cuda:2'), covar=tensor([0.0875, 0.0939, 0.0600, 0.0535, 0.0590, 0.0867, 0.0819, 0.0609], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0205, 0.0184, 0.0177, 0.0179, 0.0190, 0.0161, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:50:20,543 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 1350, loss[loss=0.2102, simple_loss=0.2837, pruned_loss=0.06834, over 4914.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2541, pruned_loss=0.06089, over 956081.77 frames. ], batch size: 36, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:50:25,902 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2388, 1.4492, 1.2967, 1.6382, 1.5198, 1.8335, 1.2961, 3.1012], + device='cuda:2'), covar=tensor([0.0675, 0.0824, 0.0918, 0.1253, 0.0709, 0.0436, 0.0780, 0.0205], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 23:50:31,135 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52911.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:50:46,658 INFO [optim.py:369] (2/7) 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 23:50:55,066 INFO [finetune.py:976] (2/7) Epoch 10, batch 1400, loss[loss=0.1895, simple_loss=0.2647, pruned_loss=0.05718, over 4865.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2581, pruned_loss=0.06251, over 955060.08 frames. ], batch size: 31, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:51:09,512 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52968.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:51:10,284 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-26 23:51:25,790 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9185, 3.7459, 2.8938, 4.4707, 3.9440, 3.9020, 1.9193, 3.8512], + device='cuda:2'), covar=tensor([0.1875, 0.1272, 0.2834, 0.1523, 0.3109, 0.1774, 0.5460, 0.2260], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0217, 0.0248, 0.0302, 0.0299, 0.0249, 0.0268, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:51:28,734 INFO [finetune.py:976] (2/7) Epoch 10, batch 1450, loss[loss=0.1952, simple_loss=0.2649, pruned_loss=0.06275, over 4892.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2593, pruned_loss=0.0628, over 954644.97 frames. ], batch size: 35, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:51:31,456 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9114, 1.4329, 1.6832, 1.7050, 1.6498, 1.3987, 0.6983, 1.3425], + device='cuda:2'), covar=tensor([0.3937, 0.4168, 0.1967, 0.2746, 0.3026, 0.3076, 0.5048, 0.2489], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0252, 0.0220, 0.0319, 0.0215, 0.0229, 0.0234, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-26 23:51:41,185 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 10, batch 1500, loss[loss=0.2315, simple_loss=0.3036, pruned_loss=0.07969, over 4829.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2596, pruned_loss=0.06308, over 953607.59 frames. ], batch size: 49, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:52:28,802 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53056.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:53:19,949 INFO [finetune.py:976] (2/7) Epoch 10, batch 1550, loss[loss=0.2168, simple_loss=0.2698, pruned_loss=0.0819, over 4838.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2598, pruned_loss=0.06294, over 951985.03 frames. ], batch size: 30, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:53:25,474 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53108.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:53:32,445 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 23:53:53,128 INFO [finetune.py:976] (2/7) Epoch 10, batch 1600, loss[loss=0.1466, simple_loss=0.2214, pruned_loss=0.03588, over 4870.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2588, pruned_loss=0.06306, over 953783.64 frames. ], batch size: 34, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:54:44,152 INFO [finetune.py:976] (2/7) Epoch 10, batch 1650, loss[loss=0.1422, simple_loss=0.2099, pruned_loss=0.03722, over 4763.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2559, pruned_loss=0.06167, over 953997.67 frames. ], batch size: 28, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:54:47,348 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8805, 2.1140, 2.0560, 2.3331, 1.9531, 2.2000, 2.1442, 2.0938], + device='cuda:2'), covar=tensor([0.5297, 0.7899, 0.6267, 0.5240, 0.6976, 0.9074, 0.7905, 0.7626], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0384, 0.0318, 0.0327, 0.0342, 0.0404, 0.0363, 0.0327], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:54:49,097 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7451, 1.3604, 4.4502, 4.1943, 3.8293, 4.0907, 4.0268, 3.9071], + device='cuda:2'), covar=tensor([0.6462, 0.5811, 0.0953, 0.1397, 0.0984, 0.1844, 0.1514, 0.1481], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0305, 0.0404, 0.0407, 0.0348, 0.0404, 0.0313, 0.0370], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-26 23:54:52,154 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53211.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:55:09,622 INFO [optim.py:369] (2/7) 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 23:55:12,825 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 1700, loss[loss=0.1567, simple_loss=0.2293, pruned_loss=0.04208, over 4891.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2544, pruned_loss=0.06142, over 955832.41 frames. ], batch size: 32, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:55:24,105 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53259.0, num_to_drop=1, layers_to_drop={0} +2023-04-26 23:55:31,329 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6157, 1.8140, 1.6846, 2.1325, 1.8993, 2.2401, 1.6686, 3.6564], + device='cuda:2'), covar=tensor([0.0596, 0.0735, 0.0819, 0.1056, 0.0600, 0.0452, 0.0752, 0.0188], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0059], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 23:55:31,353 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7852, 1.5166, 1.8675, 1.9523, 1.6043, 1.3251, 1.5789, 1.1404], + device='cuda:2'), covar=tensor([0.0559, 0.1035, 0.0591, 0.0796, 0.0762, 0.1220, 0.0697, 0.0759], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0072, 0.0070, 0.0067, 0.0075, 0.0094, 0.0077, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 23:55:39,500 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8132, 2.0633, 1.3630, 1.5540, 2.2866, 1.6900, 1.6542, 1.6773], + device='cuda:2'), covar=tensor([0.0444, 0.0298, 0.0297, 0.0463, 0.0253, 0.0459, 0.0408, 0.0477], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-26 23:55:51,480 INFO [finetune.py:976] (2/7) Epoch 10, batch 1750, loss[loss=0.212, simple_loss=0.2819, pruned_loss=0.07112, over 4828.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.256, pruned_loss=0.06225, over 953705.69 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:55:53,419 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53302.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:55:54,633 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53304.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:01,776 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 23:56:25,436 INFO [finetune.py:976] (2/7) Epoch 10, batch 1800, loss[loss=0.2043, simple_loss=0.2696, pruned_loss=0.06947, over 4824.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2592, pruned_loss=0.06338, over 953648.29 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:56:35,666 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53365.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:36,896 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53367.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:43,386 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53376.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:56:52,080 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.87 vs. limit=5.0 +2023-04-26 23:56:59,051 INFO [finetune.py:976] (2/7) Epoch 10, batch 1850, loss[loss=0.1853, simple_loss=0.2695, pruned_loss=0.05055, over 4812.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2607, pruned_loss=0.06395, over 953849.08 frames. ], batch size: 45, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:57:10,977 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53408.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:57:12,815 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3474, 1.1416, 1.5628, 1.5056, 1.2554, 1.0859, 1.1969, 0.8991], + device='cuda:2'), covar=tensor([0.0546, 0.0828, 0.0495, 0.0696, 0.0922, 0.1173, 0.0622, 0.0778], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0072, 0.0070, 0.0067, 0.0075, 0.0095, 0.0077, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-26 23:57:13,861 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53412.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:57:35,492 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53428.0, num_to_drop=1, layers_to_drop={2} +2023-04-26 23:57:45,404 INFO [optim.py:369] (2/7) 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 23:58:05,924 INFO [finetune.py:976] (2/7) Epoch 10, batch 1900, loss[loss=0.2023, simple_loss=0.2802, pruned_loss=0.06222, over 4907.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2617, pruned_loss=0.06389, over 955577.21 frames. ], batch size: 37, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:58:15,434 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53456.0, num_to_drop=0, layers_to_drop=set() +2023-04-26 23:58:48,692 INFO [finetune.py:976] (2/7) Epoch 10, batch 1950, loss[loss=0.2024, simple_loss=0.2551, pruned_loss=0.07485, over 4790.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2596, pruned_loss=0.06284, over 956420.00 frames. ], batch size: 25, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:59:01,477 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5577, 0.9332, 1.5283, 2.0876, 1.7185, 1.5198, 1.5308, 1.5504], + device='cuda:2'), covar=tensor([0.5071, 0.7497, 0.6291, 0.7026, 0.6159, 0.8191, 0.8227, 0.8224], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0414, 0.0499, 0.0517, 0.0436, 0.0455, 0.0466, 0.0465], + device='cuda:2'), out_proj_covar=tensor([9.9596e-05, 1.0255e-04, 1.1247e-04, 1.2294e-04, 1.0593e-04, 1.1005e-04, + 1.1181e-04, 1.1222e-04], device='cuda:2') +2023-04-26 23:59:11,897 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-26 23:59:12,238 INFO [optim.py:369] (2/7) 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 23:59:18,809 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-26 23:59:22,127 INFO [finetune.py:976] (2/7) Epoch 10, batch 2000, loss[loss=0.1971, simple_loss=0.2525, pruned_loss=0.0709, over 4691.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2574, pruned_loss=0.06244, over 954556.87 frames. ], batch size: 23, lr: 3.75e-03, grad_scale: 16.0 +2023-04-26 23:59:23,453 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4700, 1.7620, 1.7512, 2.0756, 1.8980, 1.9311, 1.7024, 3.3943], + device='cuda:2'), covar=tensor([0.0643, 0.0790, 0.0834, 0.1166, 0.0671, 0.0508, 0.0763, 0.0206], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-26 23:59:25,210 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.1600, 4.1725, 3.1048, 4.7991, 4.1884, 4.2244, 2.3348, 4.1494], + device='cuda:2'), covar=tensor([0.1740, 0.1142, 0.2707, 0.1269, 0.2699, 0.1854, 0.5289, 0.1976], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0219, 0.0251, 0.0304, 0.0302, 0.0251, 0.0271, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-26 23:59:54,900 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8725, 1.2224, 3.2989, 3.0537, 2.9487, 3.1744, 3.1875, 2.9013], + device='cuda:2'), covar=tensor([0.7131, 0.5484, 0.1389, 0.2225, 0.1458, 0.2077, 0.1674, 0.1755], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0310, 0.0408, 0.0412, 0.0352, 0.0406, 0.0316, 0.0373], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:00:04,621 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53597.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:00:05,764 INFO [finetune.py:976] (2/7) Epoch 10, batch 2050, loss[loss=0.1475, simple_loss=0.2164, pruned_loss=0.03927, over 4773.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2542, pruned_loss=0.06169, over 953450.19 frames. ], batch size: 28, lr: 3.75e-03, grad_scale: 16.0 +2023-04-27 00:00:25,602 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6081, 3.6321, 2.5852, 4.1495, 3.6970, 3.6917, 1.4098, 3.4909], + device='cuda:2'), covar=tensor([0.1796, 0.1334, 0.3262, 0.1902, 0.3405, 0.1846, 0.6495, 0.2572], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0218, 0.0250, 0.0304, 0.0302, 0.0252, 0.0272, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:00:34,508 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 2100, loss[loss=0.1237, simple_loss=0.1973, pruned_loss=0.02509, over 4703.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2543, pruned_loss=0.06167, over 952931.91 frames. ], batch size: 23, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:00:51,820 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53671.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:01:16,444 INFO [finetune.py:976] (2/7) Epoch 10, batch 2150, loss[loss=0.2439, simple_loss=0.3186, pruned_loss=0.08463, over 4819.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2574, pruned_loss=0.06254, over 950615.90 frames. ], batch size: 47, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:01:26,019 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53712.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:01:32,723 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53723.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 00:01:41,147 INFO [optim.py:369] (2/7) 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 00:01:49,682 INFO [finetune.py:976] (2/7) Epoch 10, batch 2200, loss[loss=0.2093, simple_loss=0.2722, pruned_loss=0.07316, over 4889.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2601, pruned_loss=0.06339, over 952262.21 frames. ], batch size: 35, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:01:57,879 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53760.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:02:22,061 INFO [finetune.py:976] (2/7) Epoch 10, batch 2250, loss[loss=0.156, simple_loss=0.23, pruned_loss=0.04099, over 4922.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2608, pruned_loss=0.06325, over 953702.24 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:02:46,629 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 2300, loss[loss=0.1829, simple_loss=0.2404, pruned_loss=0.06268, over 4886.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2607, pruned_loss=0.06301, over 952172.11 frames. ], batch size: 32, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:03:20,171 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1667, 2.5687, 0.8920, 1.4528, 1.6298, 1.8728, 1.5974, 0.8296], + device='cuda:2'), covar=tensor([0.1597, 0.1103, 0.1807, 0.1473, 0.1185, 0.1069, 0.1711, 0.1689], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0250, 0.0140, 0.0123, 0.0135, 0.0155, 0.0119, 0.0122], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 00:04:10,839 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53897.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:04:11,997 INFO [finetune.py:976] (2/7) Epoch 10, batch 2350, loss[loss=0.1682, simple_loss=0.2218, pruned_loss=0.05729, over 4724.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2571, pruned_loss=0.06165, over 952196.38 frames. ], batch size: 59, lr: 3.75e-03, grad_scale: 32.0 +2023-04-27 00:04:57,666 INFO [optim.py:369] (2/7) 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 00:05:08,688 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:05:16,065 INFO [finetune.py:976] (2/7) Epoch 10, batch 2400, loss[loss=0.2434, simple_loss=0.2854, pruned_loss=0.1007, over 4935.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2539, pruned_loss=0.0607, over 952965.38 frames. ], batch size: 33, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:05:23,337 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53971.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:05:49,429 INFO [finetune.py:976] (2/7) Epoch 10, batch 2450, loss[loss=0.1523, simple_loss=0.2153, pruned_loss=0.04468, over 4757.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2518, pruned_loss=0.06027, over 953068.41 frames. ], batch size: 28, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:05:56,752 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54019.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:06:07,935 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=54023.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:06:15,670 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 2500, loss[loss=0.1679, simple_loss=0.2493, pruned_loss=0.0432, over 4825.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2544, pruned_loss=0.06148, over 953833.89 frames. ], batch size: 39, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:06:39,518 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54071.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:06:39,571 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5418, 1.5710, 0.7280, 1.2723, 1.8013, 1.4256, 1.3185, 1.3883], + device='cuda:2'), covar=tensor([0.0566, 0.0407, 0.0405, 0.0617, 0.0296, 0.0582, 0.0539, 0.0643], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0021, 0.0029, 0.0029, 0.0030], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 00:06:54,299 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5576, 1.3908, 1.9224, 1.8359, 1.3905, 1.2365, 1.5222, 1.0519], + device='cuda:2'), covar=tensor([0.0723, 0.0826, 0.0497, 0.0802, 0.1054, 0.1500, 0.0749, 0.0871], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0072, 0.0070, 0.0066, 0.0075, 0.0095, 0.0077, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 00:06:57,697 INFO [finetune.py:976] (2/7) Epoch 10, batch 2550, loss[loss=0.1988, simple_loss=0.268, pruned_loss=0.06479, over 4866.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2591, pruned_loss=0.06341, over 951552.65 frames. ], batch size: 31, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:06:57,829 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3962, 1.1856, 1.6280, 1.5450, 1.2596, 1.1604, 1.2733, 0.8861], + device='cuda:2'), covar=tensor([0.0622, 0.0903, 0.0501, 0.0764, 0.1044, 0.1522, 0.0675, 0.0793], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0072, 0.0070, 0.0066, 0.0075, 0.0095, 0.0077, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 00:07:03,593 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-04-27 00:07:10,096 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 00:07:11,136 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9686, 2.0736, 1.6244, 1.6460, 2.0071, 1.5818, 2.4899, 1.3282], + device='cuda:2'), covar=tensor([0.3496, 0.1418, 0.4415, 0.2558, 0.1630, 0.2421, 0.1390, 0.4469], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0349, 0.0432, 0.0362, 0.0387, 0.0384, 0.0380, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:07:22,641 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 2600, loss[loss=0.1669, simple_loss=0.2431, pruned_loss=0.04541, over 4808.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2586, pruned_loss=0.06255, over 951018.31 frames. ], batch size: 39, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:07:50,793 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1578, 2.8703, 2.1377, 2.1637, 1.7038, 1.7356, 2.2322, 1.6770], + device='cuda:2'), covar=tensor([0.1561, 0.1149, 0.1445, 0.1499, 0.2120, 0.1794, 0.0990, 0.1874], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0214, 0.0170, 0.0204, 0.0204, 0.0184, 0.0160, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 00:07:53,061 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-27 00:08:04,393 INFO [finetune.py:976] (2/7) Epoch 10, batch 2650, loss[loss=0.2228, simple_loss=0.29, pruned_loss=0.07777, over 4790.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2602, pruned_loss=0.06312, over 949943.65 frames. ], batch size: 25, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:08:16,764 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7608, 1.6839, 2.1894, 2.2540, 1.5877, 1.3985, 1.8317, 1.0762], + device='cuda:2'), covar=tensor([0.0813, 0.1055, 0.0624, 0.0978, 0.1159, 0.1352, 0.0967, 0.1050], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0072, 0.0070, 0.0066, 0.0075, 0.0095, 0.0076, 0.0072], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 00:08:39,721 INFO [optim.py:369] (2/7) 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 00:08:41,125 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8526, 1.9059, 1.8567, 1.5408, 2.0796, 1.7203, 2.6543, 1.6167], + device='cuda:2'), covar=tensor([0.3545, 0.1731, 0.4256, 0.2860, 0.1545, 0.2269, 0.1254, 0.4466], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0345, 0.0427, 0.0358, 0.0385, 0.0381, 0.0376, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:08:52,970 INFO [finetune.py:976] (2/7) Epoch 10, batch 2700, loss[loss=0.1776, simple_loss=0.2445, pruned_loss=0.05538, over 4923.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2585, pruned_loss=0.0619, over 951293.17 frames. ], batch size: 33, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:08:59,672 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 00:10:03,550 INFO [finetune.py:976] (2/7) Epoch 10, batch 2750, loss[loss=0.1877, simple_loss=0.2496, pruned_loss=0.06288, over 4821.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2567, pruned_loss=0.0619, over 952238.26 frames. ], batch size: 38, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:10:49,709 INFO [optim.py:369] (2/7) 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 00:11:09,557 INFO [finetune.py:976] (2/7) Epoch 10, batch 2800, loss[loss=0.1684, simple_loss=0.2418, pruned_loss=0.04749, over 4937.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2544, pruned_loss=0.06169, over 953525.82 frames. ], batch size: 33, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:11:20,062 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6226, 1.9885, 2.5041, 3.2300, 2.4383, 1.8995, 1.9265, 2.3948], + device='cuda:2'), covar=tensor([0.3859, 0.3799, 0.1734, 0.2934, 0.3167, 0.3011, 0.4496, 0.2564], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0249, 0.0219, 0.0313, 0.0212, 0.0227, 0.0232, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:11:43,456 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-27 00:11:48,127 INFO [finetune.py:976] (2/7) Epoch 10, batch 2850, loss[loss=0.1936, simple_loss=0.2584, pruned_loss=0.06435, over 4858.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2531, pruned_loss=0.06125, over 955236.31 frames. ], batch size: 31, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:12:11,872 INFO [optim.py:369] (2/7) 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 00:12:21,812 INFO [finetune.py:976] (2/7) Epoch 10, batch 2900, loss[loss=0.1639, simple_loss=0.2376, pruned_loss=0.04515, over 4755.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2565, pruned_loss=0.0628, over 955600.34 frames. ], batch size: 26, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:12:28,667 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1467, 3.0937, 0.8175, 1.4543, 1.4495, 2.1186, 1.5976, 1.0620], + device='cuda:2'), covar=tensor([0.2015, 0.1447, 0.2571, 0.2089, 0.1585, 0.1411, 0.1906, 0.2379], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0249, 0.0140, 0.0123, 0.0135, 0.0154, 0.0119, 0.0122], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 00:12:41,621 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-04-27 00:12:55,779 INFO [finetune.py:976] (2/7) Epoch 10, batch 2950, loss[loss=0.2848, simple_loss=0.3422, pruned_loss=0.1137, over 4142.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.26, pruned_loss=0.0634, over 955925.38 frames. ], batch size: 65, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:13:19,227 INFO [optim.py:369] (2/7) 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 00:13:21,681 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3849, 1.6180, 1.7044, 1.8731, 1.7847, 1.9511, 1.8873, 1.7553], + device='cuda:2'), covar=tensor([0.5113, 0.6434, 0.5772, 0.5268, 0.6516, 0.9198, 0.6472, 0.5816], + device='cuda:2'), in_proj_covar=tensor([0.0322, 0.0377, 0.0312, 0.0322, 0.0335, 0.0397, 0.0355, 0.0320], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:13:29,120 INFO [finetune.py:976] (2/7) Epoch 10, batch 3000, loss[loss=0.1611, simple_loss=0.2396, pruned_loss=0.04127, over 4755.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2613, pruned_loss=0.064, over 954634.30 frames. ], batch size: 27, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:13:29,120 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 00:13:37,378 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1714, 1.4210, 1.5942, 1.7705, 1.7125, 1.8596, 1.6804, 1.6437], + device='cuda:2'), covar=tensor([0.4718, 0.7049, 0.6102, 0.5556, 0.6732, 0.9067, 0.6478, 0.6596], + device='cuda:2'), in_proj_covar=tensor([0.0323, 0.0378, 0.0313, 0.0323, 0.0336, 0.0398, 0.0356, 0.0321], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:13:45,198 INFO [finetune.py:1010] (2/7) Epoch 10, validation: loss=0.1531, simple_loss=0.2257, pruned_loss=0.04026, over 2265189.00 frames. +2023-04-27 00:13:45,199 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 00:14:32,347 INFO [finetune.py:976] (2/7) Epoch 10, batch 3050, loss[loss=0.1968, simple_loss=0.267, pruned_loss=0.06329, over 4807.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2617, pruned_loss=0.06331, over 955208.89 frames. ], batch size: 38, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:14:57,004 INFO [optim.py:369] (2/7) 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 00:15:05,021 INFO [finetune.py:976] (2/7) Epoch 10, batch 3100, loss[loss=0.1706, simple_loss=0.2243, pruned_loss=0.05844, over 4830.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2606, pruned_loss=0.0633, over 954682.37 frames. ], batch size: 25, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:15:27,262 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54664.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:15:57,336 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7918, 2.4064, 1.8824, 1.8258, 1.3964, 1.4163, 1.9453, 1.3552], + device='cuda:2'), covar=tensor([0.1291, 0.1121, 0.1286, 0.1463, 0.1973, 0.1612, 0.0828, 0.1738], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0204, 0.0204, 0.0184, 0.0159, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 00:16:11,742 INFO [finetune.py:976] (2/7) Epoch 10, batch 3150, loss[loss=0.1881, simple_loss=0.2412, pruned_loss=0.06751, over 4806.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2584, pruned_loss=0.06244, over 955263.30 frames. ], batch size: 51, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:16:53,622 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54725.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 00:17:04,898 INFO [optim.py:369] (2/7) 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 00:17:25,064 INFO [finetune.py:976] (2/7) Epoch 10, batch 3200, loss[loss=0.182, simple_loss=0.2516, pruned_loss=0.05616, over 4788.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2539, pruned_loss=0.06063, over 955892.71 frames. ], batch size: 29, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:18:08,047 INFO [finetune.py:976] (2/7) Epoch 10, batch 3250, loss[loss=0.2102, simple_loss=0.2762, pruned_loss=0.07209, over 4892.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2559, pruned_loss=0.06211, over 955005.80 frames. ], batch size: 32, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:18:33,642 INFO [optim.py:369] (2/7) 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 00:18:38,568 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-04-27 00:18:42,049 INFO [finetune.py:976] (2/7) Epoch 10, batch 3300, loss[loss=0.196, simple_loss=0.2736, pruned_loss=0.05924, over 4802.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2606, pruned_loss=0.06381, over 954687.29 frames. ], batch size: 29, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:18:50,584 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54862.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:19:09,157 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-04-27 00:19:11,996 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9464, 2.0087, 1.7780, 1.5250, 2.0825, 1.7464, 2.6920, 1.5827], + device='cuda:2'), covar=tensor([0.4303, 0.1984, 0.5232, 0.3427, 0.1886, 0.2765, 0.1824, 0.5137], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0348, 0.0432, 0.0362, 0.0387, 0.0383, 0.0379, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:19:15,374 INFO [finetune.py:976] (2/7) Epoch 10, batch 3350, loss[loss=0.2632, simple_loss=0.322, pruned_loss=0.1022, over 4886.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2621, pruned_loss=0.06428, over 955196.68 frames. ], batch size: 35, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:19:20,640 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 00:19:30,568 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54923.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:19:34,462 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3663, 2.9071, 0.9220, 1.6329, 1.9584, 1.6182, 4.0811, 2.2655], + device='cuda:2'), covar=tensor([0.0673, 0.0925, 0.0954, 0.1270, 0.0634, 0.0978, 0.0197, 0.0571], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0050, 0.0048, 0.0052, 0.0053, 0.0079, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-27 00:19:39,810 INFO [optim.py:369] (2/7) 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 00:19:47,702 INFO [finetune.py:976] (2/7) Epoch 10, batch 3400, loss[loss=0.2005, simple_loss=0.2712, pruned_loss=0.06488, over 4895.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2632, pruned_loss=0.0645, over 954236.83 frames. ], batch size: 37, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:20:13,386 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8311, 1.4742, 1.9969, 2.3223, 2.0221, 1.8675, 1.9892, 1.8935], + device='cuda:2'), covar=tensor([0.5623, 0.7686, 0.7540, 0.7191, 0.6616, 0.9002, 0.8336, 0.8053], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0414, 0.0496, 0.0515, 0.0436, 0.0455, 0.0465, 0.0463], + device='cuda:2'), out_proj_covar=tensor([9.9310e-05, 1.0265e-04, 1.1204e-04, 1.2246e-04, 1.0580e-04, 1.0999e-04, + 1.1155e-04, 1.1148e-04], device='cuda:2') +2023-04-27 00:20:20,499 INFO [finetune.py:976] (2/7) Epoch 10, batch 3450, loss[loss=0.2399, simple_loss=0.3039, pruned_loss=0.08793, over 4914.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2634, pruned_loss=0.06446, over 954961.75 frames. ], batch size: 36, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:20:20,630 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6765, 2.0883, 1.5451, 1.2959, 1.2358, 1.2171, 1.5258, 1.1615], + device='cuda:2'), covar=tensor([0.1588, 0.1289, 0.1603, 0.1939, 0.2384, 0.1983, 0.1101, 0.2052], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0215, 0.0170, 0.0204, 0.0204, 0.0184, 0.0160, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 00:20:34,369 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55020.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 00:20:45,488 INFO [optim.py:369] (2/7) 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 00:20:53,437 INFO [finetune.py:976] (2/7) Epoch 10, batch 3500, loss[loss=0.1722, simple_loss=0.2282, pruned_loss=0.05809, over 4761.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2607, pruned_loss=0.06405, over 952941.40 frames. ], batch size: 54, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:20:56,068 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 00:21:35,122 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9099, 1.7322, 2.1789, 2.4353, 2.1037, 1.8884, 2.0681, 2.0120], + device='cuda:2'), covar=tensor([0.5853, 0.8838, 0.9116, 0.7886, 0.7325, 1.0749, 1.1067, 1.0954], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0415, 0.0499, 0.0517, 0.0437, 0.0457, 0.0467, 0.0464], + device='cuda:2'), out_proj_covar=tensor([9.9552e-05, 1.0283e-04, 1.1269e-04, 1.2278e-04, 1.0607e-04, 1.1044e-04, + 1.1187e-04, 1.1191e-04], device='cuda:2') +2023-04-27 00:21:46,852 INFO [finetune.py:976] (2/7) Epoch 10, batch 3550, loss[loss=0.1658, simple_loss=0.2315, pruned_loss=0.05004, over 4853.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2568, pruned_loss=0.0624, over 953906.65 frames. ], batch size: 47, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:22:27,074 INFO [optim.py:369] (2/7) 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 00:22:32,299 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8162, 1.7659, 1.7083, 1.3462, 1.9311, 1.4766, 2.5153, 1.5054], + device='cuda:2'), covar=tensor([0.4175, 0.2069, 0.5389, 0.3450, 0.1918, 0.2894, 0.1496, 0.4964], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0349, 0.0433, 0.0363, 0.0388, 0.0385, 0.0381, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:22:34,762 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 00:22:35,879 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 3600, loss[loss=0.2228, simple_loss=0.2731, pruned_loss=0.08623, over 4808.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2536, pruned_loss=0.0617, over 954646.92 frames. ], batch size: 39, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:23:26,775 INFO [finetune.py:976] (2/7) Epoch 10, batch 3650, loss[loss=0.2479, simple_loss=0.2844, pruned_loss=0.1057, over 4241.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2546, pruned_loss=0.062, over 954293.14 frames. ], batch size: 18, lr: 3.74e-03, grad_scale: 32.0 +2023-04-27 00:23:26,953 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-27 00:23:33,137 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55209.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:23:34,369 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55211.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:23:39,097 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55218.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:23:50,727 INFO [optim.py:369] (2/7) 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 00:24:00,577 INFO [finetune.py:976] (2/7) Epoch 10, batch 3700, loss[loss=0.1549, simple_loss=0.2145, pruned_loss=0.04769, over 4676.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2581, pruned_loss=0.06263, over 955113.48 frames. ], batch size: 23, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:24:09,145 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 00:24:15,039 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:24:33,934 INFO [finetune.py:976] (2/7) Epoch 10, batch 3750, loss[loss=0.1798, simple_loss=0.2492, pruned_loss=0.05524, over 4824.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2603, pruned_loss=0.06319, over 952844.68 frames. ], batch size: 38, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:24:36,624 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-04-27 00:24:47,356 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55320.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:24:57,145 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 3800, loss[loss=0.1547, simple_loss=0.2295, pruned_loss=0.03996, over 4839.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2634, pruned_loss=0.06446, over 951777.78 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:25:19,728 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55368.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:25:40,047 INFO [finetune.py:976] (2/7) Epoch 10, batch 3850, loss[loss=0.1662, simple_loss=0.2338, pruned_loss=0.04928, over 4925.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2617, pruned_loss=0.06402, over 951363.71 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:25:46,623 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-04-27 00:25:52,681 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4717, 1.2246, 4.1842, 3.8998, 3.7151, 3.9740, 3.9789, 3.6752], + device='cuda:2'), covar=tensor([0.6985, 0.6143, 0.1087, 0.1714, 0.1024, 0.1516, 0.1430, 0.1431], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0312, 0.0411, 0.0415, 0.0354, 0.0411, 0.0319, 0.0374], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:26:04,612 INFO [optim.py:369] (2/7) 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 00:26:08,265 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3702, 2.1798, 2.4796, 2.7650, 2.3719, 2.1438, 2.3234, 2.2067], + device='cuda:2'), covar=tensor([0.5926, 0.7999, 0.8612, 0.8134, 0.7726, 1.0808, 1.0570, 0.9327], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0414, 0.0497, 0.0517, 0.0437, 0.0456, 0.0467, 0.0464], + device='cuda:2'), out_proj_covar=tensor([9.9328e-05, 1.0247e-04, 1.1221e-04, 1.2280e-04, 1.0602e-04, 1.1014e-04, + 1.1184e-04, 1.1166e-04], device='cuda:2') +2023-04-27 00:26:12,956 INFO [finetune.py:976] (2/7) Epoch 10, batch 3900, loss[loss=0.1708, simple_loss=0.2363, pruned_loss=0.05261, over 4751.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2588, pruned_loss=0.06336, over 953843.86 frames. ], batch size: 27, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:27:07,758 INFO [finetune.py:976] (2/7) Epoch 10, batch 3950, loss[loss=0.1684, simple_loss=0.2341, pruned_loss=0.0514, over 4822.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2558, pruned_loss=0.06245, over 954354.40 frames. ], batch size: 39, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:27:15,660 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55504.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:27:36,380 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55518.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:27:47,672 INFO [optim.py:369] (2/7) 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 00:27:56,581 INFO [finetune.py:976] (2/7) Epoch 10, batch 4000, loss[loss=0.226, simple_loss=0.3022, pruned_loss=0.07491, over 4817.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2554, pruned_loss=0.06251, over 955125.41 frames. ], batch size: 40, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:28:09,060 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55566.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:28:12,129 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55567.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:28:37,417 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.7700, 4.6909, 3.3161, 5.4986, 4.8383, 4.7483, 2.5230, 4.7911], + device='cuda:2'), covar=tensor([0.1376, 0.0993, 0.2472, 0.0830, 0.3594, 0.1576, 0.4866, 0.1905], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0215, 0.0248, 0.0305, 0.0300, 0.0249, 0.0268, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:28:43,372 INFO [finetune.py:976] (2/7) Epoch 10, batch 4050, loss[loss=0.2333, simple_loss=0.3034, pruned_loss=0.08154, over 4811.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2585, pruned_loss=0.06356, over 955691.19 frames. ], batch size: 41, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:29:09,491 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 4100, loss[loss=0.2249, simple_loss=0.2816, pruned_loss=0.08414, over 4790.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2616, pruned_loss=0.06442, over 955902.14 frames. ], batch size: 25, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:29:50,669 INFO [finetune.py:976] (2/7) Epoch 10, batch 4150, loss[loss=0.1892, simple_loss=0.2609, pruned_loss=0.05876, over 4896.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2622, pruned_loss=0.06414, over 954125.65 frames. ], batch size: 36, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:30:16,065 INFO [optim.py:369] (2/7) 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 00:30:23,239 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55748.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:30:23,766 INFO [finetune.py:976] (2/7) Epoch 10, batch 4200, loss[loss=0.1676, simple_loss=0.2319, pruned_loss=0.05161, over 4912.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2609, pruned_loss=0.06314, over 956809.14 frames. ], batch size: 37, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:30:57,481 INFO [finetune.py:976] (2/7) Epoch 10, batch 4250, loss[loss=0.1526, simple_loss=0.225, pruned_loss=0.04012, over 4757.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2589, pruned_loss=0.06283, over 957949.87 frames. ], batch size: 26, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:30:58,298 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4009, 1.1923, 4.0306, 3.7551, 3.5662, 3.8426, 3.8217, 3.5537], + device='cuda:2'), covar=tensor([0.7021, 0.6108, 0.1063, 0.1789, 0.1007, 0.1508, 0.1681, 0.1507], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0310, 0.0407, 0.0411, 0.0352, 0.0408, 0.0317, 0.0373], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:31:00,760 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55804.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:31:04,328 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55809.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:31:23,548 INFO [optim.py:369] (2/7) 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 00:31:30,791 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 4300, loss[loss=0.1896, simple_loss=0.2574, pruned_loss=0.06088, over 4784.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2558, pruned_loss=0.06105, over 958721.93 frames. ], batch size: 29, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:31:33,044 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55867.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:32:31,697 INFO [finetune.py:976] (2/7) Epoch 10, batch 4350, loss[loss=0.1849, simple_loss=0.2548, pruned_loss=0.05756, over 4927.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2526, pruned_loss=0.05953, over 959850.34 frames. ], batch size: 38, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:32:43,365 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:32:45,227 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5723, 1.9400, 2.4878, 3.1121, 2.3603, 1.8226, 1.7834, 2.2552], + device='cuda:2'), covar=tensor([0.3677, 0.3735, 0.1665, 0.2830, 0.3171, 0.2990, 0.4475, 0.2597], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0249, 0.0219, 0.0316, 0.0214, 0.0227, 0.0233, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:32:46,921 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55915.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:33:02,710 INFO [optim.py:369] (2/7) 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 00:33:10,575 INFO [finetune.py:976] (2/7) Epoch 10, batch 4400, loss[loss=0.2591, simple_loss=0.3175, pruned_loss=0.1004, over 4755.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2531, pruned_loss=0.05981, over 958196.04 frames. ], batch size: 54, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:33:21,143 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7444, 2.1589, 0.9277, 1.1002, 1.5396, 1.0777, 2.2872, 1.2630], + device='cuda:2'), covar=tensor([0.0668, 0.0797, 0.0653, 0.1078, 0.0447, 0.0958, 0.0291, 0.0642], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0050, 0.0047, 0.0052, 0.0053, 0.0079, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-27 00:33:42,938 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8153, 2.0866, 1.7022, 1.4496, 1.4274, 1.4256, 1.7347, 1.3544], + device='cuda:2'), covar=tensor([0.1740, 0.1454, 0.1551, 0.1892, 0.2491, 0.2136, 0.1183, 0.2112], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0213, 0.0169, 0.0203, 0.0204, 0.0184, 0.0159, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 00:34:18,723 INFO [finetune.py:976] (2/7) Epoch 10, batch 4450, loss[loss=0.2041, simple_loss=0.273, pruned_loss=0.06765, over 4894.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2577, pruned_loss=0.06225, over 954525.55 frames. ], batch size: 35, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:34:41,084 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8524, 1.3496, 1.6588, 1.6178, 1.6133, 1.3087, 0.7682, 1.3344], + device='cuda:2'), covar=tensor([0.3634, 0.3751, 0.1867, 0.2580, 0.2966, 0.2900, 0.4636, 0.2592], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0250, 0.0220, 0.0317, 0.0214, 0.0228, 0.0233, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:34:57,405 INFO [optim.py:369] (2/7) 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 00:35:05,273 INFO [finetune.py:976] (2/7) Epoch 10, batch 4500, loss[loss=0.2072, simple_loss=0.2753, pruned_loss=0.06959, over 4921.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2584, pruned_loss=0.0623, over 954913.63 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:35:14,480 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-04-27 00:35:16,804 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0422, 1.7038, 1.9085, 2.1167, 1.8958, 1.5316, 1.1354, 1.7569], + device='cuda:2'), covar=tensor([0.3893, 0.3670, 0.1867, 0.2474, 0.2771, 0.2948, 0.4576, 0.2201], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0249, 0.0220, 0.0316, 0.0214, 0.0227, 0.0233, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:35:26,613 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2215, 1.6244, 1.4825, 1.7948, 1.6646, 2.0088, 1.5122, 3.5142], + device='cuda:2'), covar=tensor([0.0588, 0.0763, 0.0722, 0.1075, 0.0594, 0.0478, 0.0680, 0.0132], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 00:35:38,739 INFO [finetune.py:976] (2/7) Epoch 10, batch 4550, loss[loss=0.2299, simple_loss=0.2859, pruned_loss=0.08692, over 4832.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2611, pruned_loss=0.06373, over 954685.88 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:35:41,844 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56104.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:35:49,089 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8518, 2.2519, 1.2611, 1.5343, 2.2075, 1.7022, 1.6737, 1.7058], + device='cuda:2'), covar=tensor([0.0535, 0.0359, 0.0321, 0.0563, 0.0260, 0.0542, 0.0561, 0.0600], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 00:35:58,650 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6153, 1.2269, 1.2266, 1.4327, 1.8356, 1.4541, 1.2402, 1.1560], + device='cuda:2'), covar=tensor([0.1473, 0.1388, 0.1830, 0.1252, 0.0814, 0.1600, 0.2030, 0.1916], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0323, 0.0355, 0.0300, 0.0338, 0.0321, 0.0307, 0.0361], + device='cuda:2'), out_proj_covar=tensor([6.4476e-05, 6.8404e-05, 7.6614e-05, 6.1854e-05, 7.0858e-05, 6.8610e-05, + 6.5655e-05, 7.7494e-05], device='cuda:2') +2023-04-27 00:36:03,337 INFO [optim.py:369] (2/7) 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 00:36:09,775 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 4600, loss[loss=0.2273, simple_loss=0.2853, pruned_loss=0.08462, over 4809.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2605, pruned_loss=0.06315, over 955409.62 frames. ], batch size: 39, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:36:20,016 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1111, 1.6192, 1.9116, 2.4088, 1.8313, 1.5188, 1.1429, 1.7079], + device='cuda:2'), covar=tensor([0.3404, 0.3592, 0.1810, 0.2315, 0.2761, 0.2859, 0.4709, 0.2466], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0251, 0.0221, 0.0317, 0.0214, 0.0228, 0.0234, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:36:45,854 INFO [finetune.py:976] (2/7) Epoch 10, batch 4650, loss[loss=0.1867, simple_loss=0.2499, pruned_loss=0.06175, over 4940.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2587, pruned_loss=0.0632, over 954612.85 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:36:49,602 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56204.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:36:50,846 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:37:16,888 INFO [optim.py:369] (2/7) 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 00:37:20,204 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 00:37:24,882 INFO [finetune.py:976] (2/7) Epoch 10, batch 4700, loss[loss=0.1866, simple_loss=0.2413, pruned_loss=0.06591, over 4926.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2549, pruned_loss=0.06168, over 956085.30 frames. ], batch size: 38, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:37:27,461 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-04-27 00:37:57,893 INFO [finetune.py:976] (2/7) Epoch 10, batch 4750, loss[loss=0.1931, simple_loss=0.256, pruned_loss=0.06512, over 4872.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2519, pruned_loss=0.06061, over 955189.96 frames. ], batch size: 31, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:38:08,546 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5293, 3.4382, 2.7027, 4.0896, 3.4556, 3.4836, 1.6622, 3.5232], + device='cuda:2'), covar=tensor([0.1837, 0.1423, 0.4117, 0.1634, 0.3867, 0.2008, 0.5295, 0.2552], + device='cuda:2'), in_proj_covar=tensor([0.0240, 0.0214, 0.0246, 0.0302, 0.0296, 0.0248, 0.0264, 0.0267], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:38:18,931 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5173, 1.5819, 1.3133, 0.8895, 1.1584, 1.1193, 1.2630, 1.0394], + device='cuda:2'), covar=tensor([0.1842, 0.1315, 0.1667, 0.2032, 0.2501, 0.2195, 0.1198, 0.2215], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0213, 0.0169, 0.0203, 0.0204, 0.0184, 0.0159, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 00:38:23,667 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 10, batch 4800, loss[loss=0.2454, simple_loss=0.3175, pruned_loss=0.08666, over 4849.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2582, pruned_loss=0.064, over 953738.18 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:39:33,342 INFO [finetune.py:976] (2/7) Epoch 10, batch 4850, loss[loss=0.1849, simple_loss=0.261, pruned_loss=0.05434, over 4838.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2606, pruned_loss=0.06411, over 954816.17 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:39:43,309 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56404.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:40:01,636 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0755, 2.1608, 1.9017, 1.8189, 2.4079, 1.9289, 2.8586, 1.6761], + device='cuda:2'), covar=tensor([0.3797, 0.1765, 0.4577, 0.2964, 0.1593, 0.2398, 0.1244, 0.4421], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0348, 0.0431, 0.0362, 0.0387, 0.0384, 0.0382, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:40:07,731 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6454, 1.1925, 1.7901, 2.1137, 1.7815, 1.6594, 1.7551, 1.7175], + device='cuda:2'), covar=tensor([0.5814, 0.8206, 0.7900, 0.8192, 0.7402, 0.9723, 0.9489, 0.9629], + device='cuda:2'), in_proj_covar=tensor([0.0410, 0.0413, 0.0497, 0.0517, 0.0437, 0.0458, 0.0468, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9839e-05, 1.0235e-04, 1.1229e-04, 1.2292e-04, 1.0589e-04, 1.1064e-04, + 1.1218e-04, 1.1223e-04], device='cuda:2') +2023-04-27 00:40:14,074 INFO [optim.py:369] (2/7) 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 00:40:22,379 INFO [finetune.py:976] (2/7) Epoch 10, batch 4900, loss[loss=0.1683, simple_loss=0.2326, pruned_loss=0.05201, over 4406.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2622, pruned_loss=0.06436, over 955515.84 frames. ], batch size: 19, lr: 3.73e-03, grad_scale: 32.0 +2023-04-27 00:40:24,261 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56452.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:40:56,298 INFO [finetune.py:976] (2/7) Epoch 10, batch 4950, loss[loss=0.1787, simple_loss=0.2445, pruned_loss=0.05649, over 4899.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2628, pruned_loss=0.06439, over 956185.31 frames. ], batch size: 36, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:40:57,602 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56501.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:40:59,476 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) 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 00:41:29,726 INFO [finetune.py:976] (2/7) Epoch 10, batch 5000, loss[loss=0.1859, simple_loss=0.2551, pruned_loss=0.05837, over 4767.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2603, pruned_loss=0.06367, over 954499.44 frames. ], batch size: 51, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:41:32,066 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56552.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:41:53,707 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56584.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:42:03,263 INFO [finetune.py:976] (2/7) Epoch 10, batch 5050, loss[loss=0.192, simple_loss=0.2583, pruned_loss=0.06284, over 4820.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2574, pruned_loss=0.06267, over 953224.14 frames. ], batch size: 38, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:42:56,298 INFO [optim.py:369] (2/7) 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 00:43:09,521 INFO [finetune.py:976] (2/7) Epoch 10, batch 5100, loss[loss=0.1616, simple_loss=0.2278, pruned_loss=0.04773, over 4824.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2528, pruned_loss=0.06056, over 956142.31 frames. ], batch size: 38, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:43:59,160 INFO [finetune.py:976] (2/7) Epoch 10, batch 5150, loss[loss=0.179, simple_loss=0.2569, pruned_loss=0.05053, over 4738.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2534, pruned_loss=0.06089, over 954589.16 frames. ], batch size: 27, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:44:00,529 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4089, 0.6400, 1.3932, 1.8148, 1.5124, 1.3929, 1.4284, 1.4683], + device='cuda:2'), covar=tensor([0.5135, 0.7338, 0.6986, 0.7539, 0.6506, 0.8806, 0.8100, 0.8225], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0414, 0.0497, 0.0517, 0.0438, 0.0458, 0.0469, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9684e-05, 1.0259e-04, 1.1226e-04, 1.2294e-04, 1.0601e-04, 1.1071e-04, + 1.1233e-04, 1.1224e-04], device='cuda:2') +2023-04-27 00:44:25,420 INFO [optim.py:369] (2/7) 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 00:44:26,769 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4155, 1.2386, 1.6142, 1.5121, 1.2620, 1.1306, 1.3144, 0.7942], + device='cuda:2'), covar=tensor([0.0628, 0.0843, 0.0460, 0.0815, 0.0909, 0.1308, 0.0653, 0.0792], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0073, 0.0071, 0.0068, 0.0076, 0.0096, 0.0077, 0.0073], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 00:44:33,209 INFO [finetune.py:976] (2/7) Epoch 10, batch 5200, loss[loss=0.212, simple_loss=0.2921, pruned_loss=0.06597, over 4911.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2576, pruned_loss=0.0622, over 953817.49 frames. ], batch size: 36, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:45:10,353 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0920, 1.5766, 1.9641, 2.2502, 1.9714, 1.5075, 1.1154, 1.7793], + device='cuda:2'), covar=tensor([0.3748, 0.3728, 0.1745, 0.2822, 0.2743, 0.2931, 0.4733, 0.2322], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0248, 0.0218, 0.0314, 0.0213, 0.0226, 0.0231, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:45:18,108 INFO [finetune.py:976] (2/7) Epoch 10, batch 5250, loss[loss=0.1725, simple_loss=0.2546, pruned_loss=0.04526, over 4805.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2595, pruned_loss=0.06265, over 955684.20 frames. ], batch size: 40, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:45:20,018 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56801.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:45:21,234 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56803.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:45:28,604 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-27 00:45:33,853 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0339, 1.5914, 1.9590, 2.2788, 2.3560, 1.9460, 1.5693, 2.0895], + device='cuda:2'), covar=tensor([0.0702, 0.1089, 0.0577, 0.0444, 0.0463, 0.0702, 0.0790, 0.0459], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0202, 0.0181, 0.0174, 0.0177, 0.0187, 0.0159, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:45:42,572 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7645, 2.0886, 1.6850, 1.4186, 1.2738, 1.2718, 1.7343, 1.2303], + device='cuda:2'), covar=tensor([0.1698, 0.1372, 0.1685, 0.1863, 0.2440, 0.2121, 0.1105, 0.2171], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0203, 0.0204, 0.0185, 0.0159, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 00:45:43,652 INFO [optim.py:369] (2/7) 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 00:45:50,966 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 00:45:51,393 INFO [finetune.py:976] (2/7) Epoch 10, batch 5300, loss[loss=0.24, simple_loss=0.2921, pruned_loss=0.09392, over 4808.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2607, pruned_loss=0.06317, over 954876.48 frames. ], batch size: 39, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:45:51,456 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56849.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:46:01,123 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56879.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:46:25,223 INFO [finetune.py:976] (2/7) Epoch 10, batch 5350, loss[loss=0.1683, simple_loss=0.2439, pruned_loss=0.04634, over 4764.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2602, pruned_loss=0.06264, over 954797.87 frames. ], batch size: 27, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:46:31,344 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56908.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:46:50,603 INFO [optim.py:369] (2/7) 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 00:46:57,806 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8664, 2.7873, 2.2210, 3.2645, 2.9196, 2.8171, 1.1684, 2.7436], + device='cuda:2'), covar=tensor([0.2079, 0.1711, 0.3463, 0.2899, 0.3576, 0.2240, 0.6183, 0.3055], + device='cuda:2'), in_proj_covar=tensor([0.0240, 0.0214, 0.0247, 0.0301, 0.0297, 0.0247, 0.0265, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:46:58,359 INFO [finetune.py:976] (2/7) Epoch 10, batch 5400, loss[loss=0.2105, simple_loss=0.2885, pruned_loss=0.06622, over 4905.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2581, pruned_loss=0.062, over 956138.70 frames. ], batch size: 36, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:47:11,570 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56969.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:47:15,844 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 00:47:28,552 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56993.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:47:29,793 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 5450, loss[loss=0.1593, simple_loss=0.2145, pruned_loss=0.05206, over 4773.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2561, pruned_loss=0.06202, over 956003.65 frames. ], batch size: 26, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:48:18,179 INFO [optim.py:369] (2/7) 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 00:48:38,379 INFO [finetune.py:976] (2/7) Epoch 10, batch 5500, loss[loss=0.1498, simple_loss=0.2206, pruned_loss=0.03948, over 4214.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2528, pruned_loss=0.06107, over 956146.07 frames. ], batch size: 18, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:48:47,141 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57054.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:48:48,360 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57056.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:49:09,192 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 00:49:35,989 INFO [finetune.py:976] (2/7) Epoch 10, batch 5550, loss[loss=0.1649, simple_loss=0.2488, pruned_loss=0.0405, over 4812.00 frames. ], tot_loss[loss=0.188, simple_loss=0.254, pruned_loss=0.06097, over 955099.10 frames. ], batch size: 41, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:49:47,136 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-27 00:49:59,957 INFO [optim.py:369] (2/7) 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 00:50:06,923 INFO [finetune.py:976] (2/7) Epoch 10, batch 5600, loss[loss=0.1914, simple_loss=0.2572, pruned_loss=0.0628, over 4748.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2576, pruned_loss=0.06181, over 953088.79 frames. ], batch size: 27, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:50:12,764 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57159.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:50:19,806 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 00:50:24,871 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57179.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:50:29,404 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57186.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:50:35,750 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 5650, loss[loss=0.177, simple_loss=0.2479, pruned_loss=0.05302, over 4756.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2601, pruned_loss=0.06213, over 953510.09 frames. ], batch size: 27, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:50:42,558 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7039, 1.3767, 1.8051, 2.2185, 1.8373, 1.6489, 1.7463, 1.7330], + device='cuda:2'), covar=tensor([0.6091, 0.8203, 0.8407, 0.7526, 0.7025, 1.0195, 1.0049, 0.9865], + device='cuda:2'), in_proj_covar=tensor([0.0407, 0.0411, 0.0496, 0.0515, 0.0436, 0.0456, 0.0466, 0.0465], + device='cuda:2'), 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:2') +2023-04-27 00:50:53,686 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57227.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 00:50:56,032 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9211, 3.7622, 1.1614, 2.0765, 2.2464, 2.7162, 2.2506, 1.3814], + device='cuda:2'), covar=tensor([0.1377, 0.1080, 0.2047, 0.1309, 0.1095, 0.1069, 0.1380, 0.1770], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0249, 0.0142, 0.0122, 0.0134, 0.0153, 0.0118, 0.0123], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 00:50:57,838 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5078, 2.2321, 2.4229, 2.6735, 2.7289, 2.2781, 1.9279, 2.5432], + device='cuda:2'), covar=tensor([0.0707, 0.0882, 0.0596, 0.0530, 0.0482, 0.0801, 0.0779, 0.0460], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0200, 0.0179, 0.0173, 0.0175, 0.0186, 0.0158, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:50:59,543 INFO [optim.py:369] (2/7) 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 00:51:03,843 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57244.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:51:05,616 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 10, batch 5700, loss[loss=0.1538, simple_loss=0.2133, pruned_loss=0.0471, over 4369.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2555, pruned_loss=0.06177, over 933481.51 frames. ], batch size: 19, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:51:12,250 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57258.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 00:51:15,767 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57264.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:51:38,656 INFO [finetune.py:976] (2/7) Epoch 11, batch 0, loss[loss=0.1806, simple_loss=0.2453, pruned_loss=0.05791, over 4919.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2453, pruned_loss=0.05791, over 4919.00 frames. ], batch size: 38, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:51:38,657 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 00:51:55,308 INFO [finetune.py:1010] (2/7) Epoch 11, validation: loss=0.1558, simple_loss=0.2272, pruned_loss=0.04225, over 2265189.00 frames. +2023-04-27 00:51:55,309 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 00:52:28,823 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57305.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:52:34,735 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7090, 1.2704, 1.2845, 1.4688, 1.9321, 1.5440, 1.2714, 1.2359], + device='cuda:2'), covar=tensor([0.1582, 0.1512, 0.1976, 0.1303, 0.0861, 0.1609, 0.2080, 0.2108], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0321, 0.0356, 0.0300, 0.0335, 0.0320, 0.0307, 0.0362], + device='cuda:2'), out_proj_covar=tensor([6.4488e-05, 6.8053e-05, 7.6719e-05, 6.1878e-05, 7.0120e-05, 6.8374e-05, + 6.5712e-05, 7.7799e-05], device='cuda:2') +2023-04-27 00:52:42,961 INFO [finetune.py:976] (2/7) Epoch 11, batch 50, loss[loss=0.1568, simple_loss=0.2291, pruned_loss=0.04228, over 4750.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2634, pruned_loss=0.06472, over 217110.18 frames. ], batch size: 28, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:52:49,891 INFO [optim.py:369] (2/7) 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 00:52:57,810 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57349.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:52:59,010 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57351.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:53:20,600 INFO [finetune.py:976] (2/7) Epoch 11, batch 100, loss[loss=0.1707, simple_loss=0.2419, pruned_loss=0.04977, over 4799.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2586, pruned_loss=0.06324, over 381472.85 frames. ], batch size: 45, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:53:34,834 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8759, 1.7774, 1.9869, 2.2341, 2.3694, 1.8536, 1.4962, 2.0635], + device='cuda:2'), covar=tensor([0.0904, 0.1140, 0.0653, 0.0619, 0.0545, 0.0913, 0.0898, 0.0561], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0201, 0.0180, 0.0173, 0.0176, 0.0186, 0.0158, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:53:40,241 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57398.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:54:10,095 INFO [finetune.py:976] (2/7) Epoch 11, batch 150, loss[loss=0.2033, simple_loss=0.2632, pruned_loss=0.07176, over 4890.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2529, pruned_loss=0.06139, over 509294.10 frames. ], batch size: 35, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:54:27,169 INFO [optim.py:369] (2/7) 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 00:54:29,607 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8208, 2.8225, 2.2327, 3.2597, 2.7625, 2.8368, 1.2805, 2.7842], + device='cuda:2'), covar=tensor([0.2266, 0.1817, 0.3649, 0.3301, 0.3600, 0.2262, 0.5088, 0.2773], + device='cuda:2'), in_proj_covar=tensor([0.0241, 0.0214, 0.0247, 0.0302, 0.0296, 0.0248, 0.0266, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:54:37,557 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 00:54:52,719 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:54:52,743 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:00,033 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:02,965 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6302, 1.5575, 4.5468, 4.2373, 3.9508, 4.2969, 4.2179, 4.0267], + device='cuda:2'), covar=tensor([0.6859, 0.5871, 0.1012, 0.1742, 0.1202, 0.1687, 0.1287, 0.1643], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0307, 0.0406, 0.0408, 0.0350, 0.0406, 0.0315, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:55:02,996 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8022, 1.8679, 0.7800, 1.4376, 1.9483, 1.6379, 1.4773, 1.5455], + device='cuda:2'), covar=tensor([0.0485, 0.0396, 0.0389, 0.0579, 0.0278, 0.0537, 0.0545, 0.0581], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 00:55:10,526 INFO [finetune.py:976] (2/7) Epoch 11, batch 200, loss[loss=0.2041, simple_loss=0.2702, pruned_loss=0.06899, over 4902.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2514, pruned_loss=0.06092, over 608966.58 frames. ], batch size: 36, lr: 3.72e-03, grad_scale: 32.0 +2023-04-27 00:55:18,962 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4166, 1.6280, 1.7716, 1.8953, 1.7495, 1.9323, 1.8588, 1.8325], + device='cuda:2'), covar=tensor([0.3989, 0.6092, 0.5142, 0.5037, 0.6030, 0.8361, 0.5722, 0.5857], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0379, 0.0315, 0.0326, 0.0338, 0.0402, 0.0360, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:55:20,208 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-04-27 00:55:22,549 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:30,375 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57507.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:40,654 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57523.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:55:42,939 INFO [finetune.py:976] (2/7) Epoch 11, batch 250, loss[loss=0.1738, simple_loss=0.2561, pruned_loss=0.04573, over 4731.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2517, pruned_loss=0.06033, over 685311.69 frames. ], batch size: 59, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 00:55:50,073 INFO [optim.py:369] (2/7) 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 00:55:52,339 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.6202, 4.5491, 3.1874, 5.3084, 4.6092, 4.6025, 2.2386, 4.4739], + device='cuda:2'), covar=tensor([0.1403, 0.1064, 0.3124, 0.0936, 0.3543, 0.1628, 0.5416, 0.2061], + device='cuda:2'), in_proj_covar=tensor([0.0241, 0.0214, 0.0248, 0.0301, 0.0296, 0.0248, 0.0267, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:55:54,616 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57553.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 00:56:03,126 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57556.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:56:04,900 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0569, 1.8594, 2.2678, 2.5324, 2.1685, 1.9796, 2.1108, 2.0695], + device='cuda:2'), covar=tensor([0.6003, 0.8606, 0.8599, 0.7644, 0.7994, 1.0636, 1.0750, 1.0373], + device='cuda:2'), in_proj_covar=tensor([0.0407, 0.0412, 0.0497, 0.0516, 0.0437, 0.0457, 0.0467, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9207e-05, 1.0217e-04, 1.1214e-04, 1.2278e-04, 1.0577e-04, 1.1042e-04, + 1.1193e-04, 1.1199e-04], device='cuda:2') +2023-04-27 00:56:07,263 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4363, 1.5271, 3.5469, 3.2665, 3.1716, 3.4324, 3.4290, 3.2163], + device='cuda:2'), covar=tensor([0.6828, 0.5263, 0.1349, 0.2137, 0.1470, 0.2042, 0.1477, 0.1520], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0305, 0.0405, 0.0408, 0.0349, 0.0405, 0.0314, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:56:08,500 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 300, loss[loss=0.1383, simple_loss=0.2056, pruned_loss=0.03549, over 4261.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2565, pruned_loss=0.06189, over 745971.71 frames. ], batch size: 18, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 00:56:17,558 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-27 00:56:27,335 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4374, 3.3544, 2.5820, 4.0175, 3.3782, 3.4466, 1.5091, 3.3768], + device='cuda:2'), covar=tensor([0.1737, 0.1375, 0.3749, 0.1901, 0.2832, 0.1853, 0.5676, 0.2676], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0215, 0.0249, 0.0303, 0.0297, 0.0249, 0.0268, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 00:56:32,857 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57600.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:56:34,197 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 00:56:40,097 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 350, loss[loss=0.1645, simple_loss=0.244, pruned_loss=0.04251, over 4773.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.26, pruned_loss=0.06278, over 793792.81 frames. ], batch size: 28, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 00:56:56,646 INFO [optim.py:369] (2/7) 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 00:57:11,402 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57649.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:57:13,071 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57651.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:57:46,789 INFO [finetune.py:976] (2/7) Epoch 11, batch 400, loss[loss=0.2178, simple_loss=0.2741, pruned_loss=0.08082, over 4881.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.261, pruned_loss=0.06285, over 829366.89 frames. ], batch size: 32, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 00:58:06,589 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0374, 3.0353, 2.3778, 2.5907, 2.1423, 2.5734, 2.5488, 1.9288], + device='cuda:2'), covar=tensor([0.2375, 0.1145, 0.0786, 0.1256, 0.3113, 0.1016, 0.2055, 0.2593], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0315, 0.0228, 0.0285, 0.0314, 0.0267, 0.0253, 0.0276], + device='cuda:2'), out_proj_covar=tensor([1.1856e-04, 1.2659e-04, 9.1447e-05, 1.1417e-04, 1.2835e-04, 1.0747e-04, + 1.0324e-04, 1.1079e-04], device='cuda:2') +2023-04-27 00:58:10,657 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57699.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:58:27,243 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8437, 1.5601, 1.7982, 2.0389, 1.7830, 1.4803, 1.1877, 1.5796], + device='cuda:2'), covar=tensor([0.2583, 0.2655, 0.1240, 0.1775, 0.2246, 0.2171, 0.3576, 0.1877], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0248, 0.0220, 0.0317, 0.0213, 0.0227, 0.0232, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:58:30,755 INFO [finetune.py:976] (2/7) Epoch 11, batch 450, loss[loss=0.1926, simple_loss=0.2698, pruned_loss=0.05773, over 4932.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2589, pruned_loss=0.06198, over 854697.35 frames. ], batch size: 42, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 00:58:33,254 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 00:58:34,985 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57733.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:58:37,317 INFO [optim.py:369] (2/7) 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 00:59:00,417 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57754.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:59:32,061 INFO [finetune.py:976] (2/7) Epoch 11, batch 500, loss[loss=0.1809, simple_loss=0.2371, pruned_loss=0.0623, over 4831.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2564, pruned_loss=0.06118, over 878364.70 frames. ], batch size: 33, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 00:59:36,343 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1082, 1.5973, 2.0122, 2.3393, 1.9528, 1.5749, 1.2069, 1.7150], + device='cuda:2'), covar=tensor([0.3808, 0.3836, 0.1950, 0.3007, 0.2946, 0.3081, 0.5259, 0.2629], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0250, 0.0222, 0.0319, 0.0214, 0.0229, 0.0234, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 00:59:48,111 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0884, 1.8536, 1.9941, 2.3769, 2.4573, 2.0007, 1.7453, 2.1369], + device='cuda:2'), covar=tensor([0.0709, 0.0961, 0.0610, 0.0451, 0.0471, 0.0746, 0.0751, 0.0480], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0200, 0.0180, 0.0172, 0.0175, 0.0185, 0.0157, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 00:59:54,254 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57794.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 00:59:59,390 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6197, 1.3704, 1.7418, 1.7871, 1.4232, 1.2795, 1.4476, 0.9994], + device='cuda:2'), covar=tensor([0.0520, 0.0873, 0.0514, 0.0677, 0.0790, 0.1241, 0.0741, 0.0739], + device='cuda:2'), in_proj_covar=tensor([0.0065, 0.0071, 0.0069, 0.0066, 0.0074, 0.0094, 0.0075, 0.0071], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 01:00:25,875 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57818.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:00:31,730 INFO [finetune.py:976] (2/7) Epoch 11, batch 550, loss[loss=0.2082, simple_loss=0.2551, pruned_loss=0.08068, over 4870.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2545, pruned_loss=0.06098, over 895982.43 frames. ], batch size: 31, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:00:38,241 INFO [optim.py:369] (2/7) 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 01:00:41,339 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57842.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:00:47,290 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57851.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:00:49,044 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57853.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 01:00:52,588 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6341, 3.5082, 2.6496, 4.2739, 3.6474, 3.6821, 1.6992, 3.5954], + device='cuda:2'), covar=tensor([0.2011, 0.1454, 0.3031, 0.1914, 0.3197, 0.2125, 0.5891, 0.2361], + device='cuda:2'), in_proj_covar=tensor([0.0241, 0.0215, 0.0248, 0.0301, 0.0297, 0.0249, 0.0268, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 01:01:04,583 INFO [finetune.py:976] (2/7) Epoch 11, batch 600, loss[loss=0.166, simple_loss=0.2266, pruned_loss=0.05271, over 4826.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2541, pruned_loss=0.06111, over 909437.45 frames. ], batch size: 30, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:01:13,412 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57890.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:01:20,161 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57900.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:01:20,736 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57901.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:01:38,048 INFO [finetune.py:976] (2/7) Epoch 11, batch 650, loss[loss=0.1881, simple_loss=0.2622, pruned_loss=0.05696, over 4838.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2578, pruned_loss=0.06228, over 919683.45 frames. ], batch size: 49, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:01:45,068 INFO [optim.py:369] (2/7) 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 01:01:51,799 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57948.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:01:53,668 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7837, 2.4265, 1.8488, 1.7245, 1.3013, 1.3025, 1.8558, 1.2883], + device='cuda:2'), covar=tensor([0.1555, 0.1351, 0.1415, 0.1767, 0.2274, 0.1929, 0.0972, 0.2011], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0215, 0.0170, 0.0204, 0.0204, 0.0185, 0.0159, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 01:02:11,759 INFO [finetune.py:976] (2/7) Epoch 11, batch 700, loss[loss=0.178, simple_loss=0.249, pruned_loss=0.05345, over 4771.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2593, pruned_loss=0.06253, over 927651.61 frames. ], batch size: 28, lr: 3.71e-03, grad_scale: 64.0 +2023-04-27 01:02:45,341 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9410, 2.7758, 1.7830, 2.0883, 1.4137, 1.3578, 1.8124, 1.3807], + device='cuda:2'), covar=tensor([0.1954, 0.1551, 0.1916, 0.1885, 0.2578, 0.2368, 0.1272, 0.2210], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0213, 0.0168, 0.0202, 0.0203, 0.0184, 0.0158, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 01:02:53,354 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2335, 4.1486, 3.1042, 4.9264, 4.1302, 4.1436, 1.8589, 4.2034], + device='cuda:2'), covar=tensor([0.1474, 0.1031, 0.3437, 0.0911, 0.2928, 0.1537, 0.5775, 0.1914], + device='cuda:2'), in_proj_covar=tensor([0.0240, 0.0214, 0.0248, 0.0300, 0.0295, 0.0248, 0.0267, 0.0268], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 01:02:56,395 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0525, 2.4426, 1.0069, 1.1759, 1.8849, 1.1657, 3.0504, 1.6227], + device='cuda:2'), covar=tensor([0.0691, 0.0535, 0.0760, 0.1479, 0.0505, 0.1130, 0.0358, 0.0703], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0078, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 01:02:57,541 INFO [finetune.py:976] (2/7) Epoch 11, batch 750, loss[loss=0.1975, simple_loss=0.2586, pruned_loss=0.06822, over 4911.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2605, pruned_loss=0.06312, over 933170.39 frames. ], batch size: 33, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:03:04,227 INFO [optim.py:369] (2/7) 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 01:03:14,990 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58054.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:03:31,278 INFO [finetune.py:976] (2/7) Epoch 11, batch 800, loss[loss=0.1633, simple_loss=0.2361, pruned_loss=0.04527, over 4823.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2596, pruned_loss=0.06164, over 939418.76 frames. ], batch size: 30, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:03:31,991 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9474, 0.8564, 1.0866, 1.0231, 0.9020, 0.7945, 0.8553, 0.4390], + device='cuda:2'), covar=tensor([0.0549, 0.0569, 0.0601, 0.0458, 0.0655, 0.1063, 0.0515, 0.0878], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0072, 0.0070, 0.0067, 0.0075, 0.0095, 0.0076, 0.0072], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 01:03:38,600 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58089.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:03:47,435 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58102.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:03:58,174 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58118.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:04,031 INFO [finetune.py:976] (2/7) Epoch 11, batch 850, loss[loss=0.1892, simple_loss=0.2485, pruned_loss=0.06493, over 4848.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2576, pruned_loss=0.06081, over 940532.56 frames. ], batch size: 47, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:04:10,688 INFO [optim.py:369] (2/7) 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 01:04:12,618 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58141.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:14,897 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:19,560 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58151.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:29,488 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6803, 2.1175, 1.5897, 1.3594, 1.2653, 1.2949, 1.5392, 1.2086], + device='cuda:2'), covar=tensor([0.1905, 0.1283, 0.1803, 0.2094, 0.2682, 0.2294, 0.1287, 0.2319], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0213, 0.0168, 0.0202, 0.0202, 0.0184, 0.0158, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 01:04:39,246 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:04:52,831 INFO [finetune.py:976] (2/7) Epoch 11, batch 900, loss[loss=0.2094, simple_loss=0.2743, pruned_loss=0.07228, over 4825.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2546, pruned_loss=0.0599, over 943587.05 frames. ], batch size: 40, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:05:19,788 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58199.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:05:22,296 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58202.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:05:28,580 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58205.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:05:31,793 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 01:05:54,759 INFO [finetune.py:976] (2/7) Epoch 11, batch 950, loss[loss=0.1685, simple_loss=0.2267, pruned_loss=0.05518, over 4802.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2528, pruned_loss=0.05957, over 945553.67 frames. ], batch size: 25, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:06:10,244 INFO [optim.py:369] (2/7) 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 01:06:12,964 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-04-27 01:06:43,328 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1954, 1.7768, 2.1382, 2.6887, 2.0810, 1.6426, 1.6239, 2.0658], + device='cuda:2'), covar=tensor([0.4136, 0.3837, 0.1981, 0.3187, 0.3435, 0.3240, 0.4659, 0.2725], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0249, 0.0221, 0.0317, 0.0213, 0.0227, 0.0232, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 01:06:46,847 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1905, 2.8938, 0.9563, 1.3801, 1.9596, 1.2423, 3.7875, 1.7075], + device='cuda:2'), covar=tensor([0.0697, 0.0922, 0.0978, 0.1219, 0.0556, 0.1089, 0.0232, 0.0664], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0051, 0.0053, 0.0078, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 01:06:47,387 INFO [finetune.py:976] (2/7) Epoch 11, batch 1000, loss[loss=0.2022, simple_loss=0.2734, pruned_loss=0.06549, over 4750.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2558, pruned_loss=0.06091, over 947242.00 frames. ], batch size: 54, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:07:02,934 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5429, 3.2200, 1.3568, 2.0386, 1.9482, 2.5900, 2.0523, 1.2910], + device='cuda:2'), covar=tensor([0.1137, 0.0832, 0.1405, 0.1032, 0.0922, 0.0788, 0.1294, 0.1811], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0248, 0.0139, 0.0122, 0.0133, 0.0153, 0.0118, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 01:07:20,925 INFO [finetune.py:976] (2/7) Epoch 11, batch 1050, loss[loss=0.2507, simple_loss=0.3149, pruned_loss=0.09326, over 4800.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2595, pruned_loss=0.06188, over 950010.68 frames. ], batch size: 51, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:07:28,211 INFO [optim.py:369] (2/7) 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 01:07:33,852 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7262, 2.2682, 1.7858, 2.0654, 1.6749, 1.8226, 1.8687, 1.5293], + device='cuda:2'), covar=tensor([0.2008, 0.1346, 0.0975, 0.1286, 0.3202, 0.1332, 0.1882, 0.2402], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0316, 0.0229, 0.0285, 0.0314, 0.0270, 0.0254, 0.0276], + device='cuda:2'), out_proj_covar=tensor([1.1902e-04, 1.2719e-04, 9.1677e-05, 1.1425e-04, 1.2868e-04, 1.0843e-04, + 1.0367e-04, 1.1075e-04], device='cuda:2') +2023-04-27 01:07:42,970 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-04-27 01:07:44,204 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-27 01:07:52,945 INFO [finetune.py:976] (2/7) Epoch 11, batch 1100, loss[loss=0.1866, simple_loss=0.258, pruned_loss=0.05763, over 4796.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2594, pruned_loss=0.06154, over 952284.64 frames. ], batch size: 29, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:08:01,710 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58389.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:08:18,820 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7763, 2.4382, 1.6978, 1.6841, 1.3006, 1.2863, 1.7239, 1.2201], + device='cuda:2'), covar=tensor([0.1590, 0.1289, 0.1523, 0.1781, 0.2371, 0.1958, 0.1057, 0.2038], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0214, 0.0169, 0.0204, 0.0203, 0.0184, 0.0158, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 01:08:26,953 INFO [finetune.py:976] (2/7) Epoch 11, batch 1150, loss[loss=0.1718, simple_loss=0.2426, pruned_loss=0.05053, over 4728.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2607, pruned_loss=0.0619, over 953457.74 frames. ], batch size: 59, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:08:34,047 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58437.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:08:35,083 INFO [optim.py:369] (2/7) 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 01:08:55,091 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2158, 2.4998, 1.0329, 1.3396, 1.9786, 1.2591, 3.2800, 1.7557], + device='cuda:2'), covar=tensor([0.0613, 0.0653, 0.0796, 0.1285, 0.0488, 0.1053, 0.0298, 0.0624], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0051, 0.0053, 0.0078, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 01:08:59,898 INFO [finetune.py:976] (2/7) Epoch 11, batch 1200, loss[loss=0.1902, simple_loss=0.2606, pruned_loss=0.05996, over 4907.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2584, pruned_loss=0.06111, over 952333.52 frames. ], batch size: 36, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:09:00,029 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9316, 2.6055, 1.9445, 1.9530, 1.5574, 1.4488, 2.0419, 1.4763], + device='cuda:2'), covar=tensor([0.1419, 0.1216, 0.1395, 0.1666, 0.2081, 0.1781, 0.0928, 0.1810], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0214, 0.0169, 0.0203, 0.0203, 0.0184, 0.0158, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 01:09:08,000 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2036, 1.6697, 2.0367, 2.4602, 2.0033, 1.5838, 1.2693, 1.7599], + device='cuda:2'), covar=tensor([0.3349, 0.3534, 0.1628, 0.2376, 0.2733, 0.2920, 0.4912, 0.2405], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0248, 0.0221, 0.0317, 0.0213, 0.0227, 0.0233, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 01:09:13,881 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58497.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:09:15,684 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58500.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:09:29,403 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3866, 1.3437, 1.4521, 0.9450, 1.2911, 1.1431, 1.6884, 1.2262], + device='cuda:2'), covar=tensor([0.3170, 0.1544, 0.4222, 0.2284, 0.1447, 0.2051, 0.1517, 0.4663], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0346, 0.0426, 0.0358, 0.0385, 0.0381, 0.0379, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 01:09:32,952 INFO [finetune.py:976] (2/7) Epoch 11, batch 1250, loss[loss=0.1389, simple_loss=0.2211, pruned_loss=0.02835, over 4935.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2549, pruned_loss=0.06001, over 952849.28 frames. ], batch size: 33, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:09:41,125 INFO [optim.py:369] (2/7) 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 01:10:09,823 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7400, 1.3632, 1.8067, 2.2330, 1.9045, 1.6729, 1.7903, 1.7695], + device='cuda:2'), covar=tensor([0.5602, 0.7773, 0.7957, 0.7342, 0.6488, 0.9738, 0.9317, 0.9444], + device='cuda:2'), in_proj_covar=tensor([0.0406, 0.0410, 0.0496, 0.0515, 0.0437, 0.0455, 0.0465, 0.0465], + device='cuda:2'), out_proj_covar=tensor([9.8817e-05, 1.0169e-04, 1.1178e-04, 1.2235e-04, 1.0562e-04, 1.0997e-04, + 1.1137e-04, 1.1162e-04], device='cuda:2') +2023-04-27 01:10:22,918 INFO [finetune.py:976] (2/7) Epoch 11, batch 1300, loss[loss=0.1617, simple_loss=0.2352, pruned_loss=0.04412, over 4816.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2524, pruned_loss=0.05967, over 955410.66 frames. ], batch size: 45, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:11:29,578 INFO [finetune.py:976] (2/7) Epoch 11, batch 1350, loss[loss=0.2292, simple_loss=0.3035, pruned_loss=0.07741, over 4904.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2545, pruned_loss=0.06098, over 954765.68 frames. ], batch size: 43, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:11:46,674 INFO [optim.py:369] (2/7) 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 01:12:32,200 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 01:12:34,429 INFO [finetune.py:976] (2/7) Epoch 11, batch 1400, loss[loss=0.201, simple_loss=0.2705, pruned_loss=0.06581, over 4818.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2565, pruned_loss=0.06145, over 955222.27 frames. ], batch size: 39, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:12:42,449 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58681.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:13:42,637 INFO [finetune.py:976] (2/7) Epoch 11, batch 1450, loss[loss=0.2057, simple_loss=0.284, pruned_loss=0.06372, over 4922.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2594, pruned_loss=0.06245, over 956724.02 frames. ], batch size: 42, lr: 3.71e-03, grad_scale: 32.0 +2023-04-27 01:14:01,639 INFO [optim.py:369] (2/7) 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 01:14:01,746 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9587, 1.6353, 1.8740, 2.3494, 2.3448, 1.9013, 1.5317, 2.1152], + device='cuda:2'), covar=tensor([0.0833, 0.1110, 0.0741, 0.0518, 0.0599, 0.0846, 0.0904, 0.0529], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0200, 0.0180, 0.0173, 0.0176, 0.0186, 0.0158, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 01:14:05,184 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58742.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:14:37,126 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 1500, loss[loss=0.1948, simple_loss=0.2679, pruned_loss=0.06088, over 4881.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2607, pruned_loss=0.0627, over 954258.48 frames. ], batch size: 35, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:14:52,498 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:14:54,829 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58800.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:15:02,083 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5582, 0.9372, 1.5399, 2.0497, 1.6586, 1.5012, 1.5635, 1.6110], + device='cuda:2'), covar=tensor([0.5741, 0.7908, 0.7696, 0.7508, 0.7194, 0.9351, 0.9085, 0.8682], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0413, 0.0500, 0.0519, 0.0441, 0.0460, 0.0469, 0.0468], + device='cuda:2'), out_proj_covar=tensor([9.9549e-05, 1.0249e-04, 1.1280e-04, 1.2329e-04, 1.0654e-04, 1.1114e-04, + 1.1242e-04, 1.1245e-04], device='cuda:2') +2023-04-27 01:15:11,710 INFO [finetune.py:976] (2/7) Epoch 11, batch 1550, loss[loss=0.1994, simple_loss=0.2616, pruned_loss=0.06862, over 4751.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.26, pruned_loss=0.06248, over 953561.18 frames. ], batch size: 54, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:15:15,187 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3331, 2.9288, 2.3136, 2.6813, 2.1491, 2.4199, 2.4396, 2.0215], + device='cuda:2'), covar=tensor([0.1904, 0.1150, 0.0901, 0.1164, 0.2803, 0.1390, 0.1836, 0.2595], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0318, 0.0230, 0.0286, 0.0316, 0.0271, 0.0256, 0.0277], + device='cuda:2'), out_proj_covar=tensor([1.1946e-04, 1.2774e-04, 9.2286e-05, 1.1461e-04, 1.2924e-04, 1.0909e-04, + 1.0414e-04, 1.1123e-04], device='cuda:2') +2023-04-27 01:15:15,935 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-04-27 01:15:18,197 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58836.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:15:19,336 INFO [optim.py:369] (2/7) 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 01:15:24,174 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:15:29,434 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1298, 1.9930, 1.7016, 1.5661, 1.9397, 1.6067, 2.3219, 1.4203], + device='cuda:2'), covar=tensor([0.2931, 0.1259, 0.3426, 0.2394, 0.1429, 0.2040, 0.1256, 0.4009], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0344, 0.0423, 0.0356, 0.0382, 0.0379, 0.0376, 0.0415], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 01:15:44,971 INFO [finetune.py:976] (2/7) Epoch 11, batch 1600, loss[loss=0.1568, simple_loss=0.2219, pruned_loss=0.04579, over 4745.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2576, pruned_loss=0.06178, over 953331.67 frames. ], batch size: 59, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:15:56,406 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-04-27 01:16:05,012 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58906.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:16:18,094 INFO [finetune.py:976] (2/7) Epoch 11, batch 1650, loss[loss=0.1617, simple_loss=0.2291, pruned_loss=0.04711, over 4870.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2546, pruned_loss=0.06076, over 950992.55 frames. ], batch size: 31, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:16:25,722 INFO [optim.py:369] (2/7) 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 01:16:58,088 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1204, 2.6219, 1.0517, 1.2804, 2.0621, 1.1393, 3.7081, 1.6429], + device='cuda:2'), covar=tensor([0.0684, 0.0723, 0.0825, 0.1330, 0.0546, 0.1130, 0.0252, 0.0701], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0049, 0.0047, 0.0051, 0.0053, 0.0078, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0011, 0.0008], + device='cuda:2') +2023-04-27 01:17:07,489 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58967.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:17:13,433 INFO [finetune.py:976] (2/7) Epoch 11, batch 1700, loss[loss=0.1794, simple_loss=0.2294, pruned_loss=0.06473, over 3946.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2524, pruned_loss=0.06015, over 953615.89 frames. ], batch size: 17, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:17:47,006 INFO [finetune.py:976] (2/7) Epoch 11, batch 1750, loss[loss=0.1662, simple_loss=0.2388, pruned_loss=0.04681, over 4754.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2535, pruned_loss=0.06052, over 952582.85 frames. ], batch size: 27, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:17:53,222 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59037.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:17:53,741 INFO [optim.py:369] (2/7) 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 01:17:53,867 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3988, 1.0125, 0.4098, 1.1080, 1.0324, 1.2744, 1.1765, 1.1576], + device='cuda:2'), covar=tensor([0.0566, 0.0456, 0.0436, 0.0624, 0.0326, 0.0576, 0.0560, 0.0632], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 01:17:56,692 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2822, 2.1768, 2.3535, 2.7884, 2.7694, 2.2112, 1.6856, 2.4993], + device='cuda:2'), covar=tensor([0.0977, 0.0995, 0.0754, 0.0617, 0.0617, 0.0962, 0.0985, 0.0569], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0204, 0.0183, 0.0176, 0.0179, 0.0189, 0.0161, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 01:18:30,832 INFO [finetune.py:976] (2/7) Epoch 11, batch 1800, loss[loss=0.2092, simple_loss=0.2744, pruned_loss=0.072, over 4753.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2541, pruned_loss=0.06103, over 948732.05 frames. ], batch size: 27, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:18:42,234 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7992, 1.8940, 1.7455, 2.3602, 2.1845, 2.3462, 1.8417, 4.6714], + device='cuda:2'), covar=tensor([0.0511, 0.0768, 0.0832, 0.1097, 0.0582, 0.0486, 0.0704, 0.0089], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 01:19:09,377 INFO [finetune.py:976] (2/7) Epoch 11, batch 1850, loss[loss=0.1872, simple_loss=0.2528, pruned_loss=0.06081, over 4781.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2557, pruned_loss=0.06166, over 948440.31 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:19:11,899 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59131.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:19:21,280 INFO [optim.py:369] (2/7) 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 01:19:52,228 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59164.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:20:10,992 INFO [finetune.py:976] (2/7) Epoch 11, batch 1900, loss[loss=0.2064, simple_loss=0.2691, pruned_loss=0.07184, over 4813.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2581, pruned_loss=0.06252, over 949052.22 frames. ], batch size: 38, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:20:21,664 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1325, 3.1640, 2.4287, 2.6992, 2.1746, 2.6182, 2.6773, 1.6051], + device='cuda:2'), covar=tensor([0.2784, 0.1293, 0.0863, 0.1553, 0.3131, 0.1210, 0.2233, 0.3172], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0314, 0.0229, 0.0285, 0.0314, 0.0270, 0.0254, 0.0275], + device='cuda:2'), out_proj_covar=tensor([1.1911e-04, 1.2628e-04, 9.1613e-05, 1.1399e-04, 1.2839e-04, 1.0841e-04, + 1.0362e-04, 1.1031e-04], device='cuda:2') +2023-04-27 01:21:00,115 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 1950, loss[loss=0.1894, simple_loss=0.2559, pruned_loss=0.0615, over 4886.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.257, pruned_loss=0.06166, over 949305.80 frames. ], batch size: 32, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:21:08,358 INFO [optim.py:369] (2/7) 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 01:21:24,550 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 2000, loss[loss=0.1621, simple_loss=0.2213, pruned_loss=0.05143, over 4821.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2537, pruned_loss=0.05974, over 952772.09 frames. ], batch size: 39, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:22:08,861 INFO [finetune.py:976] (2/7) Epoch 11, batch 2050, loss[loss=0.154, simple_loss=0.2253, pruned_loss=0.04138, over 4819.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2514, pruned_loss=0.05913, over 951752.29 frames. ], batch size: 25, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:22:15,378 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 01:22:48,348 INFO [finetune.py:976] (2/7) Epoch 11, batch 2100, loss[loss=0.1623, simple_loss=0.2288, pruned_loss=0.04792, over 4765.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2529, pruned_loss=0.06011, over 951611.74 frames. ], batch size: 27, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:22:48,650 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.10 vs. limit=5.0 +2023-04-27 01:22:53,374 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:23:12,106 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3423, 1.6367, 1.4488, 2.0215, 1.8967, 1.9645, 1.5784, 4.3638], + device='cuda:2'), covar=tensor([0.0630, 0.0827, 0.0858, 0.1232, 0.0628, 0.0606, 0.0778, 0.0116], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 01:23:25,578 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6316, 1.3897, 1.7510, 1.9635, 1.7648, 1.5592, 1.6429, 1.6536], + device='cuda:2'), covar=tensor([0.6415, 0.8609, 0.9660, 1.0011, 0.7639, 1.2090, 1.1674, 1.1443], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0409, 0.0498, 0.0515, 0.0437, 0.0457, 0.0465, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9089e-05, 1.0157e-04, 1.1228e-04, 1.2235e-04, 1.0591e-04, 1.1035e-04, + 1.1147e-04, 1.1200e-04], device='cuda:2') +2023-04-27 01:23:32,962 INFO [finetune.py:976] (2/7) Epoch 11, batch 2150, loss[loss=0.1682, simple_loss=0.2509, pruned_loss=0.04281, over 4786.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2569, pruned_loss=0.06128, over 952805.65 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:23:35,512 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59431.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:23:44,711 INFO [optim.py:369] (2/7) 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 01:23:54,703 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.45 vs. limit=5.0 +2023-04-27 01:24:15,853 INFO [finetune.py:976] (2/7) Epoch 11, batch 2200, loss[loss=0.2017, simple_loss=0.274, pruned_loss=0.06476, over 4895.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2602, pruned_loss=0.0625, over 953894.26 frames. ], batch size: 35, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:24:18,112 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59479.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:24:35,838 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4123, 1.7582, 1.7117, 1.8936, 1.7366, 1.8037, 1.8140, 1.7396], + device='cuda:2'), covar=tensor([0.4491, 0.6571, 0.5701, 0.5208, 0.6577, 0.8906, 0.6648, 0.6099], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0379, 0.0315, 0.0326, 0.0339, 0.0400, 0.0358, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 01:24:37,690 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59502.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:24:54,723 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:24:59,358 INFO [finetune.py:976] (2/7) Epoch 11, batch 2250, loss[loss=0.2178, simple_loss=0.2865, pruned_loss=0.07451, over 4824.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2613, pruned_loss=0.06324, over 954323.62 frames. ], batch size: 39, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:25:13,213 INFO [optim.py:369] (2/7) 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 01:25:45,851 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59562.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:25:46,508 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59563.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:25:55,720 INFO [finetune.py:976] (2/7) Epoch 11, batch 2300, loss[loss=0.1763, simple_loss=0.2401, pruned_loss=0.05629, over 4825.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2619, pruned_loss=0.06327, over 954124.20 frames. ], batch size: 30, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:26:17,899 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 2350, loss[loss=0.1346, simple_loss=0.2056, pruned_loss=0.03177, over 4839.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2581, pruned_loss=0.06158, over 954950.69 frames. ], batch size: 47, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:26:31,627 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.62 vs. limit=5.0 +2023-04-27 01:26:37,792 INFO [optim.py:369] (2/7) 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 01:27:00,876 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1396, 2.1406, 1.8769, 1.7741, 2.2350, 1.7037, 2.8177, 1.6670], + device='cuda:2'), covar=tensor([0.3815, 0.2012, 0.4596, 0.3302, 0.1771, 0.2970, 0.1488, 0.4475], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0348, 0.0427, 0.0359, 0.0384, 0.0383, 0.0381, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 01:27:02,575 INFO [finetune.py:976] (2/7) Epoch 11, batch 2400, loss[loss=0.1654, simple_loss=0.2288, pruned_loss=0.05101, over 4898.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.256, pruned_loss=0.0611, over 956974.61 frames. ], batch size: 32, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:27:36,064 INFO [finetune.py:976] (2/7) Epoch 11, batch 2450, loss[loss=0.1977, simple_loss=0.2499, pruned_loss=0.07281, over 4774.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2532, pruned_loss=0.06039, over 956607.92 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:27:43,829 INFO [optim.py:369] (2/7) 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 01:27:47,857 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8494, 1.7102, 1.9283, 2.2112, 2.2441, 1.8307, 1.4824, 2.0155], + device='cuda:2'), covar=tensor([0.0946, 0.1262, 0.0789, 0.0708, 0.0603, 0.0922, 0.0993, 0.0614], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0204, 0.0184, 0.0176, 0.0179, 0.0189, 0.0161, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 01:28:04,879 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 01:28:09,965 INFO [finetune.py:976] (2/7) Epoch 11, batch 2500, loss[loss=0.203, simple_loss=0.2787, pruned_loss=0.06361, over 4802.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2556, pruned_loss=0.06173, over 956131.63 frames. ], batch size: 41, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:28:32,781 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3399, 3.4327, 2.5724, 3.9281, 3.4044, 3.3615, 1.5906, 3.3674], + device='cuda:2'), covar=tensor([0.1811, 0.1363, 0.3060, 0.2323, 0.4690, 0.2120, 0.5676, 0.2672], + device='cuda:2'), in_proj_covar=tensor([0.0241, 0.0216, 0.0248, 0.0304, 0.0296, 0.0249, 0.0268, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 01:28:41,888 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 01:28:44,931 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 2550, loss[loss=0.1887, simple_loss=0.2602, pruned_loss=0.05866, over 4898.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2577, pruned_loss=0.06161, over 957193.19 frames. ], batch size: 43, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:29:12,510 INFO [optim.py:369] (2/7) 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 01:29:23,400 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:29:35,317 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59858.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:29:46,959 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 2600, loss[loss=0.168, simple_loss=0.242, pruned_loss=0.04705, over 4777.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2581, pruned_loss=0.06136, over 957038.38 frames. ], batch size: 29, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:30:10,824 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59889.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:30:22,127 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9901, 1.7826, 2.2409, 2.4867, 2.0196, 1.9011, 2.0206, 2.0262], + device='cuda:2'), covar=tensor([0.6293, 0.8618, 0.8849, 0.7919, 0.7658, 1.1173, 1.1921, 1.0694], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0410, 0.0497, 0.0517, 0.0439, 0.0457, 0.0466, 0.0467], + device='cuda:2'), out_proj_covar=tensor([9.9118e-05, 1.0185e-04, 1.1218e-04, 1.2272e-04, 1.0617e-04, 1.1056e-04, + 1.1183e-04, 1.1213e-04], device='cuda:2') +2023-04-27 01:30:29,463 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:30:32,469 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-04-27 01:30:51,533 INFO [finetune.py:976] (2/7) Epoch 11, batch 2650, loss[loss=0.176, simple_loss=0.2469, pruned_loss=0.05257, over 4819.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.259, pruned_loss=0.06175, over 956456.99 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:31:04,460 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:31:41,175 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 01:31:41,223 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 01:31:43,484 INFO [finetune.py:976] (2/7) Epoch 11, batch 2700, loss[loss=0.1757, simple_loss=0.2446, pruned_loss=0.05342, over 4785.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2574, pruned_loss=0.06078, over 954606.69 frames. ], batch size: 28, lr: 3.70e-03, grad_scale: 32.0 +2023-04-27 01:32:17,130 INFO [finetune.py:976] (2/7) Epoch 11, batch 2750, loss[loss=0.1966, simple_loss=0.2602, pruned_loss=0.06652, over 4871.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2556, pruned_loss=0.06113, over 955051.99 frames. ], batch size: 31, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:32:24,354 INFO [optim.py:369] (2/7) 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 01:32:25,706 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:32:33,977 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 11, batch 2800, loss[loss=0.156, simple_loss=0.22, pruned_loss=0.04597, over 4759.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2524, pruned_loss=0.06007, over 956893.32 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:32:51,103 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:32:58,432 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2028, 1.5965, 1.5732, 1.8306, 1.7721, 1.9955, 1.4060, 3.7120], + device='cuda:2'), covar=tensor([0.0648, 0.0762, 0.0772, 0.1179, 0.0585, 0.0500, 0.0750, 0.0144], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 01:33:05,645 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60101.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 01:33:11,495 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7625, 1.2838, 1.8144, 2.2142, 1.8906, 1.7582, 1.8256, 1.8263], + device='cuda:2'), covar=tensor([0.5031, 0.7386, 0.7327, 0.6880, 0.6457, 0.8719, 0.8705, 0.8316], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0411, 0.0499, 0.0518, 0.0439, 0.0458, 0.0468, 0.0468], + device='cuda:2'), out_proj_covar=tensor([9.9256e-05, 1.0193e-04, 1.1253e-04, 1.2300e-04, 1.0630e-04, 1.1080e-04, + 1.1220e-04, 1.1231e-04], device='cuda:2') +2023-04-27 01:33:14,372 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60113.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 01:33:22,796 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60126.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:33:23,283 INFO [finetune.py:976] (2/7) Epoch 11, batch 2850, loss[loss=0.2043, simple_loss=0.2795, pruned_loss=0.06448, over 4808.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2512, pruned_loss=0.06005, over 952680.68 frames. ], batch size: 51, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:33:29,986 INFO [optim.py:369] (2/7) 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 01:33:30,721 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60139.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:33:43,076 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:33:55,931 INFO [finetune.py:976] (2/7) Epoch 11, batch 2900, loss[loss=0.2287, simple_loss=0.2735, pruned_loss=0.09193, over 4761.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2527, pruned_loss=0.06044, over 954022.53 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 64.0 +2023-04-27 01:34:15,778 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60198.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:34:25,140 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:34:59,286 INFO [finetune.py:976] (2/7) Epoch 11, batch 2950, loss[loss=0.1935, simple_loss=0.2754, pruned_loss=0.05582, over 4889.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.255, pruned_loss=0.06105, over 953634.55 frames. ], batch size: 43, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:35:12,689 INFO [optim.py:369] (2/7) 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 01:35:21,700 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60245.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:35:36,405 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60259.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:35:42,783 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-04-27 01:36:04,298 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2496, 1.2564, 3.8189, 3.5226, 3.3969, 3.6552, 3.6871, 3.4144], + device='cuda:2'), covar=tensor([0.7158, 0.5783, 0.1198, 0.1996, 0.1285, 0.1686, 0.1665, 0.1524], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0308, 0.0405, 0.0408, 0.0350, 0.0408, 0.0316, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 01:36:05,414 INFO [finetune.py:976] (2/7) Epoch 11, batch 3000, loss[loss=0.1706, simple_loss=0.2411, pruned_loss=0.05008, over 4173.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2577, pruned_loss=0.06206, over 953605.44 frames. ], batch size: 18, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:36:05,414 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 01:36:27,793 INFO [finetune.py:1010] (2/7) Epoch 11, validation: loss=0.1531, simple_loss=0.2255, pruned_loss=0.04032, over 2265189.00 frames. +2023-04-27 01:36:27,794 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 01:37:00,432 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6746, 1.4878, 1.8248, 1.9338, 1.5088, 1.3033, 1.5569, 1.0400], + device='cuda:2'), covar=tensor([0.0551, 0.0819, 0.0487, 0.0610, 0.0729, 0.1216, 0.0614, 0.0755], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0071, 0.0070, 0.0067, 0.0075, 0.0096, 0.0076, 0.0071], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 01:37:31,377 INFO [finetune.py:976] (2/7) Epoch 11, batch 3050, loss[loss=0.1655, simple_loss=0.2357, pruned_loss=0.04764, over 4904.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2591, pruned_loss=0.06219, over 954581.25 frames. ], batch size: 37, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:37:45,693 INFO [optim.py:369] (2/7) 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 01:38:21,987 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 01:38:26,535 INFO [finetune.py:976] (2/7) Epoch 11, batch 3100, loss[loss=0.1923, simple_loss=0.2677, pruned_loss=0.05847, over 4816.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2562, pruned_loss=0.06091, over 954281.28 frames. ], batch size: 38, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:38:32,462 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3210, 3.3514, 0.9625, 1.6094, 1.7760, 2.3850, 1.8753, 1.0987], + device='cuda:2'), covar=tensor([0.1374, 0.0832, 0.1913, 0.1303, 0.1021, 0.0940, 0.1432, 0.1984], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0248, 0.0141, 0.0122, 0.0134, 0.0153, 0.0118, 0.0122], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 01:38:38,398 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 01:38:40,717 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60396.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 01:38:48,247 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60408.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 01:38:56,660 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60421.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:39:00,270 INFO [finetune.py:976] (2/7) Epoch 11, batch 3150, loss[loss=0.1848, simple_loss=0.2459, pruned_loss=0.06187, over 4835.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.254, pruned_loss=0.06015, over 954020.40 frames. ], batch size: 47, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:39:06,445 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60434.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:39:09,791 INFO [optim.py:369] (2/7) 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 01:39:33,959 INFO [finetune.py:976] (2/7) Epoch 11, batch 3200, loss[loss=0.1755, simple_loss=0.2422, pruned_loss=0.05442, over 4904.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2518, pruned_loss=0.05953, over 953862.15 frames. ], batch size: 36, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:39:53,405 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 3250, loss[loss=0.1498, simple_loss=0.2172, pruned_loss=0.04121, over 4792.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2518, pruned_loss=0.06004, over 951670.58 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:40:15,718 INFO [optim.py:369] (2/7) 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 01:40:20,946 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60545.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:40:25,156 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60552.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:40:26,845 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60554.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:40:52,025 INFO [finetune.py:976] (2/7) Epoch 11, batch 3300, loss[loss=0.2242, simple_loss=0.2985, pruned_loss=0.07497, over 4818.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2562, pruned_loss=0.06138, over 952518.22 frames. ], batch size: 38, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:40:53,959 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2986, 1.1907, 1.5371, 1.4566, 1.1654, 1.0577, 1.2840, 0.9016], + device='cuda:2'), covar=tensor([0.0648, 0.0776, 0.0459, 0.0604, 0.0893, 0.1364, 0.0660, 0.0730], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0072, 0.0070, 0.0067, 0.0075, 0.0096, 0.0076, 0.0071], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 01:41:14,048 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60593.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:41:58,102 INFO [finetune.py:976] (2/7) Epoch 11, batch 3350, loss[loss=0.209, simple_loss=0.2828, pruned_loss=0.06761, over 4904.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2589, pruned_loss=0.06255, over 952852.19 frames. ], batch size: 36, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:42:08,506 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 01:42:11,849 INFO [optim.py:369] (2/7) 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 01:42:43,722 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6593, 1.4122, 1.8039, 2.1242, 1.7927, 1.6567, 1.7498, 1.7111], + device='cuda:2'), covar=tensor([0.5501, 0.7582, 0.7998, 0.7425, 0.6718, 0.9130, 0.9216, 0.8953], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0409, 0.0496, 0.0515, 0.0438, 0.0456, 0.0465, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9085e-05, 1.0148e-04, 1.1197e-04, 1.2217e-04, 1.0608e-04, 1.1039e-04, + 1.1154e-04, 1.1185e-04], device='cuda:2') +2023-04-27 01:42:52,677 INFO [finetune.py:976] (2/7) Epoch 11, batch 3400, loss[loss=0.1925, simple_loss=0.2564, pruned_loss=0.06435, over 4922.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2597, pruned_loss=0.06255, over 954033.93 frames. ], batch size: 33, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:42:52,835 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4766, 1.8571, 2.3231, 2.9494, 2.2344, 1.7893, 1.7140, 2.2214], + device='cuda:2'), covar=tensor([0.3556, 0.3601, 0.1793, 0.2871, 0.3389, 0.2894, 0.4243, 0.2468], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0246, 0.0219, 0.0313, 0.0212, 0.0226, 0.0229, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 01:43:05,430 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60696.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:14,148 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60708.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:20,717 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60721.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:26,000 INFO [finetune.py:976] (2/7) Epoch 11, batch 3450, loss[loss=0.2004, simple_loss=0.2623, pruned_loss=0.06929, over 4814.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2592, pruned_loss=0.06223, over 954222.41 frames. ], batch size: 38, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:43:26,770 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6318, 1.2258, 1.7247, 2.1645, 1.7940, 1.5728, 1.6819, 1.6290], + device='cuda:2'), covar=tensor([0.5084, 0.7393, 0.7057, 0.6568, 0.6243, 0.8325, 0.8279, 0.9273], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0410, 0.0497, 0.0515, 0.0438, 0.0458, 0.0467, 0.0466], + device='cuda:2'), out_proj_covar=tensor([9.9156e-05, 1.0168e-04, 1.1203e-04, 1.2226e-04, 1.0616e-04, 1.1067e-04, + 1.1183e-04, 1.1190e-04], device='cuda:2') +2023-04-27 01:43:30,778 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60734.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:33,697 INFO [optim.py:369] (2/7) 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 01:43:34,482 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-04-27 01:43:36,746 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60744.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:45,937 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60756.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:43:54,318 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 3500, loss[loss=0.1588, simple_loss=0.2282, pruned_loss=0.04468, over 4739.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2569, pruned_loss=0.0614, over 954793.53 frames. ], batch size: 23, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:44:00,468 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60779.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:44:02,193 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60782.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:44:30,482 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6227, 3.0463, 0.9303, 1.7449, 2.1327, 1.6055, 4.3971, 2.2787], + device='cuda:2'), covar=tensor([0.0579, 0.0782, 0.0873, 0.1274, 0.0563, 0.0964, 0.0178, 0.0562], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 01:44:32,809 INFO [finetune.py:976] (2/7) Epoch 11, batch 3550, loss[loss=0.1623, simple_loss=0.2057, pruned_loss=0.05943, over 3413.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.254, pruned_loss=0.06052, over 954027.26 frames. ], batch size: 14, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:44:40,088 INFO [optim.py:369] (2/7) 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 01:44:50,266 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60854.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:45:06,039 INFO [finetune.py:976] (2/7) Epoch 11, batch 3600, loss[loss=0.1524, simple_loss=0.2202, pruned_loss=0.0423, over 4826.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2514, pruned_loss=0.05996, over 952580.54 frames. ], batch size: 30, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:45:10,553 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 01:45:21,747 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60902.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:45:39,840 INFO [finetune.py:976] (2/7) Epoch 11, batch 3650, loss[loss=0.2197, simple_loss=0.2932, pruned_loss=0.07309, over 4739.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2547, pruned_loss=0.06103, over 953966.84 frames. ], batch size: 59, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:45:47,175 INFO [optim.py:369] (2/7) 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 01:46:04,733 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0363, 2.5019, 1.0371, 1.3141, 1.9498, 1.1682, 3.3887, 1.5997], + device='cuda:2'), covar=tensor([0.0712, 0.0710, 0.0769, 0.1303, 0.0490, 0.1035, 0.0262, 0.0657], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 01:46:25,827 INFO [finetune.py:976] (2/7) Epoch 11, batch 3700, loss[loss=0.2219, simple_loss=0.2929, pruned_loss=0.07549, over 4910.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2568, pruned_loss=0.06081, over 954302.53 frames. ], batch size: 36, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:47:05,594 INFO [finetune.py:976] (2/7) Epoch 11, batch 3750, loss[loss=0.183, simple_loss=0.2499, pruned_loss=0.05808, over 4767.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2562, pruned_loss=0.05998, over 954192.64 frames. ], batch size: 27, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:47:18,728 INFO [optim.py:369] (2/7) 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 01:47:20,753 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1165, 0.7808, 0.9470, 0.8101, 1.2219, 0.9891, 0.8293, 0.9813], + device='cuda:2'), covar=tensor([0.1696, 0.1456, 0.1952, 0.1538, 0.0932, 0.1458, 0.1793, 0.2112], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0318, 0.0350, 0.0295, 0.0331, 0.0317, 0.0305, 0.0359], + device='cuda:2'), out_proj_covar=tensor([6.4072e-05, 6.7233e-05, 7.5328e-05, 6.0703e-05, 6.9103e-05, 6.7556e-05, + 6.5118e-05, 7.7168e-05], device='cuda:2') +2023-04-27 01:47:29,913 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8718, 1.4460, 1.5090, 1.6280, 2.0282, 1.6710, 1.3873, 1.4274], + device='cuda:2'), covar=tensor([0.1683, 0.1604, 0.1931, 0.1222, 0.0847, 0.1843, 0.2164, 0.2089], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0318, 0.0350, 0.0295, 0.0331, 0.0317, 0.0305, 0.0359], + device='cuda:2'), out_proj_covar=tensor([6.4105e-05, 6.7217e-05, 7.5355e-05, 6.0722e-05, 6.9122e-05, 6.7552e-05, + 6.5112e-05, 7.7167e-05], device='cuda:2') +2023-04-27 01:47:40,541 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61061.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:48:01,154 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61074.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:48:05,281 INFO [finetune.py:976] (2/7) Epoch 11, batch 3800, loss[loss=0.1729, simple_loss=0.2452, pruned_loss=0.0503, over 4921.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2569, pruned_loss=0.05989, over 953740.22 frames. ], batch size: 38, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:48:56,476 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61122.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:48:59,869 INFO [finetune.py:976] (2/7) Epoch 11, batch 3850, loss[loss=0.1881, simple_loss=0.2416, pruned_loss=0.06729, over 4800.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2553, pruned_loss=0.05945, over 954139.80 frames. ], batch size: 51, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:49:08,085 INFO [optim.py:369] (2/7) 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 01:49:33,125 INFO [finetune.py:976] (2/7) Epoch 11, batch 3900, loss[loss=0.1711, simple_loss=0.2322, pruned_loss=0.05494, over 4749.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2526, pruned_loss=0.05888, over 952718.36 frames. ], batch size: 27, lr: 3.69e-03, grad_scale: 32.0 +2023-04-27 01:50:05,883 INFO [finetune.py:976] (2/7) Epoch 11, batch 3950, loss[loss=0.1588, simple_loss=0.2245, pruned_loss=0.04652, over 4771.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2505, pruned_loss=0.0584, over 953972.69 frames. ], batch size: 28, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:50:15,545 INFO [optim.py:369] (2/7) 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 01:50:39,618 INFO [finetune.py:976] (2/7) Epoch 11, batch 4000, loss[loss=0.2311, simple_loss=0.3025, pruned_loss=0.07982, over 4822.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2504, pruned_loss=0.05872, over 954657.73 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:51:18,202 INFO [finetune.py:976] (2/7) Epoch 11, batch 4050, loss[loss=0.1676, simple_loss=0.2413, pruned_loss=0.04691, over 4826.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.254, pruned_loss=0.06053, over 953938.82 frames. ], batch size: 30, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:51:37,051 INFO [optim.py:369] (2/7) 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 01:51:37,924 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 01:52:21,695 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:52:23,441 INFO [finetune.py:976] (2/7) Epoch 11, batch 4100, loss[loss=0.2306, simple_loss=0.3042, pruned_loss=0.07847, over 4760.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2566, pruned_loss=0.0613, over 954973.61 frames. ], batch size: 59, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:53:06,242 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4773, 1.0615, 0.3748, 1.2243, 1.0598, 1.3628, 1.2412, 1.2894], + device='cuda:2'), covar=tensor([0.0540, 0.0424, 0.0453, 0.0583, 0.0332, 0.0581, 0.0547, 0.0606], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 01:53:16,502 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61417.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 01:53:25,488 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 4150, loss[loss=0.1978, simple_loss=0.2738, pruned_loss=0.06086, over 4900.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2585, pruned_loss=0.0625, over 952160.75 frames. ], batch size: 37, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:53:48,244 INFO [optim.py:369] (2/7) 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 01:54:08,543 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7708, 3.7450, 2.6478, 4.3874, 3.7722, 3.8464, 1.5196, 3.8198], + device='cuda:2'), covar=tensor([0.1832, 0.1086, 0.3200, 0.1685, 0.2745, 0.1805, 0.5841, 0.2517], + device='cuda:2'), in_proj_covar=tensor([0.0241, 0.0215, 0.0248, 0.0303, 0.0295, 0.0248, 0.0267, 0.0268], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 01:54:37,122 INFO [finetune.py:976] (2/7) Epoch 11, batch 4200, loss[loss=0.1973, simple_loss=0.2603, pruned_loss=0.06716, over 4733.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2582, pruned_loss=0.0621, over 951201.17 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:55:36,020 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0631, 2.6731, 1.1259, 1.3656, 2.1453, 1.2176, 3.4583, 1.7540], + device='cuda:2'), covar=tensor([0.0679, 0.0591, 0.0733, 0.1234, 0.0440, 0.0971, 0.0227, 0.0623], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 01:55:45,102 INFO [finetune.py:976] (2/7) Epoch 11, batch 4250, loss[loss=0.1713, simple_loss=0.2349, pruned_loss=0.05385, over 4776.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2558, pruned_loss=0.06129, over 951278.05 frames. ], batch size: 26, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:55:57,237 INFO [optim.py:369] (2/7) 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 01:56:37,316 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6608, 1.6115, 0.9027, 1.3542, 1.7429, 1.5904, 1.4240, 1.4945], + device='cuda:2'), covar=tensor([0.0507, 0.0387, 0.0366, 0.0544, 0.0275, 0.0530, 0.0518, 0.0572], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 01:56:39,812 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0700, 1.5498, 1.5309, 2.0289, 2.2369, 1.9291, 1.7859, 1.5228], + device='cuda:2'), covar=tensor([0.1834, 0.1571, 0.1569, 0.1435, 0.0880, 0.1616, 0.2051, 0.1998], + device='cuda:2'), in_proj_covar=tensor([0.0303, 0.0316, 0.0347, 0.0292, 0.0328, 0.0314, 0.0302, 0.0356], + device='cuda:2'), out_proj_covar=tensor([6.3351e-05, 6.6692e-05, 7.4626e-05, 6.0102e-05, 6.8631e-05, 6.6914e-05, + 6.4513e-05, 7.6408e-05], device='cuda:2') +2023-04-27 01:56:49,823 INFO [finetune.py:976] (2/7) Epoch 11, batch 4300, loss[loss=0.1684, simple_loss=0.2329, pruned_loss=0.05195, over 4898.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2534, pruned_loss=0.06054, over 953242.35 frames. ], batch size: 32, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:57:26,045 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0484, 1.3878, 1.2629, 1.6459, 1.5150, 1.7361, 1.2606, 2.4252], + device='cuda:2'), covar=tensor([0.0646, 0.0799, 0.0792, 0.1185, 0.0631, 0.0441, 0.0753, 0.0218], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 01:57:36,082 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1350, 2.5838, 1.0368, 1.3100, 1.9534, 1.3054, 3.5154, 1.7761], + device='cuda:2'), covar=tensor([0.0653, 0.0551, 0.0803, 0.1283, 0.0495, 0.0974, 0.0227, 0.0652], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 01:57:57,464 INFO [finetune.py:976] (2/7) Epoch 11, batch 4350, loss[loss=0.1886, simple_loss=0.2575, pruned_loss=0.05987, over 4944.00 frames. ], tot_loss[loss=0.185, simple_loss=0.251, pruned_loss=0.05947, over 953627.53 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:58:10,463 INFO [optim.py:369] (2/7) 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 01:59:02,121 INFO [finetune.py:976] (2/7) Epoch 11, batch 4400, loss[loss=0.1492, simple_loss=0.2196, pruned_loss=0.03942, over 4790.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2533, pruned_loss=0.06051, over 953948.53 frames. ], batch size: 26, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 01:59:24,038 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 01:59:55,382 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8905, 2.2082, 1.8360, 2.1047, 1.5328, 1.8431, 1.7816, 1.3452], + device='cuda:2'), covar=tensor([0.1835, 0.1190, 0.0870, 0.1163, 0.3533, 0.1226, 0.1985, 0.2670], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0314, 0.0225, 0.0283, 0.0312, 0.0270, 0.0252, 0.0274], + device='cuda:2'), out_proj_covar=tensor([1.1777e-04, 1.2611e-04, 9.0302e-05, 1.1321e-04, 1.2739e-04, 1.0822e-04, + 1.0270e-04, 1.0972e-04], device='cuda:2') +2023-04-27 01:59:57,240 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61717.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:00:08,835 INFO [finetune.py:976] (2/7) Epoch 11, batch 4450, loss[loss=0.1799, simple_loss=0.248, pruned_loss=0.05592, over 4825.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2572, pruned_loss=0.06193, over 953970.56 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:00:17,373 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61732.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:00:21,983 INFO [optim.py:369] (2/7) 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 02:00:24,560 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61743.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:00:45,387 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=61765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:01:02,752 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6280, 1.6593, 0.8200, 1.3296, 1.8834, 1.5456, 1.4096, 1.4743], + device='cuda:2'), covar=tensor([0.0529, 0.0398, 0.0381, 0.0588, 0.0286, 0.0519, 0.0521, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 02:01:03,251 INFO [finetune.py:976] (2/7) Epoch 11, batch 4500, loss[loss=0.1966, simple_loss=0.2663, pruned_loss=0.0635, over 4805.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2587, pruned_loss=0.06223, over 953598.32 frames. ], batch size: 25, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:01:19,131 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61804.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:01:42,632 INFO [finetune.py:976] (2/7) Epoch 11, batch 4550, loss[loss=0.1738, simple_loss=0.25, pruned_loss=0.04879, over 4923.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2595, pruned_loss=0.06231, over 955108.06 frames. ], batch size: 38, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:01:49,952 INFO [optim.py:369] (2/7) 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 02:01:57,205 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1447, 1.5260, 1.4242, 1.7442, 1.6552, 1.9284, 1.3258, 3.3710], + device='cuda:2'), covar=tensor([0.0593, 0.0778, 0.0772, 0.1172, 0.0605, 0.0413, 0.0744, 0.0135], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 02:02:15,316 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-04-27 02:02:16,333 INFO [finetune.py:976] (2/7) Epoch 11, batch 4600, loss[loss=0.2043, simple_loss=0.272, pruned_loss=0.06826, over 4919.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2596, pruned_loss=0.06251, over 955943.90 frames. ], batch size: 38, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:02:40,211 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61914.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:02:46,073 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6814, 1.1884, 1.7247, 2.1615, 1.7626, 1.6723, 1.7117, 1.6820], + device='cuda:2'), covar=tensor([0.5394, 0.7410, 0.7587, 0.7050, 0.6763, 0.9040, 0.8653, 0.8431], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0411, 0.0499, 0.0515, 0.0441, 0.0460, 0.0470, 0.0470], + device='cuda:2'), out_proj_covar=tensor([9.9820e-05, 1.0186e-04, 1.1256e-04, 1.2251e-04, 1.0682e-04, 1.1121e-04, + 1.1258e-04, 1.1277e-04], device='cuda:2') +2023-04-27 02:02:49,392 INFO [finetune.py:976] (2/7) Epoch 11, batch 4650, loss[loss=0.2041, simple_loss=0.2705, pruned_loss=0.06883, over 4916.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2568, pruned_loss=0.06174, over 955690.31 frames. ], batch size: 46, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:02:56,644 INFO [optim.py:369] (2/7) 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 02:03:01,034 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4236, 1.7324, 1.7261, 1.8929, 1.7221, 1.8495, 1.8012, 1.7283], + device='cuda:2'), covar=tensor([0.4858, 0.6830, 0.5975, 0.5383, 0.6687, 0.8884, 0.6949, 0.6493], + device='cuda:2'), in_proj_covar=tensor([0.0325, 0.0376, 0.0313, 0.0325, 0.0337, 0.0399, 0.0357, 0.0321], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 02:03:05,232 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 02:03:20,500 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2267, 1.4249, 1.2394, 1.3729, 1.2486, 1.1657, 1.1567, 0.9426], + device='cuda:2'), covar=tensor([0.1492, 0.1343, 0.0939, 0.1046, 0.2933, 0.1265, 0.1644, 0.1876], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0315, 0.0226, 0.0285, 0.0313, 0.0271, 0.0253, 0.0276], + device='cuda:2'), out_proj_covar=tensor([1.1875e-04, 1.2674e-04, 9.0683e-05, 1.1378e-04, 1.2779e-04, 1.0868e-04, + 1.0306e-04, 1.1035e-04], device='cuda:2') +2023-04-27 02:03:20,515 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61975.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:03:22,012 INFO [finetune.py:976] (2/7) Epoch 11, batch 4700, loss[loss=0.13, simple_loss=0.192, pruned_loss=0.03401, over 4049.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2523, pruned_loss=0.05945, over 954390.92 frames. ], batch size: 17, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:03:26,138 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1485, 1.1541, 5.0832, 4.4699, 4.4919, 4.7437, 4.2919, 4.2309], + device='cuda:2'), covar=tensor([0.8673, 0.9760, 0.1134, 0.2518, 0.1628, 0.2971, 0.2470, 0.2558], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0306, 0.0400, 0.0407, 0.0348, 0.0404, 0.0312, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:03:48,851 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6122, 3.5230, 2.5532, 4.1901, 3.6288, 3.6364, 1.3147, 3.5688], + device='cuda:2'), covar=tensor([0.1987, 0.1286, 0.3438, 0.2085, 0.3493, 0.2012, 0.6463, 0.2529], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0217, 0.0251, 0.0305, 0.0297, 0.0248, 0.0269, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 02:04:06,619 INFO [finetune.py:976] (2/7) Epoch 11, batch 4750, loss[loss=0.1558, simple_loss=0.2223, pruned_loss=0.04466, over 4904.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.251, pruned_loss=0.05922, over 955963.05 frames. ], batch size: 43, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:04:20,382 INFO [optim.py:369] (2/7) 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 02:04:27,872 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8798, 1.1134, 3.2648, 3.0361, 2.9274, 3.1698, 3.1000, 2.9324], + device='cuda:2'), covar=tensor([0.7379, 0.5619, 0.1346, 0.2073, 0.1310, 0.2208, 0.2510, 0.1549], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0304, 0.0399, 0.0406, 0.0347, 0.0404, 0.0312, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:04:56,946 INFO [finetune.py:976] (2/7) Epoch 11, batch 4800, loss[loss=0.2028, simple_loss=0.2723, pruned_loss=0.06662, over 4804.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2538, pruned_loss=0.05988, over 955676.34 frames. ], batch size: 45, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:05:05,178 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:05:11,912 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62099.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:05:27,342 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 4850, loss[loss=0.2013, simple_loss=0.2739, pruned_loss=0.06438, over 4812.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2577, pruned_loss=0.06117, over 954342.02 frames. ], batch size: 45, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:05:30,991 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 02:05:39,089 INFO [optim.py:369] (2/7) 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 02:05:47,101 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9684, 1.7328, 2.0712, 2.4125, 2.3764, 1.9277, 1.6255, 2.1310], + device='cuda:2'), covar=tensor([0.0874, 0.1070, 0.0588, 0.0540, 0.0618, 0.0849, 0.0948, 0.0549], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0205, 0.0185, 0.0176, 0.0181, 0.0190, 0.0161, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:06:08,686 INFO [finetune.py:976] (2/7) Epoch 11, batch 4900, loss[loss=0.157, simple_loss=0.2387, pruned_loss=0.0377, over 4782.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2593, pruned_loss=0.06188, over 954639.40 frames. ], batch size: 29, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:06:18,149 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 4950, loss[loss=0.1726, simple_loss=0.2108, pruned_loss=0.06723, over 3388.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2613, pruned_loss=0.06312, over 952025.78 frames. ], batch size: 14, lr: 3.68e-03, grad_scale: 64.0 +2023-04-27 02:07:01,570 INFO [optim.py:369] (2/7) 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 02:07:21,455 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62270.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:07:23,300 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 5000, loss[loss=0.174, simple_loss=0.2427, pruned_loss=0.05269, over 4914.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2584, pruned_loss=0.0617, over 951888.95 frames. ], batch size: 36, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:07:45,806 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5526, 1.7659, 1.3605, 1.1015, 1.2460, 1.1576, 1.3957, 1.1353], + device='cuda:2'), covar=tensor([0.1503, 0.1208, 0.1500, 0.1777, 0.2205, 0.1883, 0.1002, 0.2015], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0214, 0.0169, 0.0203, 0.0202, 0.0183, 0.0158, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 02:07:54,969 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 02:07:58,883 INFO [finetune.py:976] (2/7) Epoch 11, batch 5050, loss[loss=0.1557, simple_loss=0.2105, pruned_loss=0.0505, over 4239.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2555, pruned_loss=0.06088, over 952764.83 frames. ], batch size: 18, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:08:04,229 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62334.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:08:08,227 INFO [optim.py:369] (2/7) 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 02:08:32,081 INFO [finetune.py:976] (2/7) Epoch 11, batch 5100, loss[loss=0.1841, simple_loss=0.2484, pruned_loss=0.05994, over 4692.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2519, pruned_loss=0.05919, over 953742.65 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-04-27 02:08:40,384 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62388.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:08:42,184 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62391.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:08:48,004 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62399.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:09:06,026 INFO [finetune.py:976] (2/7) Epoch 11, batch 5150, loss[loss=0.1633, simple_loss=0.2342, pruned_loss=0.04625, over 4816.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2517, pruned_loss=0.05957, over 954843.19 frames. ], batch size: 33, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:09:12,052 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62436.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:09:14,850 INFO [optim.py:369] (2/7) 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 02:09:25,279 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62447.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:09:33,224 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 5200, loss[loss=0.1466, simple_loss=0.2135, pruned_loss=0.03987, over 4694.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2553, pruned_loss=0.06079, over 956484.32 frames. ], batch size: 23, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:10:02,043 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62478.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:10:40,111 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 02:10:46,292 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 02:10:58,656 INFO [finetune.py:976] (2/7) Epoch 11, batch 5250, loss[loss=0.188, simple_loss=0.2641, pruned_loss=0.05596, over 4820.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2575, pruned_loss=0.06145, over 956509.59 frames. ], batch size: 39, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:11:00,714 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 02:11:07,064 INFO [optim.py:369] (2/7) 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 02:11:07,768 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0356, 2.4364, 1.0234, 1.2759, 1.9101, 1.1769, 3.2220, 1.5746], + device='cuda:2'), covar=tensor([0.0671, 0.0664, 0.0745, 0.1255, 0.0494, 0.1007, 0.0217, 0.0655], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0049, 0.0047, 0.0051, 0.0053, 0.0077, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 02:11:10,701 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62545.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:11:28,228 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62570.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:11:32,447 INFO [finetune.py:976] (2/7) Epoch 11, batch 5300, loss[loss=0.1777, simple_loss=0.2564, pruned_loss=0.04954, over 4850.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2584, pruned_loss=0.06153, over 955174.18 frames. ], batch size: 44, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:11:35,049 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-27 02:11:37,338 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62585.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:11:52,298 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62606.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:11:57,086 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7605, 2.3963, 1.8673, 1.7607, 1.4179, 1.4181, 2.0428, 1.3788], + device='cuda:2'), covar=tensor([0.1352, 0.1244, 0.1353, 0.1610, 0.2055, 0.1654, 0.0828, 0.1859], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0214, 0.0169, 0.0203, 0.0202, 0.0183, 0.0158, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 02:12:00,029 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62618.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:05,911 INFO [finetune.py:976] (2/7) Epoch 11, batch 5350, loss[loss=0.1849, simple_loss=0.2461, pruned_loss=0.06181, over 4813.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2579, pruned_loss=0.06124, over 952410.37 frames. ], batch size: 38, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:12:06,036 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9335, 2.5012, 2.1174, 2.2084, 1.6995, 2.0675, 2.0562, 1.5874], + device='cuda:2'), covar=tensor([0.1928, 0.0993, 0.0794, 0.1165, 0.2761, 0.1001, 0.1816, 0.2413], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0315, 0.0226, 0.0287, 0.0313, 0.0271, 0.0254, 0.0276], + device='cuda:2'), out_proj_covar=tensor([1.1845e-04, 1.2664e-04, 9.0434e-05, 1.1460e-04, 1.2782e-04, 1.0868e-04, + 1.0357e-04, 1.1068e-04], device='cuda:2') +2023-04-27 02:12:07,197 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:12:15,950 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 02:12:18,132 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62646.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:32,053 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62666.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:12:39,519 INFO [finetune.py:976] (2/7) Epoch 11, batch 5400, loss[loss=0.1736, simple_loss=0.2489, pruned_loss=0.04917, over 4775.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2556, pruned_loss=0.06032, over 953010.03 frames. ], batch size: 29, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:13:06,276 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8627, 1.1325, 3.2994, 3.0414, 2.9856, 3.2288, 3.2102, 2.8980], + device='cuda:2'), covar=tensor([0.7453, 0.5806, 0.1447, 0.2285, 0.1528, 0.1958, 0.1608, 0.1622], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0302, 0.0397, 0.0402, 0.0345, 0.0402, 0.0309, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:13:12,193 INFO [finetune.py:976] (2/7) Epoch 11, batch 5450, loss[loss=0.1899, simple_loss=0.2483, pruned_loss=0.06574, over 4828.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2524, pruned_loss=0.05915, over 955276.94 frames. ], batch size: 41, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:13:12,319 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62727.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:13:20,548 INFO [optim.py:369] (2/7) 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 02:13:21,871 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1007, 2.5450, 0.8527, 1.5094, 1.5737, 1.8458, 1.6269, 0.7575], + device='cuda:2'), covar=tensor([0.1468, 0.1033, 0.1782, 0.1324, 0.1091, 0.0958, 0.1593, 0.1825], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0248, 0.0140, 0.0121, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 02:13:24,899 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62747.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:13:42,655 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 5500, loss[loss=0.2071, simple_loss=0.2747, pruned_loss=0.06978, over 4902.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2504, pruned_loss=0.05888, over 955433.15 frames. ], batch size: 37, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:13:46,177 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62778.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:13:50,890 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2747, 1.1781, 1.3372, 1.5952, 1.5534, 1.3393, 1.0594, 1.4639], + device='cuda:2'), covar=tensor([0.0980, 0.1438, 0.0931, 0.0643, 0.0804, 0.0924, 0.0955, 0.0670], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0204, 0.0183, 0.0175, 0.0179, 0.0189, 0.0159, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:14:04,839 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5361, 3.3738, 0.9537, 1.7419, 1.9966, 2.2032, 1.9116, 0.9955], + device='cuda:2'), covar=tensor([0.1426, 0.1001, 0.2181, 0.1437, 0.1141, 0.1160, 0.1761, 0.2019], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0247, 0.0140, 0.0121, 0.0133, 0.0152, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 02:14:14,802 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8938, 3.7716, 2.8179, 4.5111, 3.8622, 3.8667, 1.7029, 3.7992], + device='cuda:2'), covar=tensor([0.1675, 0.1199, 0.3579, 0.1238, 0.3546, 0.1637, 0.5823, 0.2385], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0217, 0.0249, 0.0304, 0.0298, 0.0247, 0.0269, 0.0268], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 02:14:18,345 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 11, batch 5550, loss[loss=0.1953, simple_loss=0.2657, pruned_loss=0.06246, over 4915.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2527, pruned_loss=0.05973, over 956320.17 frames. ], batch size: 37, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:14:22,722 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3634, 1.6071, 1.5382, 2.1423, 2.3551, 1.9195, 1.9362, 1.6717], + device='cuda:2'), covar=tensor([0.1492, 0.2042, 0.2100, 0.1603, 0.1288, 0.2194, 0.2212, 0.2263], + device='cuda:2'), in_proj_covar=tensor([0.0304, 0.0317, 0.0349, 0.0294, 0.0329, 0.0315, 0.0304, 0.0356], + device='cuda:2'), out_proj_covar=tensor([6.3754e-05, 6.6784e-05, 7.4978e-05, 6.0380e-05, 6.8801e-05, 6.7256e-05, + 6.4837e-05, 7.6344e-05], device='cuda:2') +2023-04-27 02:14:23,730 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:14:28,601 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0343, 2.5006, 1.2415, 1.7369, 2.4675, 1.8989, 1.8397, 1.8757], + device='cuda:2'), covar=tensor([0.0506, 0.0320, 0.0317, 0.0538, 0.0221, 0.0525, 0.0529, 0.0572], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 02:15:05,422 INFO [finetune.py:976] (2/7) Epoch 11, batch 5600, loss[loss=0.1943, simple_loss=0.2744, pruned_loss=0.05712, over 4744.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.256, pruned_loss=0.06057, over 956229.20 frames. ], batch size: 59, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:15:15,846 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8272, 1.1106, 3.2551, 3.0104, 2.9256, 3.1429, 3.1278, 2.8448], + device='cuda:2'), covar=tensor([0.7614, 0.5679, 0.1389, 0.2232, 0.1392, 0.2213, 0.2057, 0.1854], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0303, 0.0398, 0.0404, 0.0346, 0.0404, 0.0311, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:15:28,131 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2570, 1.5533, 1.3146, 1.4930, 1.2829, 1.2544, 1.3138, 1.0289], + device='cuda:2'), covar=tensor([0.1677, 0.1230, 0.0979, 0.1222, 0.3400, 0.1320, 0.1628, 0.2204], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0312, 0.0223, 0.0282, 0.0309, 0.0266, 0.0251, 0.0273], + device='cuda:2'), out_proj_covar=tensor([1.1678e-04, 1.2515e-04, 8.9263e-05, 1.1294e-04, 1.2619e-04, 1.0684e-04, + 1.0200e-04, 1.0908e-04], device='cuda:2') +2023-04-27 02:15:37,400 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62901.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:15:50,471 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 02:16:10,654 INFO [finetune.py:976] (2/7) Epoch 11, batch 5650, loss[loss=0.163, simple_loss=0.2288, pruned_loss=0.04858, over 4702.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2595, pruned_loss=0.06151, over 957045.55 frames. ], batch size: 23, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:16:11,895 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:16:23,344 INFO [optim.py:369] (2/7) 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 02:16:23,981 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:17:03,859 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-27 02:17:06,579 INFO [finetune.py:976] (2/7) Epoch 11, batch 5700, loss[loss=0.2139, simple_loss=0.2542, pruned_loss=0.08673, over 3968.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2559, pruned_loss=0.06144, over 939656.71 frames. ], batch size: 17, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:17:06,614 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:17:38,294 INFO [finetune.py:976] (2/7) Epoch 12, batch 0, loss[loss=0.22, simple_loss=0.2836, pruned_loss=0.07818, over 4866.00 frames. ], tot_loss[loss=0.22, simple_loss=0.2836, pruned_loss=0.07818, over 4866.00 frames. ], batch size: 34, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:17:38,294 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 02:17:47,542 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2194, 1.4301, 1.6984, 1.8610, 1.7581, 1.9284, 1.7208, 1.7524], + device='cuda:2'), covar=tensor([0.4323, 0.6107, 0.5438, 0.5231, 0.6219, 0.8086, 0.6295, 0.5712], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0379, 0.0315, 0.0327, 0.0339, 0.0400, 0.0359, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 02:17:53,690 INFO [finetune.py:1010] (2/7) Epoch 12, validation: loss=0.1544, simple_loss=0.2267, pruned_loss=0.04099, over 2265189.00 frames. +2023-04-27 02:17:53,690 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 02:18:07,089 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.01 vs. limit=5.0 +2023-04-27 02:18:09,361 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0691, 1.8353, 2.1762, 2.4352, 2.5146, 1.9528, 1.7600, 2.2047], + device='cuda:2'), covar=tensor([0.0973, 0.1163, 0.0670, 0.0678, 0.0593, 0.1014, 0.0855, 0.0631], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0204, 0.0183, 0.0175, 0.0179, 0.0189, 0.0159, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:18:18,013 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63022.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:18:39,643 INFO [optim.py:369] (2/7) 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 02:18:43,421 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-04-27 02:18:44,986 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63047.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:18:54,570 INFO [finetune.py:976] (2/7) Epoch 12, batch 50, loss[loss=0.2045, simple_loss=0.2715, pruned_loss=0.06871, over 4899.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2583, pruned_loss=0.06182, over 215836.35 frames. ], batch size: 36, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:19:04,815 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4702, 1.3758, 1.7418, 1.7561, 1.3345, 1.0834, 1.2458, 0.8203], + device='cuda:2'), covar=tensor([0.0564, 0.0723, 0.0474, 0.0664, 0.0801, 0.1624, 0.0738, 0.0886], + device='cuda:2'), in_proj_covar=tensor([0.0066, 0.0071, 0.0070, 0.0066, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 02:19:30,955 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63095.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:19:37,861 INFO [finetune.py:976] (2/7) Epoch 12, batch 100, loss[loss=0.1425, simple_loss=0.2055, pruned_loss=0.03973, over 4824.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2521, pruned_loss=0.06035, over 381605.06 frames. ], batch size: 25, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:19:54,524 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:20:01,184 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 12, batch 150, loss[loss=0.2171, simple_loss=0.2728, pruned_loss=0.08073, over 4817.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2479, pruned_loss=0.059, over 510493.51 frames. ], batch size: 39, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:20:58,535 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63201.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:01,804 INFO [finetune.py:976] (2/7) Epoch 12, batch 200, loss[loss=0.1525, simple_loss=0.2215, pruned_loss=0.04174, over 4795.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2468, pruned_loss=0.05867, over 609014.16 frames. ], batch size: 25, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:21:09,017 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 02:21:13,192 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:21:25,232 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63241.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:30,506 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63249.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:34,071 INFO [finetune.py:976] (2/7) Epoch 12, batch 250, loss[loss=0.1708, simple_loss=0.2355, pruned_loss=0.05308, over 4827.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2507, pruned_loss=0.05945, over 687795.18 frames. ], batch size: 25, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:21:52,393 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63282.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:21:53,171 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-27 02:21:56,554 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:22:12,122 INFO [finetune.py:976] (2/7) Epoch 12, batch 300, loss[loss=0.1952, simple_loss=0.2623, pruned_loss=0.06408, over 4889.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2541, pruned_loss=0.0606, over 746109.54 frames. ], batch size: 35, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:22:21,955 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8626, 2.3857, 1.8576, 1.7092, 1.3769, 1.3874, 1.9305, 1.2504], + device='cuda:2'), covar=tensor([0.1776, 0.1428, 0.1520, 0.1814, 0.2497, 0.2132, 0.1019, 0.2192], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0204, 0.0203, 0.0185, 0.0159, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 02:22:36,086 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:23:01,680 INFO [finetune.py:976] (2/7) Epoch 12, batch 350, loss[loss=0.1825, simple_loss=0.2527, pruned_loss=0.05615, over 4751.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2564, pruned_loss=0.06149, over 791708.14 frames. ], batch size: 27, lr: 3.67e-03, grad_scale: 32.0 +2023-04-27 02:23:19,008 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63370.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:23:23,103 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5571, 0.6185, 1.3573, 1.9045, 1.6216, 1.4397, 1.4090, 1.4961], + device='cuda:2'), covar=tensor([0.5126, 0.6906, 0.7129, 0.7089, 0.6174, 0.8299, 0.8357, 0.8146], + device='cuda:2'), in_proj_covar=tensor([0.0407, 0.0408, 0.0494, 0.0512, 0.0438, 0.0457, 0.0465, 0.0467], + device='cuda:2'), 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:2') +2023-04-27 02:23:57,902 INFO [finetune.py:976] (2/7) Epoch 12, batch 400, loss[loss=0.1877, simple_loss=0.2556, pruned_loss=0.05985, over 4896.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2588, pruned_loss=0.0625, over 828354.83 frames. ], batch size: 43, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:23:59,855 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2489, 1.5157, 1.4178, 1.7429, 1.6592, 1.7244, 1.3957, 3.1085], + device='cuda:2'), covar=tensor([0.0660, 0.0788, 0.0846, 0.1161, 0.0653, 0.0520, 0.0755, 0.0199], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 02:24:11,156 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7473, 1.3416, 1.8178, 2.1589, 1.8790, 1.6624, 1.7027, 1.7426], + device='cuda:2'), covar=tensor([0.5417, 0.8333, 0.8147, 0.7036, 0.6472, 0.9982, 0.9762, 0.9493], + device='cuda:2'), in_proj_covar=tensor([0.0408, 0.0408, 0.0495, 0.0513, 0.0439, 0.0458, 0.0467, 0.0468], + device='cuda:2'), 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:2') +2023-04-27 02:24:18,829 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63422.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:24:31,387 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63429.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:24:43,211 INFO [optim.py:369] (2/7) 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 02:25:03,610 INFO [finetune.py:976] (2/7) Epoch 12, batch 450, loss[loss=0.1332, simple_loss=0.2044, pruned_loss=0.03097, over 4776.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2579, pruned_loss=0.06166, over 856779.43 frames. ], batch size: 27, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:25:38,225 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63477.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:25:46,557 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63481.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:25:47,777 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:26:01,236 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1130, 2.4855, 0.9683, 1.3582, 1.7099, 1.2460, 3.0874, 1.6733], + device='cuda:2'), covar=tensor([0.0666, 0.0643, 0.0786, 0.1243, 0.0518, 0.0947, 0.0285, 0.0618], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 02:26:12,256 INFO [finetune.py:976] (2/7) Epoch 12, batch 500, loss[loss=0.1799, simple_loss=0.2434, pruned_loss=0.05814, over 4928.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2544, pruned_loss=0.05987, over 879510.19 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:26:40,111 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0989, 1.0548, 5.0243, 4.4660, 4.4314, 4.6211, 4.2838, 4.1952], + device='cuda:2'), covar=tensor([0.8575, 0.9639, 0.1655, 0.3443, 0.1683, 0.2807, 0.3181, 0.2999], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0303, 0.0401, 0.0405, 0.0348, 0.0404, 0.0313, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:26:41,244 INFO [optim.py:369] (2/7) 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 02:26:50,386 INFO [finetune.py:976] (2/7) Epoch 12, batch 550, loss[loss=0.1845, simple_loss=0.2527, pruned_loss=0.0582, over 4767.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2519, pruned_loss=0.05947, over 895758.59 frames. ], batch size: 28, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:26:59,472 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63569.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:05,806 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:12,217 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63586.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:18,912 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 02:27:23,755 INFO [finetune.py:976] (2/7) Epoch 12, batch 600, loss[loss=0.2234, simple_loss=0.2889, pruned_loss=0.079, over 4865.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2523, pruned_loss=0.06007, over 909511.62 frames. ], batch size: 31, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:27:41,086 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63630.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:27:48,533 INFO [optim.py:369] (2/7) 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 02:27:52,950 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 650, loss[loss=0.1776, simple_loss=0.253, pruned_loss=0.0511, over 4912.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2557, pruned_loss=0.06065, over 919561.16 frames. ], batch size: 36, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:28:15,856 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 02:28:42,667 INFO [finetune.py:976] (2/7) Epoch 12, batch 700, loss[loss=0.191, simple_loss=0.2633, pruned_loss=0.05932, over 4805.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2571, pruned_loss=0.06047, over 927588.61 frames. ], batch size: 40, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:29:23,284 INFO [optim.py:369] (2/7) 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 02:29:26,938 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8468, 1.3308, 4.5645, 4.3161, 4.0398, 4.2738, 4.0650, 4.0801], + device='cuda:2'), covar=tensor([0.6939, 0.5817, 0.1055, 0.1547, 0.0993, 0.1106, 0.2643, 0.1580], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0303, 0.0401, 0.0405, 0.0348, 0.0404, 0.0313, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:29:32,955 INFO [finetune.py:976] (2/7) Epoch 12, batch 750, loss[loss=0.1765, simple_loss=0.2694, pruned_loss=0.04177, over 4756.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2581, pruned_loss=0.06085, over 934320.39 frames. ], batch size: 51, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:29:45,873 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63776.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:29:47,599 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63778.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:30:04,986 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-27 02:30:06,645 INFO [finetune.py:976] (2/7) Epoch 12, batch 800, loss[loss=0.2045, simple_loss=0.2658, pruned_loss=0.07163, over 4836.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2582, pruned_loss=0.06057, over 941296.47 frames. ], batch size: 44, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:30:09,831 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63810.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:30:37,933 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6294, 1.5542, 0.8241, 1.2615, 1.8111, 1.5307, 1.3709, 1.3716], + device='cuda:2'), covar=tensor([0.0510, 0.0381, 0.0379, 0.0566, 0.0275, 0.0505, 0.0494, 0.0602], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 02:30:40,780 INFO [optim.py:369] (2/7) 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 02:30:49,742 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63845.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:30:52,167 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4396, 3.3033, 0.9867, 1.9221, 1.8689, 2.3554, 1.8665, 0.9716], + device='cuda:2'), covar=tensor([0.1499, 0.0930, 0.2065, 0.1303, 0.1117, 0.0997, 0.1566, 0.1895], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0248, 0.0141, 0.0122, 0.0134, 0.0153, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 02:31:01,223 INFO [finetune.py:976] (2/7) Epoch 12, batch 850, loss[loss=0.1995, simple_loss=0.2616, pruned_loss=0.06869, over 4844.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.256, pruned_loss=0.06023, over 943368.11 frames. ], batch size: 47, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:31:03,794 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3124, 1.6940, 2.1505, 2.7390, 2.1036, 1.6943, 1.5179, 1.8747], + device='cuda:2'), covar=tensor([0.3525, 0.3460, 0.1687, 0.2338, 0.3099, 0.2965, 0.4597, 0.2504], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0250, 0.0223, 0.0316, 0.0215, 0.0229, 0.0231, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 02:31:21,245 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63871.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:31:24,880 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63877.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:31:33,723 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1019, 1.6479, 1.4772, 1.8514, 1.8797, 1.9461, 1.4202, 3.6235], + device='cuda:2'), covar=tensor([0.0649, 0.0794, 0.0815, 0.1181, 0.0582, 0.0479, 0.0772, 0.0157], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 02:32:06,882 INFO [finetune.py:976] (2/7) Epoch 12, batch 900, loss[loss=0.1795, simple_loss=0.2511, pruned_loss=0.05396, over 4790.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2525, pruned_loss=0.05912, over 945843.37 frames. ], batch size: 51, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:32:07,617 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63906.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:19,220 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:24,718 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:32:24,727 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:32:36,449 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 950, loss[loss=0.1537, simple_loss=0.2205, pruned_loss=0.04339, over 4798.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2497, pruned_loss=0.05787, over 948126.20 frames. ], batch size: 25, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:32:53,540 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4424, 1.2411, 3.9582, 3.6759, 3.4942, 3.7290, 3.6909, 3.4899], + device='cuda:2'), covar=tensor([0.7597, 0.6280, 0.1187, 0.1966, 0.1307, 0.1889, 0.2423, 0.1786], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0305, 0.0401, 0.0408, 0.0349, 0.0405, 0.0314, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:32:58,581 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 02:32:59,651 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:33:20,846 INFO [finetune.py:976] (2/7) Epoch 12, batch 1000, loss[loss=0.1713, simple_loss=0.2475, pruned_loss=0.04754, over 4914.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2533, pruned_loss=0.05886, over 950923.41 frames. ], batch size: 36, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:33:32,494 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64023.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:33:43,224 INFO [optim.py:369] (2/7) 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 02:33:58,121 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.57 vs. limit=5.0 +2023-04-27 02:33:59,166 INFO [finetune.py:976] (2/7) Epoch 12, batch 1050, loss[loss=0.2041, simple_loss=0.2732, pruned_loss=0.06756, over 4822.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2551, pruned_loss=0.05917, over 951285.51 frames. ], batch size: 40, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:34:29,323 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64076.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:34:30,571 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:34:40,115 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64084.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:35:05,821 INFO [finetune.py:976] (2/7) Epoch 12, batch 1100, loss[loss=0.1772, simple_loss=0.2603, pruned_loss=0.04708, over 4868.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2577, pruned_loss=0.06053, over 951205.45 frames. ], batch size: 34, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:35:17,120 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6635, 1.9555, 0.9037, 1.4257, 2.2126, 1.6162, 1.4834, 1.5979], + device='cuda:2'), covar=tensor([0.0527, 0.0367, 0.0326, 0.0573, 0.0250, 0.0533, 0.0508, 0.0595], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 02:35:28,513 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:35:29,731 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64126.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:35:31,259 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 02:35:38,343 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 12, batch 1150, loss[loss=0.2075, simple_loss=0.266, pruned_loss=0.07447, over 4817.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2589, pruned_loss=0.06093, over 952337.59 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:35:57,681 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:36:02,467 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6667, 1.8418, 0.8501, 1.3753, 1.8574, 1.5633, 1.4152, 1.5042], + device='cuda:2'), covar=tensor([0.0525, 0.0376, 0.0380, 0.0577, 0.0286, 0.0568, 0.0527, 0.0629], + device='cuda:2'), in_proj_covar=tensor([0.0030, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 02:36:11,613 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6239, 1.3744, 4.3072, 4.0026, 3.7801, 4.0826, 3.9559, 3.8055], + device='cuda:2'), covar=tensor([0.7131, 0.5764, 0.1018, 0.1777, 0.1163, 0.1541, 0.2020, 0.1622], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0304, 0.0398, 0.0405, 0.0346, 0.0403, 0.0314, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:36:19,868 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 1200, loss[loss=0.1959, simple_loss=0.2555, pruned_loss=0.06815, over 4850.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2566, pruned_loss=0.0601, over 953821.35 frames. ], batch size: 44, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:36:36,065 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:36:40,395 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:36:50,586 INFO [optim.py:369] (2/7) 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 02:36:51,916 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64242.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:10,837 INFO [finetune.py:976] (2/7) Epoch 12, batch 1250, loss[loss=0.1832, simple_loss=0.2462, pruned_loss=0.06007, over 4904.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2541, pruned_loss=0.05895, over 955354.50 frames. ], batch size: 32, lr: 3.66e-03, grad_scale: 64.0 +2023-04-27 02:37:33,224 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:33,823 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:55,439 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64290.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:37:57,338 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:38:16,205 INFO [finetune.py:976] (2/7) Epoch 12, batch 1300, loss[loss=0.1839, simple_loss=0.2614, pruned_loss=0.05318, over 4828.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.251, pruned_loss=0.05804, over 954767.42 frames. ], batch size: 38, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:38:32,617 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 02:38:40,957 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64345.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:38:45,293 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4981, 3.5033, 0.8639, 1.8699, 1.8685, 2.4481, 1.8956, 0.9127], + device='cuda:2'), covar=tensor([0.1475, 0.0841, 0.2132, 0.1352, 0.1203, 0.1073, 0.1504, 0.2247], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0246, 0.0139, 0.0121, 0.0133, 0.0152, 0.0118, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 02:38:49,942 INFO [finetune.py:976] (2/7) Epoch 12, batch 1350, loss[loss=0.1981, simple_loss=0.2728, pruned_loss=0.06174, over 4917.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2505, pruned_loss=0.05772, over 955722.32 frames. ], batch size: 36, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:39:07,141 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 1400, loss[loss=0.2053, simple_loss=0.2791, pruned_loss=0.06576, over 4739.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2544, pruned_loss=0.05879, over 956678.35 frames. ], batch size: 59, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:39:23,702 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64406.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:39:37,455 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3994, 1.3297, 1.4420, 1.0646, 1.3659, 1.1631, 1.6940, 1.1651], + device='cuda:2'), covar=tensor([0.3922, 0.1968, 0.5100, 0.2785, 0.1670, 0.2436, 0.1798, 0.5267], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0346, 0.0427, 0.0358, 0.0383, 0.0381, 0.0373, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:39:48,147 INFO [optim.py:369] (2/7) 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 02:40:07,658 INFO [finetune.py:976] (2/7) Epoch 12, batch 1450, loss[loss=0.1774, simple_loss=0.2575, pruned_loss=0.04866, over 4775.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2575, pruned_loss=0.05995, over 957582.57 frames. ], batch size: 29, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:40:21,237 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64465.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:40:21,843 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:40:55,245 INFO [zipformer.py:1188] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.1904, 1.5803, 1.4923, 1.8254, 1.7412, 2.1036, 1.4203, 3.9487], + device='cuda:2'), covar=tensor([0.0587, 0.0781, 0.0767, 0.1174, 0.0648, 0.0525, 0.0776, 0.0118], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 02:40:57,604 INFO [finetune.py:976] (2/7) Epoch 12, batch 1500, loss[loss=0.1541, simple_loss=0.2196, pruned_loss=0.04431, over 4898.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2581, pruned_loss=0.06046, over 957220.24 frames. ], batch size: 36, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:41:03,985 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64514.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:12,849 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64526.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:22,341 INFO [optim.py:369] (2/7) 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 02:41:27,204 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64549.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:30,772 INFO [finetune.py:976] (2/7) Epoch 12, batch 1550, loss[loss=0.2011, simple_loss=0.2715, pruned_loss=0.06531, over 4779.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.258, pruned_loss=0.06032, over 953454.74 frames. ], batch size: 51, lr: 3.66e-03, grad_scale: 32.0 +2023-04-27 02:41:42,605 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64572.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:53,767 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64588.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:41:54,938 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 1600, loss[loss=0.2173, simple_loss=0.289, pruned_loss=0.07284, over 4867.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2557, pruned_loss=0.05981, over 954994.97 frames. ], batch size: 34, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:42:29,956 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4955, 1.4926, 0.6688, 1.2035, 1.5502, 1.3050, 1.2691, 1.3043], + device='cuda:2'), covar=tensor([0.0642, 0.0365, 0.0409, 0.0664, 0.0318, 0.0687, 0.0671, 0.0657], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 02:42:31,727 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64620.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:42:46,405 INFO [optim.py:369] (2/7) 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 02:42:56,562 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64648.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:42:57,774 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 1650, loss[loss=0.1676, simple_loss=0.2339, pruned_loss=0.0506, over 4920.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2523, pruned_loss=0.05828, over 956424.27 frames. ], batch size: 37, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:43:36,877 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:09,957 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64701.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:11,220 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 1700, loss[loss=0.2222, simple_loss=0.286, pruned_loss=0.0792, over 4821.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2508, pruned_loss=0.05823, over 958829.95 frames. ], batch size: 38, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:44:13,248 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-27 02:44:14,916 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64709.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:27,405 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:44:39,610 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 02:44:40,036 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64745.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:41,140 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1723, 1.6379, 2.0562, 2.4856, 2.0525, 1.5923, 1.2779, 1.8509], + device='cuda:2'), covar=tensor([0.3617, 0.3691, 0.1844, 0.2867, 0.2922, 0.2990, 0.4712, 0.2268], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0247, 0.0221, 0.0314, 0.0212, 0.0227, 0.0229, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 02:44:44,764 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64752.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:44:46,482 INFO [finetune.py:976] (2/7) Epoch 12, batch 1750, loss[loss=0.1224, simple_loss=0.1986, pruned_loss=0.02306, over 4756.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2538, pruned_loss=0.05962, over 956873.99 frames. ], batch size: 26, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:44:52,005 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64764.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:45:20,207 INFO [finetune.py:976] (2/7) Epoch 12, batch 1800, loss[loss=0.2265, simple_loss=0.2912, pruned_loss=0.08091, over 4796.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2571, pruned_loss=0.06028, over 957936.18 frames. ], batch size: 29, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:45:20,917 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64806.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:45:25,145 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:45:37,028 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4640, 1.6389, 1.2883, 1.0445, 1.1481, 1.1119, 1.2444, 1.0519], + device='cuda:2'), covar=tensor([0.1648, 0.1383, 0.1705, 0.1972, 0.2383, 0.2105, 0.1195, 0.2174], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0214, 0.0170, 0.0204, 0.0202, 0.0184, 0.0158, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 02:45:37,645 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7109, 2.2428, 1.6007, 1.5074, 1.2649, 1.2736, 1.6233, 1.1622], + device='cuda:2'), covar=tensor([0.1649, 0.1423, 0.1573, 0.1908, 0.2281, 0.1987, 0.1074, 0.2076], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0214, 0.0170, 0.0204, 0.0202, 0.0184, 0.0158, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 02:45:37,748 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-04-27 02:46:00,004 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 12, batch 1850, loss[loss=0.2359, simple_loss=0.2942, pruned_loss=0.08879, over 4819.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2589, pruned_loss=0.06163, over 958953.13 frames. ], batch size: 38, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:46:56,846 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0426, 1.9160, 1.7160, 1.6610, 2.1463, 1.7622, 2.6750, 1.5342], + device='cuda:2'), covar=tensor([0.3992, 0.2106, 0.5362, 0.3541, 0.1802, 0.2604, 0.1359, 0.5062], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0347, 0.0430, 0.0358, 0.0385, 0.0382, 0.0376, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:46:58,527 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:47:03,897 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 02:47:10,162 INFO [finetune.py:976] (2/7) Epoch 12, batch 1900, loss[loss=0.2051, simple_loss=0.2723, pruned_loss=0.06895, over 4816.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2608, pruned_loss=0.06238, over 960322.02 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:47:35,553 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:47:39,044 INFO [optim.py:369] (2/7) 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 02:47:48,360 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 1950, loss[loss=0.1855, simple_loss=0.2284, pruned_loss=0.07125, over 4308.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2593, pruned_loss=0.06189, over 956616.96 frames. ], batch size: 18, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:48:19,691 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9929, 1.5288, 4.4157, 4.1746, 3.9425, 4.1226, 3.8894, 3.9226], + device='cuda:2'), covar=tensor([0.6071, 0.5302, 0.0974, 0.1640, 0.0927, 0.1374, 0.3300, 0.1251], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0300, 0.0395, 0.0402, 0.0343, 0.0399, 0.0310, 0.0362], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 02:48:19,824 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 02:48:31,528 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64976.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:48:44,181 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 02:48:49,861 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 12, batch 2000, loss[loss=0.203, simple_loss=0.2591, pruned_loss=0.07344, over 4866.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2558, pruned_loss=0.0606, over 954401.18 frames. ], batch size: 31, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:49:12,525 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65037.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 02:49:15,251 INFO [optim.py:369] (2/7) 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 02:49:21,190 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 2050, loss[loss=0.1678, simple_loss=0.2288, pruned_loss=0.05342, over 4865.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.252, pruned_loss=0.05934, over 955376.90 frames. ], batch size: 31, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:49:26,442 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:49:28,850 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65101.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:00,020 INFO [finetune.py:976] (2/7) Epoch 12, batch 2100, loss[loss=0.1649, simple_loss=0.2303, pruned_loss=0.04976, over 4898.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2526, pruned_loss=0.05976, over 955935.34 frames. ], batch size: 32, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:50:01,929 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65108.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:02,612 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3878, 1.7101, 1.6181, 2.2446, 2.3814, 1.9955, 1.9578, 1.6929], + device='cuda:2'), covar=tensor([0.1523, 0.1664, 0.2155, 0.1680, 0.1373, 0.2549, 0.2479, 0.2277], + device='cuda:2'), in_proj_covar=tensor([0.0304, 0.0316, 0.0351, 0.0294, 0.0331, 0.0317, 0.0304, 0.0359], + device='cuda:2'), out_proj_covar=tensor([6.3643e-05, 6.6506e-05, 7.5323e-05, 6.0411e-05, 6.9127e-05, 6.7353e-05, + 6.4911e-05, 7.6999e-05], device='cuda:2') +2023-04-27 02:50:06,883 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:10,309 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65121.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:50:22,434 INFO [optim.py:369] (2/7) 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 02:50:33,403 INFO [finetune.py:976] (2/7) Epoch 12, batch 2150, loss[loss=0.1628, simple_loss=0.2269, pruned_loss=0.04939, over 4748.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2553, pruned_loss=0.06084, over 955879.74 frames. ], batch size: 27, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:50:42,395 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65169.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:51:05,066 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 02:51:05,956 INFO [finetune.py:976] (2/7) Epoch 12, batch 2200, loss[loss=0.2304, simple_loss=0.2912, pruned_loss=0.08485, over 4820.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2578, pruned_loss=0.06204, over 953599.97 frames. ], batch size: 33, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:51:25,624 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8607, 2.4697, 2.1688, 2.2694, 1.6252, 2.0962, 1.9887, 1.6327], + device='cuda:2'), covar=tensor([0.2234, 0.1402, 0.0743, 0.1237, 0.3480, 0.1105, 0.1955, 0.2753], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0310, 0.0222, 0.0279, 0.0309, 0.0263, 0.0250, 0.0272], + device='cuda:2'), out_proj_covar=tensor([1.1650e-04, 1.2408e-04, 8.8694e-05, 1.1165e-04, 1.2597e-04, 1.0558e-04, + 1.0168e-04, 1.0863e-04], device='cuda:2') +2023-04-27 02:51:56,236 INFO [optim.py:369] (2/7) 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 02:51:58,808 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 2250, loss[loss=0.1655, simple_loss=0.2332, pruned_loss=0.04887, over 4908.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2599, pruned_loss=0.06263, over 953076.54 frames. ], batch size: 38, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:52:53,407 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65285.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:53:03,645 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:53:21,424 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65304.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:53:21,920 INFO [finetune.py:976] (2/7) Epoch 12, batch 2300, loss[loss=0.1845, simple_loss=0.2586, pruned_loss=0.05517, over 4759.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2586, pruned_loss=0.06136, over 955206.29 frames. ], batch size: 26, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:53:56,249 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65332.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 02:54:07,783 INFO [optim.py:369] (2/7) 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 02:54:16,353 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65346.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:20,876 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65352.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:22,655 INFO [finetune.py:976] (2/7) Epoch 12, batch 2350, loss[loss=0.1944, simple_loss=0.2662, pruned_loss=0.06126, over 4905.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2563, pruned_loss=0.06029, over 956645.92 frames. ], batch size: 46, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:54:25,635 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65359.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:41,561 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6172, 1.4353, 1.8560, 1.8446, 1.4394, 1.3229, 1.5028, 0.8994], + device='cuda:2'), covar=tensor([0.0593, 0.0809, 0.0427, 0.0789, 0.0891, 0.1028, 0.0706, 0.0802], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0072, 0.0071, 0.0068, 0.0076, 0.0097, 0.0076, 0.0071], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 02:54:42,727 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2854, 1.7332, 1.6793, 2.0601, 1.9858, 2.2023, 1.6108, 4.3876], + device='cuda:2'), covar=tensor([0.0580, 0.0832, 0.0762, 0.1179, 0.0603, 0.0492, 0.0732, 0.0108], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 02:54:53,204 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65401.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:55,972 INFO [finetune.py:976] (2/7) Epoch 12, batch 2400, loss[loss=0.139, simple_loss=0.2045, pruned_loss=0.03676, over 4758.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2527, pruned_loss=0.05913, over 956179.88 frames. ], batch size: 59, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:54:57,713 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65407.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:54:58,379 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65408.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:00,170 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65411.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:20,648 INFO [optim.py:369] (2/7) 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 02:55:25,586 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65449.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:29,227 INFO [finetune.py:976] (2/7) Epoch 12, batch 2450, loss[loss=0.1923, simple_loss=0.2665, pruned_loss=0.0591, over 4828.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2497, pruned_loss=0.05825, over 956576.97 frames. ], batch size: 33, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:55:30,348 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65456.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:55:46,735 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1000, 2.3917, 1.0842, 1.3362, 1.7730, 1.2536, 3.0014, 1.6493], + device='cuda:2'), covar=tensor([0.0637, 0.0634, 0.0701, 0.1350, 0.0498, 0.1041, 0.0407, 0.0687], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0051, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 02:56:02,902 INFO [finetune.py:976] (2/7) Epoch 12, batch 2500, loss[loss=0.2425, simple_loss=0.2822, pruned_loss=0.1014, over 4896.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2522, pruned_loss=0.05941, over 956078.32 frames. ], batch size: 35, lr: 3.65e-03, grad_scale: 32.0 +2023-04-27 02:56:16,847 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1920, 3.2282, 2.5317, 3.7136, 3.1584, 3.2110, 1.5313, 3.2040], + device='cuda:2'), covar=tensor([0.1951, 0.1463, 0.3960, 0.2620, 0.3004, 0.2031, 0.5368, 0.2731], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0250, 0.0305, 0.0300, 0.0249, 0.0271, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 02:56:19,308 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 02:56:22,231 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65532.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 02:56:28,027 INFO [optim.py:369] (2/7) 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 02:56:33,192 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 02:56:36,599 INFO [finetune.py:976] (2/7) Epoch 12, batch 2550, loss[loss=0.2122, simple_loss=0.2803, pruned_loss=0.0721, over 4853.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2564, pruned_loss=0.06044, over 954598.65 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:56:52,702 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 02:57:02,306 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65593.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 02:57:10,044 INFO [finetune.py:976] (2/7) Epoch 12, batch 2600, loss[loss=0.1765, simple_loss=0.2541, pruned_loss=0.04943, over 4919.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.258, pruned_loss=0.06057, over 957221.66 frames. ], batch size: 42, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:57:29,309 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65632.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:57:29,954 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7455, 1.9405, 1.0240, 1.4317, 2.2039, 1.6422, 1.5700, 1.5770], + device='cuda:2'), covar=tensor([0.0489, 0.0339, 0.0326, 0.0539, 0.0264, 0.0471, 0.0449, 0.0532], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 02:57:34,725 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65641.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:57:35,235 INFO [optim.py:369] (2/7) 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 02:57:54,585 INFO [finetune.py:976] (2/7) Epoch 12, batch 2650, loss[loss=0.2175, simple_loss=0.282, pruned_loss=0.07649, over 4812.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2585, pruned_loss=0.06042, over 955507.53 frames. ], batch size: 39, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:57:57,767 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6143, 1.4298, 1.9157, 1.9131, 1.4459, 1.3651, 1.5981, 0.9829], + device='cuda:2'), covar=tensor([0.0565, 0.0772, 0.0449, 0.0660, 0.0901, 0.1199, 0.0705, 0.0825], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0071, 0.0070, 0.0067, 0.0075, 0.0096, 0.0076, 0.0071], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 02:58:19,222 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2603, 2.8391, 0.9889, 1.3266, 2.1989, 1.2068, 3.9167, 1.7751], + device='cuda:2'), covar=tensor([0.0715, 0.0856, 0.0864, 0.1333, 0.0503, 0.1066, 0.0321, 0.0651], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0067, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 02:58:26,148 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:58:54,677 INFO [finetune.py:976] (2/7) Epoch 12, batch 2700, loss[loss=0.1651, simple_loss=0.2304, pruned_loss=0.04986, over 4708.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2571, pruned_loss=0.05996, over 955024.17 frames. ], batch size: 23, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 02:59:03,987 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65711.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 02:59:30,221 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.26 vs. limit=5.0 +2023-04-27 02:59:42,189 INFO [optim.py:369] (2/7) 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 03:00:01,387 INFO [finetune.py:976] (2/7) Epoch 12, batch 2750, loss[loss=0.1747, simple_loss=0.24, pruned_loss=0.05475, over 4890.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2552, pruned_loss=0.06007, over 954416.98 frames. ], batch size: 32, lr: 3.65e-03, grad_scale: 16.0 +2023-04-27 03:00:09,047 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65759.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:00:44,414 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6637, 3.2808, 1.0838, 1.8540, 1.8805, 2.2833, 1.9176, 1.1690], + device='cuda:2'), covar=tensor([0.1347, 0.1172, 0.1966, 0.1339, 0.1159, 0.1162, 0.1606, 0.1724], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0250, 0.0142, 0.0123, 0.0135, 0.0155, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:01:07,122 INFO [finetune.py:976] (2/7) Epoch 12, batch 2800, loss[loss=0.1814, simple_loss=0.2598, pruned_loss=0.05145, over 4809.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2515, pruned_loss=0.05869, over 954767.58 frames. ], batch size: 38, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:01:51,947 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65838.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:01:59,669 INFO [optim.py:369] (2/7) 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 03:02:05,217 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5068, 2.4117, 1.9169, 2.1819, 2.4500, 2.0830, 3.1857, 1.7864], + device='cuda:2'), covar=tensor([0.3791, 0.2334, 0.4817, 0.3694, 0.1928, 0.2672, 0.2032, 0.4728], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0347, 0.0428, 0.0358, 0.0383, 0.0382, 0.0376, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:02:08,130 INFO [finetune.py:976] (2/7) Epoch 12, batch 2850, loss[loss=0.2235, simple_loss=0.2818, pruned_loss=0.08259, over 4022.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2501, pruned_loss=0.05854, over 955042.01 frames. ], batch size: 65, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:02:29,540 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65888.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:02:31,925 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2370, 1.5074, 1.4275, 1.7865, 1.6153, 1.8010, 1.3570, 3.5016], + device='cuda:2'), covar=tensor([0.0636, 0.0769, 0.0814, 0.1145, 0.0625, 0.0548, 0.0753, 0.0151], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:02:38,193 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65899.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:02:38,776 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2328, 2.7756, 2.0930, 2.2040, 1.5609, 1.4473, 2.3521, 1.4479], + device='cuda:2'), covar=tensor([0.1817, 0.1839, 0.1530, 0.1894, 0.2657, 0.2134, 0.1051, 0.2210], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0214, 0.0170, 0.0203, 0.0202, 0.0184, 0.0157, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 03:02:41,704 INFO [finetune.py:976] (2/7) Epoch 12, batch 2900, loss[loss=0.196, simple_loss=0.2683, pruned_loss=0.06184, over 4835.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2538, pruned_loss=0.05988, over 955329.24 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:03:03,866 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6911, 1.3748, 1.8093, 1.8372, 1.4641, 1.3862, 1.4492, 0.8615], + device='cuda:2'), covar=tensor([0.0469, 0.0783, 0.0468, 0.0605, 0.0694, 0.1132, 0.0632, 0.0735], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0072, 0.0071, 0.0068, 0.0076, 0.0097, 0.0077, 0.0072], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:03:05,042 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:03:05,559 INFO [optim.py:369] (2/7) 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 03:03:15,314 INFO [finetune.py:976] (2/7) Epoch 12, batch 2950, loss[loss=0.2282, simple_loss=0.2807, pruned_loss=0.08789, over 4740.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.256, pruned_loss=0.06008, over 955763.41 frames. ], batch size: 59, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:03:27,469 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2540, 1.4425, 1.3781, 1.7109, 1.5295, 1.8689, 1.3317, 3.2163], + device='cuda:2'), covar=tensor([0.0669, 0.0943, 0.0899, 0.1319, 0.0739, 0.0598, 0.0896, 0.0214], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:03:36,944 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:03:49,752 INFO [finetune.py:976] (2/7) Epoch 12, batch 3000, loss[loss=0.187, simple_loss=0.2564, pruned_loss=0.05877, over 4848.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2578, pruned_loss=0.06135, over 955182.89 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:03:49,753 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 03:03:55,304 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2579, 1.6088, 1.4784, 1.8144, 1.6804, 1.7863, 1.4253, 3.0623], + device='cuda:2'), covar=tensor([0.0625, 0.0807, 0.0756, 0.1271, 0.0622, 0.0461, 0.0741, 0.0195], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:03:56,817 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5366, 1.6044, 3.7667, 3.5162, 3.4271, 3.5239, 3.6638, 3.3501], + device='cuda:2'), covar=tensor([0.6409, 0.4271, 0.1023, 0.1603, 0.1125, 0.1362, 0.0826, 0.1343], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0306, 0.0404, 0.0409, 0.0351, 0.0410, 0.0317, 0.0371], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:04:00,410 INFO [finetune.py:1010] (2/7) Epoch 12, validation: loss=0.1529, simple_loss=0.2247, pruned_loss=0.04052, over 2265189.00 frames. +2023-04-27 03:04:00,411 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 03:04:15,164 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2067, 3.8828, 3.0138, 4.7749, 3.9703, 4.1039, 1.8848, 4.1722], + device='cuda:2'), covar=tensor([0.1433, 0.1129, 0.3838, 0.0982, 0.2385, 0.1628, 0.5384, 0.1916], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0220, 0.0253, 0.0307, 0.0303, 0.0252, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 03:04:27,773 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0588, 1.4427, 1.2861, 1.7127, 1.5506, 1.6940, 1.3079, 3.0888], + device='cuda:2'), covar=tensor([0.0699, 0.0869, 0.0869, 0.1284, 0.0657, 0.0531, 0.0796, 0.0198], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:04:29,086 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 03:04:29,442 INFO [optim.py:369] (2/7) 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 03:04:38,278 INFO [finetune.py:976] (2/7) Epoch 12, batch 3050, loss[loss=0.1665, simple_loss=0.2446, pruned_loss=0.04414, over 4763.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2584, pruned_loss=0.06108, over 956390.09 frames. ], batch size: 26, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:05:10,512 INFO [finetune.py:976] (2/7) Epoch 12, batch 3100, loss[loss=0.1434, simple_loss=0.2152, pruned_loss=0.03584, over 4705.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2554, pruned_loss=0.0596, over 956013.33 frames. ], batch size: 59, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:05:40,810 INFO [optim.py:369] (2/7) 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 03:05:48,050 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:05:54,638 INFO [finetune.py:976] (2/7) Epoch 12, batch 3150, loss[loss=0.2483, simple_loss=0.2945, pruned_loss=0.101, over 4898.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2529, pruned_loss=0.05893, over 956735.21 frames. ], batch size: 43, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:05:58,709 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66160.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:06:07,462 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6403, 1.4362, 1.9073, 1.8551, 1.4245, 1.3372, 1.5408, 1.0451], + device='cuda:2'), covar=tensor([0.0569, 0.0861, 0.0453, 0.0730, 0.0970, 0.1304, 0.0792, 0.0836], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0072, 0.0071, 0.0067, 0.0076, 0.0097, 0.0077, 0.0072], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:06:27,530 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66188.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:06:30,551 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 03:06:36,567 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66194.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:06:49,145 INFO [finetune.py:976] (2/7) Epoch 12, batch 3200, loss[loss=0.1666, simple_loss=0.2398, pruned_loss=0.04672, over 4872.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.25, pruned_loss=0.0581, over 956792.13 frames. ], batch size: 34, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:06:49,251 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66205.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:07:12,333 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66221.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:07:34,236 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66236.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:07:43,560 INFO [optim.py:369] (2/7) 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 03:07:57,821 INFO [finetune.py:976] (2/7) Epoch 12, batch 3250, loss[loss=0.1682, simple_loss=0.2345, pruned_loss=0.05095, over 4795.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2518, pruned_loss=0.05911, over 958283.75 frames. ], batch size: 25, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:07:59,795 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66258.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:08:20,726 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4068, 1.5453, 1.6238, 2.1194, 2.2683, 1.8920, 1.9053, 1.7516], + device='cuda:2'), covar=tensor([0.2014, 0.2151, 0.2055, 0.2021, 0.1472, 0.2387, 0.2941, 0.2195], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0322, 0.0356, 0.0298, 0.0337, 0.0320, 0.0309, 0.0363], + device='cuda:2'), out_proj_covar=tensor([6.4866e-05, 6.7869e-05, 7.6371e-05, 6.1257e-05, 7.0320e-05, 6.8084e-05, + 6.5726e-05, 7.7630e-05], device='cuda:2') +2023-04-27 03:08:47,973 INFO [finetune.py:976] (2/7) Epoch 12, batch 3300, loss[loss=0.1757, simple_loss=0.2516, pruned_loss=0.04994, over 4829.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2563, pruned_loss=0.06035, over 959032.10 frames. ], batch size: 47, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:08:57,560 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:09:13,362 INFO [optim.py:369] (2/7) 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 03:09:21,690 INFO [finetune.py:976] (2/7) Epoch 12, batch 3350, loss[loss=0.1403, simple_loss=0.2161, pruned_loss=0.03227, over 4697.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2586, pruned_loss=0.0616, over 956445.40 frames. ], batch size: 23, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:09:54,777 INFO [finetune.py:976] (2/7) Epoch 12, batch 3400, loss[loss=0.184, simple_loss=0.2612, pruned_loss=0.05339, over 4784.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2578, pruned_loss=0.06107, over 955085.90 frames. ], batch size: 51, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:10:20,216 INFO [optim.py:369] (2/7) 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 03:10:28,165 INFO [finetune.py:976] (2/7) Epoch 12, batch 3450, loss[loss=0.1994, simple_loss=0.2609, pruned_loss=0.06895, over 4846.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.257, pruned_loss=0.06075, over 954634.75 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:10:55,250 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66494.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:10:58,906 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66500.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:11:01,857 INFO [finetune.py:976] (2/7) Epoch 12, batch 3500, loss[loss=0.1541, simple_loss=0.2108, pruned_loss=0.04865, over 4247.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2552, pruned_loss=0.06029, over 953241.12 frames. ], batch size: 18, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:11:09,121 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66516.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:11:12,213 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66521.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:11:26,636 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66542.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:11:40,349 INFO [finetune.py:976] (2/7) Epoch 12, batch 3550, loss[loss=0.2056, simple_loss=0.2745, pruned_loss=0.06833, over 4880.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2523, pruned_loss=0.05969, over 953124.27 frames. ], batch size: 35, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:12:02,901 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4483, 1.6537, 1.8557, 1.9705, 1.8585, 2.0424, 1.9133, 1.8610], + device='cuda:2'), covar=tensor([0.4372, 0.6343, 0.5508, 0.5393, 0.6345, 0.8339, 0.5986, 0.5939], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0380, 0.0318, 0.0328, 0.0339, 0.0401, 0.0359, 0.0327], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 03:12:12,787 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66582.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:12:45,654 INFO [finetune.py:976] (2/7) Epoch 12, batch 3600, loss[loss=0.1614, simple_loss=0.2357, pruned_loss=0.04357, over 4795.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2499, pruned_loss=0.05863, over 952576.92 frames. ], batch size: 29, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:12:57,600 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66614.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:13:16,155 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 03:13:31,412 INFO [optim.py:369] (2/7) 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 03:13:45,503 INFO [finetune.py:976] (2/7) Epoch 12, batch 3650, loss[loss=0.1713, simple_loss=0.2507, pruned_loss=0.04591, over 4842.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2524, pruned_loss=0.0595, over 951199.88 frames. ], batch size: 49, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:13:57,246 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-04-27 03:13:59,532 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3325, 1.2836, 1.3759, 1.5978, 1.6614, 1.3625, 1.0265, 1.5453], + device='cuda:2'), covar=tensor([0.0813, 0.1185, 0.0846, 0.0603, 0.0606, 0.0786, 0.0839, 0.0543], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0202, 0.0184, 0.0173, 0.0179, 0.0187, 0.0157, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:14:15,353 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0588, 2.6542, 2.1345, 2.3946, 1.8701, 2.3325, 2.4188, 1.7111], + device='cuda:2'), covar=tensor([0.2301, 0.1321, 0.0849, 0.1546, 0.3246, 0.1095, 0.2013, 0.2647], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0311, 0.0223, 0.0282, 0.0310, 0.0261, 0.0251, 0.0270], + device='cuda:2'), out_proj_covar=tensor([1.1722e-04, 1.2417e-04, 8.9115e-05, 1.1257e-04, 1.2670e-04, 1.0465e-04, + 1.0222e-04, 1.0788e-04], device='cuda:2') +2023-04-27 03:14:19,347 INFO [finetune.py:976] (2/7) Epoch 12, batch 3700, loss[loss=0.1741, simple_loss=0.2385, pruned_loss=0.05489, over 4713.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2559, pruned_loss=0.06021, over 951908.02 frames. ], batch size: 23, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:14:35,823 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 03:14:46,405 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5121, 1.3571, 0.5261, 1.2268, 1.4061, 1.3982, 1.2793, 1.3285], + device='cuda:2'), covar=tensor([0.0509, 0.0383, 0.0415, 0.0563, 0.0303, 0.0520, 0.0521, 0.0562], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 03:14:52,769 INFO [finetune.py:976] (2/7) Epoch 12, batch 3750, loss[loss=0.1572, simple_loss=0.2327, pruned_loss=0.04084, over 4825.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2574, pruned_loss=0.06036, over 951709.08 frames. ], batch size: 39, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:14:55,464 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 03:15:16,262 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66792.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:15:21,667 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66800.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:15:25,454 INFO [finetune.py:976] (2/7) Epoch 12, batch 3800, loss[loss=0.2099, simple_loss=0.2938, pruned_loss=0.06305, over 4815.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2592, pruned_loss=0.06162, over 949986.75 frames. ], batch size: 40, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:15:32,286 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66816.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:15:34,093 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6276, 1.3827, 4.1913, 3.8851, 3.7035, 3.9145, 3.7723, 3.7574], + device='cuda:2'), covar=tensor([0.6626, 0.5393, 0.0901, 0.1482, 0.1083, 0.1525, 0.2481, 0.1100], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0307, 0.0402, 0.0409, 0.0350, 0.0407, 0.0314, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:15:48,321 INFO [optim.py:369] (2/7) 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 03:15:52,969 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 3850, loss[loss=0.1979, simple_loss=0.2662, pruned_loss=0.06477, over 4829.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2578, pruned_loss=0.06074, over 952182.08 frames. ], batch size: 47, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:16:04,613 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66864.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:16:08,873 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8856, 2.3068, 0.9518, 1.2598, 1.6244, 1.0732, 2.9816, 1.5503], + device='cuda:2'), covar=tensor([0.0734, 0.0611, 0.0750, 0.1193, 0.0533, 0.1003, 0.0251, 0.0623], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0049, 0.0047, 0.0051, 0.0053, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 03:16:12,476 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66877.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:16:31,348 INFO [finetune.py:976] (2/7) Epoch 12, batch 3900, loss[loss=0.2478, simple_loss=0.3015, pruned_loss=0.097, over 4209.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2562, pruned_loss=0.06077, over 953634.99 frames. ], batch size: 66, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:16:38,417 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66914.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:16:39,034 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7624, 1.9762, 0.9938, 1.5549, 2.2351, 1.7125, 1.6918, 1.6758], + device='cuda:2'), covar=tensor([0.0515, 0.0364, 0.0308, 0.0563, 0.0232, 0.0511, 0.0489, 0.0583], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 03:16:51,763 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9050, 1.3898, 1.5186, 1.6115, 2.1298, 1.7448, 1.4271, 1.4406], + device='cuda:2'), covar=tensor([0.1610, 0.1521, 0.2089, 0.1320, 0.0843, 0.1400, 0.2092, 0.2043], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0319, 0.0353, 0.0297, 0.0332, 0.0316, 0.0305, 0.0361], + device='cuda:2'), out_proj_covar=tensor([6.4102e-05, 6.7232e-05, 7.5756e-05, 6.1193e-05, 6.9374e-05, 6.7176e-05, + 6.4987e-05, 7.7302e-05], device='cuda:2') +2023-04-27 03:17:12,162 INFO [optim.py:369] (2/7) 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 03:17:32,060 INFO [finetune.py:976] (2/7) Epoch 12, batch 3950, loss[loss=0.1866, simple_loss=0.2566, pruned_loss=0.05832, over 4848.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2522, pruned_loss=0.05924, over 952885.01 frames. ], batch size: 49, lr: 3.64e-03, grad_scale: 16.0 +2023-04-27 03:17:33,416 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66957.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:17:42,917 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66962.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:18:39,567 INFO [finetune.py:976] (2/7) Epoch 12, batch 4000, loss[loss=0.1736, simple_loss=0.2448, pruned_loss=0.05119, over 4818.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2517, pruned_loss=0.05948, over 954696.93 frames. ], batch size: 41, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:18:58,944 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67018.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:19:13,721 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 03:19:14,011 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 12, batch 4050, loss[loss=0.2066, simple_loss=0.2668, pruned_loss=0.07318, over 4757.00 frames. ], tot_loss[loss=0.189, simple_loss=0.256, pruned_loss=0.06102, over 953706.91 frames. ], batch size: 28, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:19:24,975 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-04-27 03:19:42,050 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67083.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:19:44,508 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67087.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:19:46,353 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2603, 1.5482, 1.4431, 1.7031, 1.5802, 1.8633, 1.4190, 3.5902], + device='cuda:2'), covar=tensor([0.0634, 0.0792, 0.0771, 0.1216, 0.0687, 0.0572, 0.0766, 0.0134], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:19:56,230 INFO [finetune.py:976] (2/7) Epoch 12, batch 4100, loss[loss=0.2265, simple_loss=0.2998, pruned_loss=0.07657, over 4816.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2585, pruned_loss=0.06146, over 954664.29 frames. ], batch size: 40, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:19:58,846 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 03:20:04,605 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5080, 1.5157, 1.4845, 1.9507, 1.7741, 1.8813, 1.5555, 4.1057], + device='cuda:2'), covar=tensor([0.0655, 0.0963, 0.0901, 0.1315, 0.0739, 0.0671, 0.0847, 0.0198], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:20:17,041 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4722, 1.2162, 1.6748, 1.6761, 1.3015, 1.1712, 1.3230, 0.8783], + device='cuda:2'), covar=tensor([0.0655, 0.0894, 0.0548, 0.0760, 0.0807, 0.1365, 0.0778, 0.0769], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0071, 0.0071, 0.0067, 0.0076, 0.0097, 0.0076, 0.0071], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:20:21,139 INFO [optim.py:369] (2/7) 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 03:20:21,275 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0493, 2.8507, 3.3825, 3.5170, 3.4404, 2.9511, 2.5758, 3.2330], + device='cuda:2'), covar=tensor([0.0884, 0.0832, 0.0436, 0.0451, 0.0486, 0.0778, 0.0789, 0.0484], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0200, 0.0182, 0.0173, 0.0177, 0.0185, 0.0155, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:20:22,508 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:20:29,538 INFO [finetune.py:976] (2/7) Epoch 12, batch 4150, loss[loss=0.2201, simple_loss=0.2848, pruned_loss=0.07768, over 4907.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2595, pruned_loss=0.06191, over 951135.00 frames. ], batch size: 37, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:20:37,307 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1206, 1.3785, 1.2527, 1.5921, 1.4952, 1.7080, 1.2873, 2.9381], + device='cuda:2'), covar=tensor([0.0704, 0.0803, 0.0820, 0.1269, 0.0663, 0.0568, 0.0770, 0.0191], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:20:45,419 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67177.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:21:03,552 INFO [finetune.py:976] (2/7) Epoch 12, batch 4200, loss[loss=0.1943, simple_loss=0.2686, pruned_loss=0.06003, over 4918.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2596, pruned_loss=0.06132, over 950977.51 frames. ], batch size: 42, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:21:18,110 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67225.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:21:28,834 INFO [optim.py:369] (2/7) 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 03:21:37,113 INFO [finetune.py:976] (2/7) Epoch 12, batch 4250, loss[loss=0.2195, simple_loss=0.273, pruned_loss=0.08299, over 4794.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2566, pruned_loss=0.06055, over 952981.29 frames. ], batch size: 51, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:22:10,186 INFO [finetune.py:976] (2/7) Epoch 12, batch 4300, loss[loss=0.1757, simple_loss=0.2323, pruned_loss=0.05955, over 4700.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.254, pruned_loss=0.05964, over 954949.46 frames. ], batch size: 23, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:22:15,630 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67313.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:22:36,119 INFO [optim.py:369] (2/7) 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 03:22:43,959 INFO [finetune.py:976] (2/7) Epoch 12, batch 4350, loss[loss=0.161, simple_loss=0.2269, pruned_loss=0.04762, over 4798.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2511, pruned_loss=0.05856, over 956037.51 frames. ], batch size: 29, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:23:29,030 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67387.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:23:51,834 INFO [finetune.py:976] (2/7) Epoch 12, batch 4400, loss[loss=0.2231, simple_loss=0.2747, pruned_loss=0.08572, over 4739.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2507, pruned_loss=0.05837, over 955091.97 frames. ], batch size: 27, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:24:15,019 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0692, 0.7065, 0.9063, 0.7507, 1.2346, 0.9425, 0.8880, 0.9088], + device='cuda:2'), covar=tensor([0.1540, 0.1243, 0.1846, 0.1369, 0.0862, 0.1215, 0.1379, 0.1947], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0318, 0.0352, 0.0295, 0.0332, 0.0315, 0.0305, 0.0361], + device='cuda:2'), out_proj_covar=tensor([6.4040e-05, 6.6979e-05, 7.5505e-05, 6.0679e-05, 6.9404e-05, 6.6934e-05, + 6.4883e-05, 7.7344e-05], device='cuda:2') +2023-04-27 03:24:26,227 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67435.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:24:34,555 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67439.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:24:36,311 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 12, batch 4450, loss[loss=0.1888, simple_loss=0.2529, pruned_loss=0.06235, over 4783.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2543, pruned_loss=0.05957, over 954522.96 frames. ], batch size: 26, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:25:33,490 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.63 vs. limit=5.0 +2023-04-27 03:25:48,055 INFO [finetune.py:976] (2/7) Epoch 12, batch 4500, loss[loss=0.206, simple_loss=0.2715, pruned_loss=0.07024, over 4795.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2562, pruned_loss=0.06029, over 954962.33 frames. ], batch size: 25, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:25:53,063 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67513.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:26:12,842 INFO [optim.py:369] (2/7) 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 03:26:21,703 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4600, 1.3774, 4.0771, 3.8198, 3.5897, 3.8712, 3.8110, 3.6465], + device='cuda:2'), covar=tensor([0.6912, 0.5508, 0.1051, 0.1630, 0.1152, 0.1721, 0.1825, 0.1252], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0307, 0.0403, 0.0412, 0.0351, 0.0410, 0.0314, 0.0370], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:26:22,245 INFO [finetune.py:976] (2/7) Epoch 12, batch 4550, loss[loss=0.1817, simple_loss=0.2495, pruned_loss=0.05695, over 4801.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2572, pruned_loss=0.06015, over 955760.55 frames. ], batch size: 40, lr: 3.63e-03, grad_scale: 32.0 +2023-04-27 03:26:28,002 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.76 vs. limit=5.0 +2023-04-27 03:26:34,575 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67574.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:26:51,644 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 03:26:56,263 INFO [finetune.py:976] (2/7) Epoch 12, batch 4600, loss[loss=0.2033, simple_loss=0.2755, pruned_loss=0.06556, over 4807.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2574, pruned_loss=0.06036, over 956301.70 frames. ], batch size: 39, lr: 3.63e-03, grad_scale: 32.0 +2023-04-27 03:27:01,312 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:27:15,110 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5145, 0.9719, 1.2666, 1.2015, 1.6265, 1.3008, 1.1069, 1.2248], + device='cuda:2'), covar=tensor([0.1418, 0.1203, 0.1658, 0.1204, 0.0710, 0.1302, 0.1715, 0.1726], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0317, 0.0352, 0.0295, 0.0333, 0.0315, 0.0304, 0.0361], + device='cuda:2'), 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:2') +2023-04-27 03:27:20,971 INFO [optim.py:369] (2/7) 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 03:27:29,856 INFO [finetune.py:976] (2/7) Epoch 12, batch 4650, loss[loss=0.1494, simple_loss=0.2195, pruned_loss=0.0396, over 4788.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2553, pruned_loss=0.0597, over 957540.52 frames. ], batch size: 29, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:27:34,153 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67661.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:27:36,916 INFO [scaling.py:679] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:27:54,875 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:28:04,217 INFO [finetune.py:976] (2/7) Epoch 12, batch 4700, loss[loss=0.1647, simple_loss=0.2163, pruned_loss=0.05658, over 4735.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2515, pruned_loss=0.05814, over 957939.03 frames. ], batch size: 23, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:28:26,034 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7819, 1.2995, 4.2899, 4.0588, 3.7376, 3.9708, 3.9001, 3.8414], + device='cuda:2'), covar=tensor([0.6988, 0.5827, 0.1018, 0.1696, 0.1199, 0.1681, 0.2203, 0.1486], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0307, 0.0403, 0.0412, 0.0350, 0.0409, 0.0315, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:28:37,287 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67739.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:28:38,543 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67741.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:28:40,069 INFO [optim.py:369] (2/7) 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 03:28:58,108 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67753.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:28:58,806 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 03:28:59,232 INFO [finetune.py:976] (2/7) Epoch 12, batch 4750, loss[loss=0.206, simple_loss=0.269, pruned_loss=0.0715, over 4864.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2499, pruned_loss=0.05801, over 956961.30 frames. ], batch size: 44, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:29:11,198 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8066, 1.4271, 1.9099, 2.3230, 1.9205, 1.8080, 1.8328, 1.8375], + device='cuda:2'), covar=tensor([0.5380, 0.7180, 0.7460, 0.6206, 0.6357, 0.8947, 0.8761, 0.8284], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0406, 0.0496, 0.0514, 0.0441, 0.0462, 0.0469, 0.0471], + device='cuda:2'), out_proj_covar=tensor([9.9664e-05, 1.0074e-04, 1.1173e-04, 1.2194e-04, 1.0679e-04, 1.1151e-04, + 1.1234e-04, 1.1262e-04], device='cuda:2') +2023-04-27 03:29:23,303 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 03:29:42,892 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67787.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:29:58,539 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4090, 1.6625, 1.7508, 1.8613, 1.6384, 1.7707, 1.8219, 1.7209], + device='cuda:2'), covar=tensor([0.4863, 0.6810, 0.5836, 0.5404, 0.7073, 0.9443, 0.7521, 0.7022], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0376, 0.0314, 0.0326, 0.0339, 0.0398, 0.0357, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 03:30:06,230 INFO [finetune.py:976] (2/7) Epoch 12, batch 4800, loss[loss=0.2305, simple_loss=0.2926, pruned_loss=0.08422, over 4241.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2533, pruned_loss=0.05966, over 956267.37 frames. ], batch size: 65, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:30:23,008 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0684, 1.4533, 1.3119, 1.6090, 1.4946, 1.7249, 1.3324, 3.4536], + device='cuda:2'), covar=tensor([0.0736, 0.0966, 0.0965, 0.1399, 0.0779, 0.0699, 0.0915, 0.0165], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:30:35,557 INFO [optim.py:369] (2/7) 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 03:30:43,895 INFO [finetune.py:976] (2/7) Epoch 12, batch 4850, loss[loss=0.185, simple_loss=0.2417, pruned_loss=0.06419, over 4885.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2572, pruned_loss=0.06049, over 957431.36 frames. ], batch size: 32, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:30:54,465 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67869.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:31:05,484 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5237, 1.8514, 1.2953, 1.0847, 1.1738, 1.1480, 1.2856, 1.0730], + device='cuda:2'), covar=tensor([0.1986, 0.1206, 0.1700, 0.1914, 0.2649, 0.2302, 0.1175, 0.2170], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0214, 0.0170, 0.0204, 0.0203, 0.0185, 0.0157, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 03:31:17,819 INFO [finetune.py:976] (2/7) Epoch 12, batch 4900, loss[loss=0.2224, simple_loss=0.2777, pruned_loss=0.08355, over 4777.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2586, pruned_loss=0.0613, over 958142.13 frames. ], batch size: 26, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:31:43,017 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 12, batch 4950, loss[loss=0.1938, simple_loss=0.2666, pruned_loss=0.06051, over 4821.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.26, pruned_loss=0.06123, over 958627.72 frames. ], batch size: 39, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:32:01,238 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5484, 1.4088, 1.8234, 1.8418, 1.3362, 1.1856, 1.4910, 1.0050], + device='cuda:2'), covar=tensor([0.0709, 0.0893, 0.0471, 0.0708, 0.0996, 0.1441, 0.0857, 0.0843], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0071, 0.0070, 0.0067, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:32:09,670 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:32:26,418 INFO [finetune.py:976] (2/7) Epoch 12, batch 5000, loss[loss=0.1631, simple_loss=0.2312, pruned_loss=0.04753, over 4918.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2576, pruned_loss=0.06016, over 958168.19 frames. ], batch size: 36, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:32:36,355 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6309, 1.3970, 1.7947, 1.8669, 1.4149, 1.3066, 1.4745, 1.0056], + device='cuda:2'), covar=tensor([0.0648, 0.0903, 0.0483, 0.0663, 0.0868, 0.1359, 0.0698, 0.0764], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0070, 0.0070, 0.0067, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:32:47,884 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:32:51,637 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68042.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:32:52,122 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68048.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:32:59,803 INFO [finetune.py:976] (2/7) Epoch 12, batch 5050, loss[loss=0.1518, simple_loss=0.2226, pruned_loss=0.0405, over 4761.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2544, pruned_loss=0.05899, over 956349.46 frames. ], batch size: 27, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:33:22,295 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68087.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:33:23,464 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5911, 1.5501, 4.3981, 4.1680, 3.8702, 4.2127, 4.0016, 3.8310], + device='cuda:2'), covar=tensor([0.6793, 0.5499, 0.0905, 0.1524, 0.0993, 0.1409, 0.1477, 0.1366], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0308, 0.0404, 0.0412, 0.0352, 0.0410, 0.0316, 0.0373], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:33:26,546 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9724, 2.5955, 2.0215, 1.8257, 1.4001, 1.4218, 2.0305, 1.3443], + device='cuda:2'), covar=tensor([0.1654, 0.1391, 0.1424, 0.1901, 0.2373, 0.1971, 0.1017, 0.2103], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0212, 0.0170, 0.0203, 0.0202, 0.0184, 0.0157, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 03:33:27,713 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8102, 3.7322, 2.8028, 4.3811, 3.8494, 3.7954, 1.5807, 3.7142], + device='cuda:2'), covar=tensor([0.1835, 0.1209, 0.3176, 0.1769, 0.3085, 0.2005, 0.6391, 0.2395], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0215, 0.0248, 0.0303, 0.0298, 0.0247, 0.0271, 0.0267], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 03:33:33,613 INFO [finetune.py:976] (2/7) Epoch 12, batch 5100, loss[loss=0.1784, simple_loss=0.2494, pruned_loss=0.0537, over 4834.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.251, pruned_loss=0.0579, over 957861.07 frames. ], batch size: 30, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:33:43,239 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1830, 2.0032, 1.7975, 1.6877, 2.1546, 1.8025, 2.3817, 1.6087], + device='cuda:2'), covar=tensor([0.3073, 0.1320, 0.3810, 0.2373, 0.1229, 0.1974, 0.1475, 0.3891], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0345, 0.0427, 0.0355, 0.0378, 0.0379, 0.0373, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:34:11,406 INFO [optim.py:369] (2/7) 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 03:34:14,584 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68148.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:34:18,644 INFO [finetune.py:976] (2/7) Epoch 12, batch 5150, loss[loss=0.1954, simple_loss=0.2744, pruned_loss=0.05818, over 4912.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2505, pruned_loss=0.05771, over 955944.26 frames. ], batch size: 37, lr: 3.63e-03, grad_scale: 16.0 +2023-04-27 03:34:25,263 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8614, 1.3897, 4.9820, 4.7107, 4.4325, 4.7216, 4.3492, 4.4021], + device='cuda:2'), covar=tensor([0.7339, 0.6323, 0.0963, 0.1722, 0.1125, 0.1613, 0.1449, 0.1493], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0307, 0.0403, 0.0410, 0.0351, 0.0409, 0.0315, 0.0372], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:34:34,048 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68169.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:35:16,279 INFO [finetune.py:976] (2/7) Epoch 12, batch 5200, loss[loss=0.178, simple_loss=0.2533, pruned_loss=0.05134, over 4781.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2541, pruned_loss=0.05919, over 953918.34 frames. ], batch size: 28, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:35:26,340 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4977, 0.7249, 1.4899, 1.8832, 1.6358, 1.5019, 1.5044, 1.5452], + device='cuda:2'), covar=tensor([0.4509, 0.6058, 0.5889, 0.6301, 0.5492, 0.7269, 0.7020, 0.6988], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0406, 0.0495, 0.0511, 0.0440, 0.0461, 0.0468, 0.0469], + device='cuda:2'), out_proj_covar=tensor([9.9326e-05, 1.0074e-04, 1.1154e-04, 1.2128e-04, 1.0645e-04, 1.1108e-04, + 1.1202e-04, 1.1235e-04], device='cuda:2') +2023-04-27 03:35:34,921 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68217.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:36:09,271 INFO [optim.py:369] (2/7) 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 03:36:18,617 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4952, 1.3535, 4.3191, 4.0553, 3.8028, 4.0952, 3.9420, 3.8136], + device='cuda:2'), covar=tensor([0.7199, 0.5777, 0.0965, 0.1541, 0.1066, 0.1283, 0.1686, 0.1367], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0308, 0.0404, 0.0411, 0.0352, 0.0411, 0.0316, 0.0373], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:36:22,258 INFO [finetune.py:976] (2/7) Epoch 12, batch 5250, loss[loss=0.2058, simple_loss=0.2764, pruned_loss=0.06765, over 4920.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2558, pruned_loss=0.05932, over 953376.57 frames. ], batch size: 33, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:36:29,961 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8519, 2.2856, 0.9197, 1.2024, 1.6688, 1.1769, 2.5248, 1.3538], + device='cuda:2'), covar=tensor([0.0716, 0.0563, 0.0662, 0.1318, 0.0433, 0.0974, 0.0278, 0.0676], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0049, 0.0047, 0.0051, 0.0052, 0.0077, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 03:37:07,992 INFO [finetune.py:976] (2/7) Epoch 12, batch 5300, loss[loss=0.1766, simple_loss=0.2516, pruned_loss=0.05083, over 4913.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2574, pruned_loss=0.06005, over 952398.31 frames. ], batch size: 42, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:37:29,917 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:37:30,484 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:37:34,487 INFO [optim.py:369] (2/7) 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 03:37:37,608 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 12, batch 5350, loss[loss=0.1969, simple_loss=0.2649, pruned_loss=0.06441, over 4919.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2562, pruned_loss=0.05887, over 954384.56 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:37:54,993 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8931, 1.2702, 1.5499, 1.5283, 2.0005, 1.6468, 1.4129, 1.5000], + device='cuda:2'), covar=tensor([0.1853, 0.1668, 0.1993, 0.1449, 0.1028, 0.1544, 0.2416, 0.2365], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0320, 0.0354, 0.0298, 0.0335, 0.0317, 0.0309, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.4599e-05, 6.7359e-05, 7.6041e-05, 6.1071e-05, 6.9907e-05, 6.7277e-05, + 6.5792e-05, 7.7938e-05], device='cuda:2') +2023-04-27 03:38:00,874 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:38:01,531 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0851, 2.2310, 1.8762, 1.8843, 2.4401, 1.8758, 2.9279, 1.5989], + device='cuda:2'), covar=tensor([0.3308, 0.1488, 0.4228, 0.2526, 0.1282, 0.2402, 0.0951, 0.4417], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0339, 0.0423, 0.0351, 0.0376, 0.0377, 0.0368, 0.0413], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:38:10,025 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:38:11,933 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68399.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:38:15,663 INFO [finetune.py:976] (2/7) Epoch 12, batch 5400, loss[loss=0.2009, simple_loss=0.265, pruned_loss=0.06845, over 4805.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2535, pruned_loss=0.05824, over 955418.87 frames. ], batch size: 51, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:38:41,371 INFO [optim.py:369] (2/7) 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 03:38:41,453 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68443.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:38:49,104 INFO [finetune.py:976] (2/7) Epoch 12, batch 5450, loss[loss=0.1496, simple_loss=0.2222, pruned_loss=0.03849, over 4790.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.251, pruned_loss=0.05779, over 955445.67 frames. ], batch size: 29, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:38:52,254 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68460.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:39:33,462 INFO [finetune.py:976] (2/7) Epoch 12, batch 5500, loss[loss=0.1527, simple_loss=0.2252, pruned_loss=0.04012, over 4927.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2481, pruned_loss=0.05666, over 956046.13 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:39:38,627 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.52 vs. limit=2.0 +2023-04-27 03:39:58,193 INFO [optim.py:369] (2/7) 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 03:40:06,516 INFO [finetune.py:976] (2/7) Epoch 12, batch 5550, loss[loss=0.164, simple_loss=0.2373, pruned_loss=0.04532, over 4895.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2485, pruned_loss=0.05719, over 954571.02 frames. ], batch size: 32, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:40:21,100 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68578.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:40:54,927 INFO [finetune.py:976] (2/7) Epoch 12, batch 5600, loss[loss=0.1716, simple_loss=0.2395, pruned_loss=0.05181, over 4771.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2532, pruned_loss=0.05851, over 955682.08 frames. ], batch size: 26, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:41:36,476 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68637.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:41:37,653 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68639.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:41:46,139 INFO [optim.py:369] (2/7) 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 03:41:58,625 INFO [finetune.py:976] (2/7) Epoch 12, batch 5650, loss[loss=0.1996, simple_loss=0.2858, pruned_loss=0.05668, over 4805.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2574, pruned_loss=0.05991, over 955620.77 frames. ], batch size: 45, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:42:22,436 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9077, 2.5302, 1.9591, 1.8130, 1.3439, 1.2933, 1.9947, 1.3277], + device='cuda:2'), covar=tensor([0.1705, 0.1230, 0.1513, 0.1755, 0.2331, 0.1966, 0.1057, 0.2065], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0214, 0.0170, 0.0204, 0.0203, 0.0185, 0.0157, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 03:42:23,632 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1476, 1.5460, 1.6690, 1.8842, 2.3320, 1.9537, 1.6163, 1.5741], + device='cuda:2'), covar=tensor([0.1534, 0.1590, 0.2057, 0.1168, 0.0838, 0.1286, 0.2040, 0.2216], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0322, 0.0356, 0.0298, 0.0336, 0.0319, 0.0311, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.4889e-05, 6.7810e-05, 7.6402e-05, 6.1188e-05, 7.0143e-05, 6.7871e-05, + 6.6156e-05, 7.7893e-05], device='cuda:2') +2023-04-27 03:42:33,432 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:42:33,500 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:42:45,295 INFO [finetune.py:976] (2/7) Epoch 12, batch 5700, loss[loss=0.1598, simple_loss=0.2166, pruned_loss=0.05148, over 4414.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2533, pruned_loss=0.05888, over 942613.25 frames. ], batch size: 19, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:42:47,793 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3769, 2.1969, 2.5078, 2.8435, 2.4584, 2.2233, 2.3481, 2.3084], + device='cuda:2'), covar=tensor([0.5122, 0.6809, 0.7948, 0.5989, 0.6444, 0.9381, 0.8969, 0.9069], + device='cuda:2'), in_proj_covar=tensor([0.0413, 0.0409, 0.0500, 0.0515, 0.0443, 0.0464, 0.0472, 0.0475], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:43:16,427 INFO [finetune.py:976] (2/7) Epoch 13, batch 0, loss[loss=0.222, simple_loss=0.2751, pruned_loss=0.08446, over 4850.00 frames. ], tot_loss[loss=0.222, simple_loss=0.2751, pruned_loss=0.08446, over 4850.00 frames. ], batch size: 31, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:43:16,427 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 03:43:36,473 INFO [finetune.py:1010] (2/7) Epoch 13, validation: loss=0.1542, simple_loss=0.2264, pruned_loss=0.04102, over 2265189.00 frames. +2023-04-27 03:43:36,474 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 03:43:49,890 INFO [optim.py:369] (2/7) 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 03:43:50,012 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68743.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:43:51,895 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68746.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:43:56,108 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4349, 1.6157, 1.6830, 2.3016, 2.4706, 2.0280, 1.9833, 1.8327], + device='cuda:2'), covar=tensor([0.1545, 0.1850, 0.1904, 0.1620, 0.1144, 0.1869, 0.2303, 0.1917], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0318, 0.0352, 0.0294, 0.0332, 0.0316, 0.0307, 0.0361], + device='cuda:2'), 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:2') +2023-04-27 03:43:57,284 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68755.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:44:12,770 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9630, 1.3043, 1.7048, 2.4366, 2.5779, 1.8023, 1.5849, 1.9200], + device='cuda:2'), covar=tensor([0.1015, 0.1772, 0.1066, 0.0598, 0.0551, 0.1077, 0.1059, 0.0804], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0201, 0.0182, 0.0173, 0.0178, 0.0184, 0.0155, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:44:15,627 INFO [finetune.py:976] (2/7) Epoch 13, batch 50, loss[loss=0.2006, simple_loss=0.2678, pruned_loss=0.06673, over 4813.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.2648, pruned_loss=0.06611, over 217386.09 frames. ], batch size: 39, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:44:21,434 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68791.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:44:22,082 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6852, 1.6729, 1.7013, 1.3123, 1.7980, 1.4857, 2.2858, 1.4607], + device='cuda:2'), covar=tensor([0.3902, 0.1855, 0.4950, 0.3304, 0.1735, 0.2395, 0.1427, 0.4451], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0342, 0.0426, 0.0354, 0.0378, 0.0377, 0.0371, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:44:26,389 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9788, 1.7152, 2.0902, 2.3496, 2.0797, 1.8233, 1.9424, 1.9733], + device='cuda:2'), covar=tensor([0.5441, 0.7393, 0.8129, 0.6461, 0.6114, 0.9381, 0.9734, 0.9622], + device='cuda:2'), in_proj_covar=tensor([0.0412, 0.0407, 0.0499, 0.0514, 0.0442, 0.0462, 0.0470, 0.0473], + device='cuda:2'), 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:2') +2023-04-27 03:44:46,962 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3954, 1.9080, 2.3121, 2.8808, 2.2907, 1.8322, 1.7696, 2.2310], + device='cuda:2'), covar=tensor([0.2932, 0.3202, 0.1589, 0.2509, 0.2915, 0.2610, 0.4020, 0.2277], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0248, 0.0223, 0.0317, 0.0215, 0.0229, 0.0231, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 03:44:48,040 INFO [finetune.py:976] (2/7) Epoch 13, batch 100, loss[loss=0.1345, simple_loss=0.2096, pruned_loss=0.02972, over 4786.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2542, pruned_loss=0.06074, over 380869.35 frames. ], batch size: 29, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:44:55,551 INFO [optim.py:369] (2/7) 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 03:44:57,639 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 03:45:21,008 INFO [finetune.py:976] (2/7) Epoch 13, batch 150, loss[loss=0.17, simple_loss=0.2462, pruned_loss=0.04694, over 4934.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2489, pruned_loss=0.05896, over 506467.68 frames. ], batch size: 33, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:45:47,023 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6428, 1.6440, 1.1217, 1.3525, 1.8832, 1.5169, 1.4272, 1.4398], + device='cuda:2'), covar=tensor([0.0475, 0.0336, 0.0316, 0.0512, 0.0293, 0.0481, 0.0453, 0.0531], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0037, 0.0050, 0.0038, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 03:45:53,957 INFO [finetune.py:976] (2/7) Epoch 13, batch 200, loss[loss=0.1691, simple_loss=0.2231, pruned_loss=0.05754, over 4908.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2465, pruned_loss=0.05868, over 605374.28 frames. ], batch size: 32, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:45:55,118 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68934.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:46:01,420 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 13, batch 250, loss[loss=0.1904, simple_loss=0.2637, pruned_loss=0.05851, over 4897.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2507, pruned_loss=0.06006, over 682356.15 frames. ], batch size: 35, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:47:12,552 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3620, 3.0912, 0.9552, 1.7852, 1.6658, 2.3467, 1.8124, 1.0186], + device='cuda:2'), covar=tensor([0.1458, 0.0955, 0.1937, 0.1275, 0.1216, 0.0929, 0.1495, 0.2024], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0247, 0.0138, 0.0120, 0.0132, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:47:36,760 INFO [finetune.py:976] (2/7) Epoch 13, batch 300, loss[loss=0.2013, simple_loss=0.2771, pruned_loss=0.06271, over 4832.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2554, pruned_loss=0.06119, over 742010.38 frames. ], batch size: 30, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:47:47,290 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 03:47:48,556 INFO [zipformer.py:1188] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.3688, 1.3500, 1.7340, 1.6334, 1.3064, 1.1879, 1.3913, 0.9396], + device='cuda:2'), covar=tensor([0.0617, 0.0627, 0.0397, 0.0624, 0.0817, 0.1146, 0.0585, 0.0637], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0071, 0.0071, 0.0067, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:48:05,315 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 350, loss[loss=0.2026, simple_loss=0.2684, pruned_loss=0.06835, over 4704.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2586, pruned_loss=0.06224, over 788612.36 frames. ], batch size: 59, lr: 3.62e-03, grad_scale: 16.0 +2023-04-27 03:48:59,537 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69103.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:49:00,174 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 400, loss[loss=0.25, simple_loss=0.2878, pruned_loss=0.1061, over 4068.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2581, pruned_loss=0.06084, over 825494.59 frames. ], batch size: 65, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:49:24,720 INFO [optim.py:369] (2/7) 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 03:49:38,258 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 03:49:38,760 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69162.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:49:42,421 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69168.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:49:51,429 INFO [finetune.py:976] (2/7) Epoch 13, batch 450, loss[loss=0.181, simple_loss=0.2533, pruned_loss=0.05433, over 4825.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2567, pruned_loss=0.05993, over 855451.65 frames. ], batch size: 30, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:49:57,545 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8836, 2.5452, 1.8743, 1.7679, 1.3311, 1.3495, 1.9366, 1.2407], + device='cuda:2'), covar=tensor([0.1631, 0.1317, 0.1467, 0.1819, 0.2372, 0.1939, 0.0975, 0.2046], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0214, 0.0170, 0.0204, 0.0204, 0.0185, 0.0157, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 03:49:58,751 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-27 03:50:19,104 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69223.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:20,344 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:22,771 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:25,118 INFO [finetune.py:976] (2/7) Epoch 13, batch 500, loss[loss=0.1401, simple_loss=0.217, pruned_loss=0.03154, over 4810.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2536, pruned_loss=0.05843, over 878332.08 frames. ], batch size: 39, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:50:25,827 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 03:50:32,583 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 03:50:32,982 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69245.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:50:37,665 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7736, 2.0639, 1.9420, 2.1879, 1.9528, 2.1382, 2.0211, 2.0205], + device='cuda:2'), covar=tensor([0.4498, 0.6872, 0.5768, 0.4840, 0.6466, 0.8079, 0.7252, 0.6909], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0377, 0.0316, 0.0328, 0.0341, 0.0398, 0.0358, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 03:50:52,433 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:50:57,799 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 550, loss[loss=0.1636, simple_loss=0.2335, pruned_loss=0.04681, over 4764.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2502, pruned_loss=0.05756, over 896062.93 frames. ], batch size: 54, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:51:00,263 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5882, 1.3675, 4.4768, 4.1711, 3.9151, 4.1845, 4.1236, 3.9744], + device='cuda:2'), covar=tensor([0.6688, 0.5981, 0.0853, 0.1455, 0.1043, 0.1515, 0.1288, 0.1417], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0305, 0.0399, 0.0407, 0.0348, 0.0407, 0.0312, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:51:00,301 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69286.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:51:14,437 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69306.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 03:51:16,595 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6758, 1.7179, 1.8812, 1.3072, 1.9143, 1.5256, 2.3885, 1.6179], + device='cuda:2'), covar=tensor([0.3510, 0.1694, 0.4038, 0.2957, 0.1531, 0.2154, 0.1344, 0.4112], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0343, 0.0425, 0.0356, 0.0378, 0.0378, 0.0371, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 03:51:24,241 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:51:30,305 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2783, 1.5210, 1.5030, 1.8929, 1.7114, 1.8152, 1.4248, 3.4430], + device='cuda:2'), covar=tensor([0.0678, 0.0930, 0.0911, 0.1248, 0.0684, 0.0532, 0.0870, 0.0189], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0058], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 03:51:32,073 INFO [finetune.py:976] (2/7) Epoch 13, batch 600, loss[loss=0.1507, simple_loss=0.2179, pruned_loss=0.04177, over 4915.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.251, pruned_loss=0.05838, over 911076.52 frames. ], batch size: 36, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:51:32,811 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69334.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:51:37,046 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69341.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:51:38,135 INFO [optim.py:369] (2/7) 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 03:51:56,869 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69368.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:02,370 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.37 vs. limit=5.0 +2023-04-27 03:52:04,775 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 650, loss[loss=0.1985, simple_loss=0.2835, pruned_loss=0.05674, over 4932.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2534, pruned_loss=0.05901, over 919997.88 frames. ], batch size: 42, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:52:06,028 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3834, 1.3169, 1.6137, 1.6146, 1.2843, 1.1428, 1.3077, 0.9539], + device='cuda:2'), covar=tensor([0.0657, 0.0626, 0.0528, 0.0607, 0.0917, 0.1219, 0.0703, 0.0729], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0070, 0.0071, 0.0067, 0.0075, 0.0096, 0.0075, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:52:08,471 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4710, 1.6740, 1.2818, 1.0043, 1.1120, 1.0656, 1.2571, 1.0141], + device='cuda:2'), covar=tensor([0.1745, 0.1317, 0.1625, 0.1884, 0.2390, 0.2079, 0.1165, 0.2111], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0214, 0.0170, 0.0204, 0.0204, 0.0185, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 03:52:09,604 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69389.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:12,666 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69394.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:15,686 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69399.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:36,499 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69429.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:52:44,419 INFO [finetune.py:976] (2/7) Epoch 13, batch 700, loss[loss=0.2282, simple_loss=0.2902, pruned_loss=0.08312, over 4197.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2558, pruned_loss=0.05998, over 925004.37 frames. ], batch size: 65, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:52:56,107 INFO [optim.py:369] (2/7) 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 03:53:09,897 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 750, loss[loss=0.2152, simple_loss=0.2846, pruned_loss=0.07287, over 4861.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2559, pruned_loss=0.05993, over 930106.27 frames. ], batch size: 34, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:53:59,662 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 03:54:42,286 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69518.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:54:46,349 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 800, loss[loss=0.2011, simple_loss=0.2813, pruned_loss=0.06049, over 4806.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2563, pruned_loss=0.05979, over 934967.72 frames. ], batch size: 41, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:55:09,401 INFO [optim.py:369] (2/7) 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 03:55:15,770 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 03:55:18,173 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 03:55:35,210 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69581.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:55:36,382 INFO [finetune.py:976] (2/7) Epoch 13, batch 850, loss[loss=0.1698, simple_loss=0.236, pruned_loss=0.05181, over 4824.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2542, pruned_loss=0.05919, over 939219.54 frames. ], batch size: 30, lr: 3.61e-03, grad_scale: 16.0 +2023-04-27 03:55:47,513 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69601.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:56:07,728 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69629.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 03:56:10,108 INFO [finetune.py:976] (2/7) Epoch 13, batch 900, loss[loss=0.1485, simple_loss=0.2183, pruned_loss=0.03933, over 4774.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2515, pruned_loss=0.05869, over 944503.66 frames. ], batch size: 28, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:56:16,242 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69674.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 03:56:39,834 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 950, loss[loss=0.1992, simple_loss=0.2566, pruned_loss=0.07092, over 4763.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2507, pruned_loss=0.05875, over 949482.34 frames. ], batch size: 23, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:56:49,082 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5913, 1.6195, 0.9776, 1.3603, 1.5264, 1.4836, 1.4489, 1.4487], + device='cuda:2'), covar=tensor([0.0462, 0.0323, 0.0376, 0.0489, 0.0335, 0.0471, 0.0418, 0.0514], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0049, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 03:56:53,916 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69699.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:57:11,108 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69724.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:57:18,036 INFO [finetune.py:976] (2/7) Epoch 13, batch 1000, loss[loss=0.1488, simple_loss=0.2344, pruned_loss=0.03156, over 4763.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2525, pruned_loss=0.05884, over 950798.02 frames. ], batch size: 28, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:57:19,387 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69735.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 03:57:24,160 INFO [optim.py:369] (2/7) 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 03:57:26,749 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69747.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:57:28,610 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 1050, loss[loss=0.1671, simple_loss=0.2379, pruned_loss=0.04816, over 4892.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2563, pruned_loss=0.05997, over 952804.66 frames. ], batch size: 32, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:58:13,061 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69818.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:58:16,680 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69824.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:58:23,479 INFO [finetune.py:976] (2/7) Epoch 13, batch 1100, loss[loss=0.193, simple_loss=0.2541, pruned_loss=0.06596, over 4703.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.257, pruned_loss=0.06019, over 953346.66 frames. ], batch size: 59, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:58:30,701 INFO [optim.py:369] (2/7) 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 03:58:39,447 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9824, 1.9844, 2.3755, 2.4427, 1.9144, 1.6334, 1.9407, 1.0299], + device='cuda:2'), covar=tensor([0.0824, 0.0945, 0.0560, 0.0855, 0.0798, 0.1253, 0.0877, 0.0965], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0070, 0.0071, 0.0066, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 03:58:50,906 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69866.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:59:00,301 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69872.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:59:12,138 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69881.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 03:59:13,755 INFO [finetune.py:976] (2/7) Epoch 13, batch 1150, loss[loss=0.1866, simple_loss=0.2624, pruned_loss=0.05541, over 4925.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2593, pruned_loss=0.06098, over 954560.36 frames. ], batch size: 41, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 03:59:41,631 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69901.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 04:00:04,397 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7838, 2.0376, 1.3387, 1.6246, 2.2886, 1.6844, 1.7383, 1.7226], + device='cuda:2'), covar=tensor([0.0446, 0.0301, 0.0320, 0.0460, 0.0287, 0.0458, 0.0443, 0.0488], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0037, 0.0049, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 04:00:16,222 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:00:16,251 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69929.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 04:00:18,639 INFO [finetune.py:976] (2/7) Epoch 13, batch 1200, loss[loss=0.176, simple_loss=0.2405, pruned_loss=0.05573, over 4752.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2569, pruned_loss=0.05974, over 955389.06 frames. ], batch size: 28, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:00:32,298 INFO [optim.py:369] (2/7) 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 04:00:36,032 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69949.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:00:47,615 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6507, 1.5868, 0.8547, 1.3724, 1.7004, 1.5325, 1.4727, 1.4827], + device='cuda:2'), covar=tensor([0.0507, 0.0371, 0.0361, 0.0546, 0.0295, 0.0513, 0.0490, 0.0574], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0030, 0.0020, 0.0029, 0.0029, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 04:00:48,211 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6514, 1.7609, 5.6167, 5.2976, 4.9971, 5.2942, 4.8836, 4.9333], + device='cuda:2'), covar=tensor([0.6115, 0.6007, 0.1028, 0.1709, 0.0919, 0.1125, 0.0906, 0.1500], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0302, 0.0398, 0.0406, 0.0345, 0.0404, 0.0310, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:00:52,492 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:00:54,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5029, 1.0093, 1.2968, 1.2435, 1.6762, 1.3605, 1.1225, 1.2714], + device='cuda:2'), covar=tensor([0.1692, 0.1469, 0.2036, 0.1466, 0.0783, 0.1554, 0.1907, 0.2241], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0317, 0.0349, 0.0291, 0.0329, 0.0314, 0.0304, 0.0358], + device='cuda:2'), out_proj_covar=tensor([6.3851e-05, 6.6598e-05, 7.4808e-05, 5.9668e-05, 6.8746e-05, 6.6659e-05, + 6.4779e-05, 7.6530e-05], device='cuda:2') +2023-04-27 04:00:57,153 INFO [finetune.py:976] (2/7) Epoch 13, batch 1250, loss[loss=0.1808, simple_loss=0.2528, pruned_loss=0.05443, over 4866.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2529, pruned_loss=0.0579, over 955854.18 frames. ], batch size: 31, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:01:26,387 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:01:26,413 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70030.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:01:31,777 INFO [finetune.py:976] (2/7) Epoch 13, batch 1300, loss[loss=0.1522, simple_loss=0.2286, pruned_loss=0.03783, over 4827.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2498, pruned_loss=0.05696, over 956071.74 frames. ], batch size: 39, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:01:39,418 INFO [optim.py:369] (2/7) 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 04:01:40,130 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5577, 3.4373, 1.0185, 2.0614, 1.7812, 2.4916, 1.9839, 0.9959], + device='cuda:2'), covar=tensor([0.1342, 0.0923, 0.1874, 0.1086, 0.1172, 0.0977, 0.1349, 0.2026], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0245, 0.0138, 0.0121, 0.0133, 0.0152, 0.0116, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:01:44,719 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 04:01:45,190 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70050.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:01:58,566 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70072.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:01:59,202 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8409, 4.1365, 0.8093, 2.2887, 2.2467, 2.7730, 2.4045, 0.9382], + device='cuda:2'), covar=tensor([0.1361, 0.1087, 0.2153, 0.1223, 0.1095, 0.1036, 0.1431, 0.2159], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0245, 0.0138, 0.0121, 0.0133, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:02:05,230 INFO [finetune.py:976] (2/7) Epoch 13, batch 1350, loss[loss=0.1997, simple_loss=0.2819, pruned_loss=0.05871, over 4927.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2501, pruned_loss=0.05758, over 955577.81 frames. ], batch size: 42, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:02:16,792 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70098.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:02:24,493 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:02:38,384 INFO [finetune.py:976] (2/7) Epoch 13, batch 1400, loss[loss=0.1884, simple_loss=0.2633, pruned_loss=0.05672, over 4901.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2514, pruned_loss=0.0577, over 956660.99 frames. ], batch size: 37, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:02:43,372 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9050, 1.4214, 1.7404, 1.7533, 1.7024, 1.3810, 0.7658, 1.3412], + device='cuda:2'), covar=tensor([0.3788, 0.3862, 0.1920, 0.2565, 0.2958, 0.2983, 0.4516, 0.2577], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0247, 0.0222, 0.0314, 0.0212, 0.0227, 0.0230, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 04:02:45,033 INFO [optim.py:369] (2/7) 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 04:03:04,727 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:03:06,556 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9335, 2.4475, 1.9955, 2.3269, 1.5951, 2.0641, 2.1863, 1.7228], + device='cuda:2'), covar=tensor([0.1850, 0.1128, 0.0853, 0.1025, 0.3139, 0.0997, 0.1557, 0.2230], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0314, 0.0227, 0.0286, 0.0317, 0.0266, 0.0257, 0.0274], + device='cuda:2'), out_proj_covar=tensor([1.1897e-04, 1.2554e-04, 9.0594e-05, 1.1440e-04, 1.2933e-04, 1.0662e-04, + 1.0442e-04, 1.0962e-04], device='cuda:2') +2023-04-27 04:03:11,826 INFO [finetune.py:976] (2/7) Epoch 13, batch 1450, loss[loss=0.182, simple_loss=0.2532, pruned_loss=0.05545, over 4792.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2522, pruned_loss=0.05754, over 956427.32 frames. ], batch size: 45, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:03:16,822 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4782, 1.7163, 1.7890, 1.9432, 1.7814, 1.9397, 1.9554, 1.8413], + device='cuda:2'), covar=tensor([0.4645, 0.6218, 0.5175, 0.4751, 0.5965, 0.7716, 0.6255, 0.5837], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0375, 0.0315, 0.0328, 0.0340, 0.0397, 0.0357, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:03:18,055 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7401, 1.2966, 1.8293, 2.2737, 1.8492, 1.7317, 1.7710, 1.7769], + device='cuda:2'), covar=tensor([0.5246, 0.7756, 0.7381, 0.7071, 0.7089, 0.9376, 0.9261, 0.8874], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0404, 0.0491, 0.0511, 0.0439, 0.0459, 0.0465, 0.0469], + device='cuda:2'), 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:2') +2023-04-27 04:03:31,614 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-27 04:03:32,085 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9829, 1.9026, 1.6510, 1.5087, 2.0276, 1.5885, 2.4272, 1.4492], + device='cuda:2'), covar=tensor([0.3552, 0.1732, 0.4765, 0.2930, 0.1486, 0.2358, 0.1235, 0.4377], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0345, 0.0427, 0.0358, 0.0381, 0.0381, 0.0372, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:03:34,591 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1835, 2.8741, 2.4805, 2.6653, 1.9950, 2.5601, 2.6800, 2.0367], + device='cuda:2'), covar=tensor([0.2267, 0.1372, 0.0750, 0.1239, 0.3116, 0.1116, 0.1868, 0.2631], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0314, 0.0227, 0.0287, 0.0317, 0.0266, 0.0258, 0.0274], + device='cuda:2'), out_proj_covar=tensor([1.1914e-04, 1.2569e-04, 9.0769e-05, 1.1456e-04, 1.2922e-04, 1.0658e-04, + 1.0454e-04, 1.0969e-04], device='cuda:2') +2023-04-27 04:03:39,420 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1545, 1.6847, 2.0529, 2.4853, 2.1032, 1.6004, 1.3351, 1.9688], + device='cuda:2'), covar=tensor([0.3451, 0.3552, 0.1803, 0.2545, 0.2821, 0.2834, 0.4553, 0.2331], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0246, 0.0221, 0.0314, 0.0212, 0.0227, 0.0229, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 04:03:45,322 INFO [finetune.py:976] (2/7) Epoch 13, batch 1500, loss[loss=0.2596, simple_loss=0.3191, pruned_loss=0.1001, over 4759.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2556, pruned_loss=0.05935, over 957060.36 frames. ], batch size: 54, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:03:51,423 INFO [optim.py:369] (2/7) 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 04:04:00,072 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0940, 2.4842, 0.9702, 1.4930, 2.0903, 1.3323, 3.5579, 1.8173], + device='cuda:2'), covar=tensor([0.0659, 0.0643, 0.0825, 0.1374, 0.0494, 0.1002, 0.0275, 0.0627], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 04:04:10,480 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:04:40,470 INFO [finetune.py:976] (2/7) Epoch 13, batch 1550, loss[loss=0.202, simple_loss=0.2677, pruned_loss=0.06815, over 4886.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2554, pruned_loss=0.05874, over 957687.23 frames. ], batch size: 32, lr: 3.61e-03, grad_scale: 32.0 +2023-04-27 04:04:57,302 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-04-27 04:05:10,360 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-27 04:05:17,656 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:05:28,447 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70330.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:05:30,176 INFO [finetune.py:976] (2/7) Epoch 13, batch 1600, loss[loss=0.1168, simple_loss=0.1897, pruned_loss=0.02193, over 4746.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2543, pruned_loss=0.05867, over 957879.17 frames. ], batch size: 23, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:05:41,009 INFO [optim.py:369] (2/7) 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 04:05:50,512 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9371, 1.1004, 3.2085, 2.7997, 2.9086, 2.9434, 3.0052, 2.7488], + device='cuda:2'), covar=tensor([0.9616, 0.8062, 0.2432, 0.4020, 0.2882, 0.4131, 0.3841, 0.3669], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0304, 0.0399, 0.0407, 0.0345, 0.0405, 0.0310, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:06:16,252 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70378.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:06:19,131 INFO [finetune.py:976] (2/7) Epoch 13, batch 1650, loss[loss=0.1684, simple_loss=0.2315, pruned_loss=0.05266, over 4860.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2519, pruned_loss=0.05879, over 957399.93 frames. ], batch size: 44, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:06:52,946 INFO [finetune.py:976] (2/7) Epoch 13, batch 1700, loss[loss=0.2009, simple_loss=0.2695, pruned_loss=0.06618, over 4924.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2489, pruned_loss=0.05728, over 957787.15 frames. ], batch size: 37, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:06:59,067 INFO [optim.py:369] (2/7) 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 04:07:12,884 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8718, 1.0959, 1.5005, 1.6373, 1.5963, 1.6998, 1.5618, 1.5191], + device='cuda:2'), covar=tensor([0.4013, 0.5239, 0.4566, 0.4659, 0.5585, 0.7567, 0.5208, 0.4694], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0378, 0.0317, 0.0329, 0.0342, 0.0399, 0.0359, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:07:16,023 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:07:26,839 INFO [finetune.py:976] (2/7) Epoch 13, batch 1750, loss[loss=0.1839, simple_loss=0.2491, pruned_loss=0.05932, over 4739.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2517, pruned_loss=0.05849, over 956519.62 frames. ], batch size: 23, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:07:26,971 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4311, 1.6574, 1.3429, 1.0344, 1.1705, 1.0660, 1.3883, 1.0855], + device='cuda:2'), covar=tensor([0.1424, 0.1361, 0.1324, 0.1798, 0.2050, 0.1729, 0.0920, 0.1829], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0214, 0.0169, 0.0204, 0.0202, 0.0184, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 04:08:00,061 INFO [finetune.py:976] (2/7) Epoch 13, batch 1800, loss[loss=0.1902, simple_loss=0.2524, pruned_loss=0.064, over 4877.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2539, pruned_loss=0.05828, over 956313.35 frames. ], batch size: 32, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:08:06,007 INFO [optim.py:369] (2/7) 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 04:08:12,439 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 04:08:18,948 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 04:08:33,360 INFO [finetune.py:976] (2/7) Epoch 13, batch 1850, loss[loss=0.1641, simple_loss=0.2259, pruned_loss=0.05109, over 4787.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2549, pruned_loss=0.05868, over 954596.88 frames. ], batch size: 26, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:08:38,542 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-27 04:08:42,676 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5335, 1.0100, 1.2294, 1.1878, 1.6385, 1.3398, 1.0417, 1.2129], + device='cuda:2'), covar=tensor([0.1549, 0.1466, 0.1825, 0.1388, 0.0846, 0.1437, 0.2098, 0.2175], + device='cuda:2'), in_proj_covar=tensor([0.0304, 0.0314, 0.0347, 0.0289, 0.0326, 0.0312, 0.0303, 0.0357], + device='cuda:2'), out_proj_covar=tensor([6.3449e-05, 6.5908e-05, 7.4554e-05, 5.9093e-05, 6.8074e-05, 6.6222e-05, + 6.4399e-05, 7.6374e-05], device='cuda:2') +2023-04-27 04:08:54,137 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:09:06,409 INFO [finetune.py:976] (2/7) Epoch 13, batch 1900, loss[loss=0.1916, simple_loss=0.2661, pruned_loss=0.05848, over 4893.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2557, pruned_loss=0.05864, over 954961.09 frames. ], batch size: 37, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:09:12,480 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 13, batch 1950, loss[loss=0.1443, simple_loss=0.2191, pruned_loss=0.03478, over 4814.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2539, pruned_loss=0.05846, over 954338.78 frames. ], batch size: 39, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:10:17,390 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7064, 2.0123, 1.7772, 1.3098, 1.2411, 1.2979, 1.8391, 1.2317], + device='cuda:2'), covar=tensor([0.1732, 0.1555, 0.1455, 0.1908, 0.2454, 0.2060, 0.0940, 0.2074], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0214, 0.0169, 0.0203, 0.0202, 0.0184, 0.0157, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 04:10:39,638 INFO [finetune.py:976] (2/7) Epoch 13, batch 2000, loss[loss=0.1828, simple_loss=0.2428, pruned_loss=0.06139, over 4820.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2514, pruned_loss=0.05807, over 955575.36 frames. ], batch size: 40, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:10:52,178 INFO [optim.py:369] (2/7) 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 04:11:15,905 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 2050, loss[loss=0.1989, simple_loss=0.2648, pruned_loss=0.06655, over 4856.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2484, pruned_loss=0.05709, over 955032.48 frames. ], batch size: 49, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:11:48,333 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70814.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:12:01,813 INFO [finetune.py:976] (2/7) Epoch 13, batch 2100, loss[loss=0.2024, simple_loss=0.2761, pruned_loss=0.0644, over 4781.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2484, pruned_loss=0.0572, over 955151.29 frames. ], batch size: 54, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:12:08,818 INFO [optim.py:369] (2/7) 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 04:12:35,717 INFO [finetune.py:976] (2/7) Epoch 13, batch 2150, loss[loss=0.2189, simple_loss=0.2859, pruned_loss=0.07597, over 4928.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.253, pruned_loss=0.05895, over 955875.85 frames. ], batch size: 33, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:12:43,009 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-27 04:12:56,789 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:13:01,704 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2572, 1.2206, 1.3261, 1.5168, 1.5270, 1.2468, 1.0073, 1.4144], + device='cuda:2'), covar=tensor([0.0934, 0.1334, 0.0894, 0.0707, 0.0735, 0.0871, 0.0883, 0.0660], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0201, 0.0182, 0.0173, 0.0178, 0.0184, 0.0156, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:13:08,519 INFO [finetune.py:976] (2/7) Epoch 13, batch 2200, loss[loss=0.2156, simple_loss=0.2761, pruned_loss=0.07755, over 4788.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2557, pruned_loss=0.05962, over 955730.13 frames. ], batch size: 29, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:13:16,545 INFO [optim.py:369] (2/7) 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 04:13:16,788 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 04:13:24,030 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2157, 2.9667, 2.0695, 2.3315, 1.6263, 1.5763, 2.2031, 1.6111], + device='cuda:2'), covar=tensor([0.1855, 0.1505, 0.1617, 0.1829, 0.2520, 0.2215, 0.1156, 0.2199], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0214, 0.0169, 0.0204, 0.0203, 0.0184, 0.0157, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 04:13:29,541 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 2250, loss[loss=0.2388, simple_loss=0.3023, pruned_loss=0.08765, over 4805.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2575, pruned_loss=0.05986, over 954616.78 frames. ], batch size: 40, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:13:49,273 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 13, batch 2300, loss[loss=0.1991, simple_loss=0.2657, pruned_loss=0.0663, over 4762.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2563, pruned_loss=0.0588, over 953474.57 frames. ], batch size: 27, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:14:23,496 INFO [optim.py:369] (2/7) 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 04:14:30,614 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71053.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:15:10,013 INFO [finetune.py:976] (2/7) Epoch 13, batch 2350, loss[loss=0.1808, simple_loss=0.244, pruned_loss=0.05878, over 4736.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2546, pruned_loss=0.05842, over 955225.34 frames. ], batch size: 54, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:15:20,870 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71091.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:15:46,153 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4904, 2.1883, 2.6159, 2.9032, 2.8294, 2.2077, 1.9404, 2.4356], + device='cuda:2'), covar=tensor([0.0856, 0.1069, 0.0628, 0.0564, 0.0608, 0.0945, 0.0895, 0.0659], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0201, 0.0182, 0.0173, 0.0178, 0.0184, 0.0155, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:15:46,259 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 04:16:15,378 INFO [finetune.py:976] (2/7) Epoch 13, batch 2400, loss[loss=0.1358, simple_loss=0.2073, pruned_loss=0.03212, over 4819.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.251, pruned_loss=0.05708, over 956394.72 frames. ], batch size: 40, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:16:16,730 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2159, 2.0897, 1.8369, 1.8253, 2.1873, 1.7347, 2.5634, 1.5684], + device='cuda:2'), covar=tensor([0.3432, 0.1736, 0.3871, 0.2971, 0.1609, 0.2445, 0.1704, 0.4199], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0341, 0.0424, 0.0354, 0.0378, 0.0377, 0.0371, 0.0415], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:16:26,844 INFO [optim.py:369] (2/7) 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 04:16:34,297 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7636, 1.2585, 1.5273, 2.1125, 2.1658, 1.5823, 1.3343, 1.9233], + device='cuda:2'), covar=tensor([0.0860, 0.1891, 0.1166, 0.0607, 0.0553, 0.1011, 0.0924, 0.0604], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0202, 0.0182, 0.0173, 0.0178, 0.0184, 0.0156, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:16:54,089 INFO [finetune.py:976] (2/7) Epoch 13, batch 2450, loss[loss=0.144, simple_loss=0.2126, pruned_loss=0.0377, over 4770.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2482, pruned_loss=0.05611, over 955080.74 frames. ], batch size: 28, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:17:02,041 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6645, 0.9822, 1.6044, 2.0208, 1.6841, 1.5744, 1.5971, 1.6804], + device='cuda:2'), covar=tensor([0.5617, 0.7608, 0.7101, 0.7874, 0.6779, 0.9266, 0.9411, 0.9494], + device='cuda:2'), in_proj_covar=tensor([0.0413, 0.0409, 0.0496, 0.0516, 0.0443, 0.0463, 0.0470, 0.0474], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:17:07,381 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 04:17:28,066 INFO [finetune.py:976] (2/7) Epoch 13, batch 2500, loss[loss=0.2085, simple_loss=0.2751, pruned_loss=0.07092, over 4823.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2512, pruned_loss=0.05741, over 954735.55 frames. ], batch size: 40, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:17:34,107 INFO [optim.py:369] (2/7) 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 04:17:53,039 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4459, 3.0117, 1.1166, 1.8145, 1.6548, 2.3230, 1.7953, 1.0728], + device='cuda:2'), covar=tensor([0.1356, 0.1247, 0.1848, 0.1295, 0.1157, 0.0885, 0.1469, 0.1881], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0247, 0.0140, 0.0122, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:17:56,831 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 04:18:01,395 INFO [finetune.py:976] (2/7) Epoch 13, batch 2550, loss[loss=0.1604, simple_loss=0.2395, pruned_loss=0.04061, over 4810.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2528, pruned_loss=0.05721, over 956885.19 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:18:03,464 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 04:18:34,378 INFO [finetune.py:976] (2/7) Epoch 13, batch 2600, loss[loss=0.1907, simple_loss=0.2547, pruned_loss=0.06337, over 4792.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2544, pruned_loss=0.05796, over 955279.26 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:18:40,523 INFO [optim.py:369] (2/7) 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 04:18:43,641 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71348.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:18:55,001 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6648, 1.6952, 0.7772, 1.3447, 1.6780, 1.5222, 1.4173, 1.4695], + device='cuda:2'), covar=tensor([0.0500, 0.0380, 0.0370, 0.0571, 0.0289, 0.0528, 0.0517, 0.0598], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 04:18:55,047 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3443, 1.5392, 1.8036, 1.9540, 1.8583, 1.8986, 1.8904, 1.8633], + device='cuda:2'), covar=tensor([0.4428, 0.6348, 0.5081, 0.5036, 0.5750, 0.7496, 0.5862, 0.5380], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0378, 0.0317, 0.0331, 0.0341, 0.0399, 0.0359, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:19:05,373 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-27 04:19:08,105 INFO [finetune.py:976] (2/7) Epoch 13, batch 2650, loss[loss=0.176, simple_loss=0.2513, pruned_loss=0.05034, over 4885.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2556, pruned_loss=0.05819, over 953867.97 frames. ], batch size: 35, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:19:10,020 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71386.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:19:17,413 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1953, 2.5815, 0.8695, 1.4662, 1.5898, 1.9168, 1.7098, 0.8349], + device='cuda:2'), covar=tensor([0.1535, 0.1166, 0.1835, 0.1436, 0.1197, 0.1016, 0.1526, 0.1603], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0246, 0.0139, 0.0122, 0.0133, 0.0153, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:19:41,957 INFO [finetune.py:976] (2/7) Epoch 13, batch 2700, loss[loss=0.1894, simple_loss=0.2514, pruned_loss=0.0637, over 4902.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2541, pruned_loss=0.05778, over 954452.27 frames. ], batch size: 36, lr: 3.60e-03, grad_scale: 32.0 +2023-04-27 04:19:48,072 INFO [optim.py:369] (2/7) 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 04:19:57,114 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5077, 1.3285, 4.3859, 4.0488, 3.7984, 4.1390, 3.9905, 3.8522], + device='cuda:2'), covar=tensor([0.7297, 0.5928, 0.1041, 0.1889, 0.1241, 0.1554, 0.1781, 0.1710], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0307, 0.0403, 0.0409, 0.0348, 0.0406, 0.0313, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:19:58,761 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9566, 2.6258, 1.0523, 1.3586, 2.0229, 1.1192, 3.3673, 1.6464], + device='cuda:2'), covar=tensor([0.0703, 0.0667, 0.0819, 0.1297, 0.0511, 0.1055, 0.0249, 0.0647], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0049, 0.0047, 0.0051, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 04:19:58,784 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9724, 2.0609, 1.7327, 1.6604, 2.0904, 1.6756, 2.6502, 1.4967], + device='cuda:2'), covar=tensor([0.3564, 0.1792, 0.4619, 0.2910, 0.1698, 0.2411, 0.1246, 0.4448], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0342, 0.0426, 0.0356, 0.0378, 0.0379, 0.0372, 0.0415], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:20:30,300 INFO [finetune.py:976] (2/7) Epoch 13, batch 2750, loss[loss=0.1771, simple_loss=0.2331, pruned_loss=0.06057, over 4757.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2517, pruned_loss=0.05741, over 953993.02 frames. ], batch size: 28, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:21:21,453 INFO [finetune.py:976] (2/7) Epoch 13, batch 2800, loss[loss=0.1586, simple_loss=0.2244, pruned_loss=0.04638, over 4903.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2488, pruned_loss=0.05618, over 955544.28 frames. ], batch size: 36, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:21:33,187 INFO [optim.py:369] (2/7) 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 04:21:45,750 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4485, 3.4844, 1.0589, 1.8196, 1.8642, 2.4874, 1.9451, 0.9610], + device='cuda:2'), covar=tensor([0.1566, 0.0952, 0.2021, 0.1355, 0.1243, 0.1062, 0.1540, 0.2124], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0247, 0.0140, 0.0122, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:22:06,641 INFO [finetune.py:976] (2/7) Epoch 13, batch 2850, loss[loss=0.2159, simple_loss=0.255, pruned_loss=0.0884, over 4823.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2482, pruned_loss=0.0561, over 954296.61 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:22:21,788 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 04:22:28,001 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:22:36,057 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6689, 1.5485, 0.7759, 1.3689, 1.4986, 1.5347, 1.4234, 1.4752], + device='cuda:2'), covar=tensor([0.0474, 0.0371, 0.0376, 0.0565, 0.0302, 0.0535, 0.0513, 0.0563], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0048, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 04:22:40,693 INFO [finetune.py:976] (2/7) Epoch 13, batch 2900, loss[loss=0.1956, simple_loss=0.2775, pruned_loss=0.05683, over 4929.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2511, pruned_loss=0.05697, over 952288.40 frames. ], batch size: 42, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:22:46,790 INFO [optim.py:369] (2/7) 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 04:22:49,912 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71648.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:09,466 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 2950, loss[loss=0.186, simple_loss=0.2496, pruned_loss=0.06115, over 4896.00 frames. ], tot_loss[loss=0.184, simple_loss=0.253, pruned_loss=0.05756, over 951634.57 frames. ], batch size: 32, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:23:15,382 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71686.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:21,433 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71696.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:23:45,592 INFO [finetune.py:976] (2/7) Epoch 13, batch 3000, loss[loss=0.1623, simple_loss=0.2186, pruned_loss=0.05299, over 4206.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2538, pruned_loss=0.05782, over 951522.69 frames. ], batch size: 18, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:23:45,592 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 04:23:47,482 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6868, 1.0756, 1.6810, 2.2002, 1.8617, 1.6609, 1.6701, 1.6821], + device='cuda:2'), covar=tensor([0.5320, 0.7233, 0.6898, 0.6614, 0.6642, 0.8614, 0.8674, 0.8578], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0408, 0.0492, 0.0514, 0.0441, 0.0462, 0.0468, 0.0472], + device='cuda:2'), out_proj_covar=tensor([9.9688e-05, 1.0086e-04, 1.1092e-04, 1.2192e-04, 1.0659e-04, 1.1131e-04, + 1.1198e-04, 1.1277e-04], device='cuda:2') +2023-04-27 04:23:54,279 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3893, 3.4376, 2.5975, 3.8625, 3.4456, 3.4740, 1.4829, 3.3722], + device='cuda:2'), covar=tensor([0.1799, 0.1262, 0.3001, 0.2186, 0.2850, 0.2038, 0.5243, 0.2406], + device='cuda:2'), in_proj_covar=tensor([0.0241, 0.0213, 0.0247, 0.0301, 0.0296, 0.0245, 0.0268, 0.0268], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:23:55,115 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7030, 2.1651, 1.8244, 2.1585, 1.6748, 1.7602, 1.7824, 1.4757], + device='cuda:2'), covar=tensor([0.1781, 0.1046, 0.0888, 0.0967, 0.3051, 0.1026, 0.1745, 0.2235], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0313, 0.0226, 0.0284, 0.0314, 0.0267, 0.0256, 0.0273], + device='cuda:2'), out_proj_covar=tensor([1.1831e-04, 1.2499e-04, 9.0531e-05, 1.1365e-04, 1.2819e-04, 1.0673e-04, + 1.0384e-04, 1.0918e-04], device='cuda:2') +2023-04-27 04:23:56,065 INFO [finetune.py:1010] (2/7) Epoch 13, validation: loss=0.1517, simple_loss=0.224, pruned_loss=0.03973, over 2265189.00 frames. +2023-04-27 04:23:56,065 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 04:23:56,742 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71734.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:24:03,110 INFO [optim.py:369] (2/7) 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 04:24:05,208 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.18 vs. limit=5.0 +2023-04-27 04:24:16,836 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-27 04:24:27,544 INFO [finetune.py:976] (2/7) Epoch 13, batch 3050, loss[loss=0.2184, simple_loss=0.279, pruned_loss=0.0789, over 4918.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.254, pruned_loss=0.05752, over 951452.67 frames. ], batch size: 33, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:24:33,885 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5953, 3.6238, 0.7575, 1.8273, 1.9193, 2.5776, 1.9979, 0.9534], + device='cuda:2'), covar=tensor([0.1435, 0.0975, 0.2210, 0.1340, 0.1134, 0.0985, 0.1477, 0.2087], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0247, 0.0139, 0.0122, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:25:00,526 INFO [finetune.py:976] (2/7) Epoch 13, batch 3100, loss[loss=0.1908, simple_loss=0.2482, pruned_loss=0.06669, over 4912.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.253, pruned_loss=0.05732, over 952878.06 frames. ], batch size: 43, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:25:08,969 INFO [optim.py:369] (2/7) 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 04:25:54,815 INFO [finetune.py:976] (2/7) Epoch 13, batch 3150, loss[loss=0.1716, simple_loss=0.2472, pruned_loss=0.04801, over 4824.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2512, pruned_loss=0.05727, over 953309.78 frames. ], batch size: 41, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:26:48,033 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71919.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:27:02,055 INFO [finetune.py:976] (2/7) Epoch 13, batch 3200, loss[loss=0.1793, simple_loss=0.2564, pruned_loss=0.05106, over 4791.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2482, pruned_loss=0.05651, over 954217.25 frames. ], batch size: 45, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:27:08,183 INFO [optim.py:369] (2/7) 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 04:27:28,806 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:27:33,768 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:27:35,466 INFO [finetune.py:976] (2/7) Epoch 13, batch 3250, loss[loss=0.2375, simple_loss=0.3029, pruned_loss=0.08602, over 4749.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2488, pruned_loss=0.05667, over 954692.29 frames. ], batch size: 54, lr: 3.59e-03, grad_scale: 64.0 +2023-04-27 04:28:10,359 INFO [finetune.py:976] (2/7) Epoch 13, batch 3300, loss[loss=0.1826, simple_loss=0.2554, pruned_loss=0.05489, over 4906.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2517, pruned_loss=0.05753, over 954386.02 frames. ], batch size: 36, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:28:12,306 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0086, 2.1113, 1.6814, 1.6850, 2.1165, 1.7273, 2.6618, 1.5127], + device='cuda:2'), covar=tensor([0.4311, 0.1880, 0.5599, 0.3196, 0.2016, 0.2598, 0.1692, 0.5192], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0346, 0.0428, 0.0358, 0.0380, 0.0382, 0.0373, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:28:16,994 INFO [optim.py:369] (2/7) 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 04:28:28,724 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-04-27 04:28:43,989 INFO [finetune.py:976] (2/7) Epoch 13, batch 3350, loss[loss=0.2407, simple_loss=0.3057, pruned_loss=0.08784, over 4122.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2539, pruned_loss=0.058, over 954619.54 frames. ], batch size: 65, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:29:15,410 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:29:17,730 INFO [finetune.py:976] (2/7) Epoch 13, batch 3400, loss[loss=0.1381, simple_loss=0.22, pruned_loss=0.02812, over 4747.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2563, pruned_loss=0.05843, over 957141.40 frames. ], batch size: 27, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:29:24,412 INFO [optim.py:369] (2/7) 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 04:29:40,671 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 04:29:49,104 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1066, 1.3787, 1.8436, 2.3562, 2.0270, 1.4962, 1.1705, 1.6325], + device='cuda:2'), covar=tensor([0.3656, 0.4192, 0.2075, 0.2724, 0.2914, 0.2954, 0.4530, 0.2429], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0247, 0.0223, 0.0316, 0.0214, 0.0229, 0.0229, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 04:29:51,377 INFO [finetune.py:976] (2/7) Epoch 13, batch 3450, loss[loss=0.1952, simple_loss=0.2553, pruned_loss=0.06756, over 4862.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2562, pruned_loss=0.05869, over 958045.00 frames. ], batch size: 34, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:29:55,711 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72190.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:30:13,361 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6443, 1.7963, 1.8776, 1.4460, 1.8660, 1.5092, 2.3630, 1.6226], + device='cuda:2'), covar=tensor([0.3164, 0.1389, 0.3797, 0.2298, 0.1239, 0.1974, 0.1303, 0.4046], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0341, 0.0423, 0.0351, 0.0375, 0.0378, 0.0368, 0.0413], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:30:24,562 INFO [finetune.py:976] (2/7) Epoch 13, batch 3500, loss[loss=0.173, simple_loss=0.2408, pruned_loss=0.05262, over 3997.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2525, pruned_loss=0.05731, over 957595.52 frames. ], batch size: 17, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:30:31,098 INFO [optim.py:369] (2/7) 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 04:30:58,758 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6100, 0.6614, 1.4922, 2.0012, 1.6825, 1.5000, 1.5213, 1.5203], + device='cuda:2'), covar=tensor([0.4676, 0.6110, 0.6049, 0.6354, 0.5528, 0.7733, 0.7505, 0.8015], + device='cuda:2'), in_proj_covar=tensor([0.0409, 0.0406, 0.0493, 0.0511, 0.0439, 0.0461, 0.0468, 0.0470], + device='cuda:2'), out_proj_covar=tensor([9.9199e-05, 1.0038e-04, 1.1085e-04, 1.2131e-04, 1.0615e-04, 1.1118e-04, + 1.1173e-04, 1.1255e-04], device='cuda:2') +2023-04-27 04:31:08,030 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:31:09,865 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2860, 1.2702, 3.8350, 3.5278, 3.4324, 3.6600, 3.6569, 3.3860], + device='cuda:2'), covar=tensor([0.7746, 0.5952, 0.1369, 0.2400, 0.1297, 0.1906, 0.1513, 0.1598], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0304, 0.0400, 0.0405, 0.0345, 0.0404, 0.0311, 0.0362], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:31:09,867 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72275.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:31:20,404 INFO [finetune.py:976] (2/7) Epoch 13, batch 3550, loss[loss=0.1623, simple_loss=0.2331, pruned_loss=0.04573, over 4909.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2503, pruned_loss=0.05704, over 956680.30 frames. ], batch size: 35, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:31:21,320 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 04:32:06,131 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:32:16,762 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 3600, loss[loss=0.14, simple_loss=0.2156, pruned_loss=0.03216, over 4934.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2485, pruned_loss=0.05661, over 956328.95 frames. ], batch size: 38, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:32:33,138 INFO [optim.py:369] (2/7) 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 04:32:59,908 INFO [finetune.py:976] (2/7) Epoch 13, batch 3650, loss[loss=0.1385, simple_loss=0.2105, pruned_loss=0.03328, over 4769.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2505, pruned_loss=0.05721, over 955632.80 frames. ], batch size: 26, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:33:02,556 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72387.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:33:11,911 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6595, 1.6924, 1.6359, 1.2310, 1.7731, 1.4075, 2.2305, 1.4127], + device='cuda:2'), covar=tensor([0.3763, 0.1691, 0.5250, 0.3042, 0.1553, 0.2416, 0.1423, 0.4980], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0343, 0.0425, 0.0354, 0.0377, 0.0379, 0.0370, 0.0415], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:33:33,725 INFO [finetune.py:976] (2/7) Epoch 13, batch 3700, loss[loss=0.2384, simple_loss=0.3092, pruned_loss=0.08385, over 4819.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2531, pruned_loss=0.05802, over 954475.00 frames. ], batch size: 45, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:33:36,427 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.96 vs. limit=5.0 +2023-04-27 04:33:40,456 INFO [optim.py:369] (2/7) 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 04:33:43,014 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8718, 1.6565, 2.0666, 2.2591, 1.9677, 1.7606, 1.8892, 1.8565], + device='cuda:2'), covar=tensor([0.5465, 0.7435, 0.7517, 0.7231, 0.6588, 0.9977, 1.0209, 1.0438], + device='cuda:2'), in_proj_covar=tensor([0.0410, 0.0406, 0.0494, 0.0511, 0.0440, 0.0462, 0.0468, 0.0471], + device='cuda:2'), out_proj_covar=tensor([9.9410e-05, 1.0054e-04, 1.1109e-04, 1.2136e-04, 1.0631e-04, 1.1122e-04, + 1.1175e-04, 1.1272e-04], device='cuda:2') +2023-04-27 04:33:50,300 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2111, 1.2796, 1.3963, 1.6159, 1.5895, 1.1823, 0.8590, 1.4451], + device='cuda:2'), covar=tensor([0.0912, 0.1287, 0.0882, 0.0574, 0.0694, 0.0914, 0.0918, 0.0582], + device='cuda:2'), in_proj_covar=tensor([0.0190, 0.0200, 0.0180, 0.0172, 0.0177, 0.0182, 0.0154, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:33:52,683 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0462, 2.4443, 2.4964, 2.6839, 2.5764, 2.6452, 2.3468, 4.9662], + device='cuda:2'), covar=tensor([0.0500, 0.0665, 0.0676, 0.0994, 0.0525, 0.0432, 0.0606, 0.0114], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 04:33:58,576 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2317, 2.2004, 1.8058, 1.8675, 2.3192, 1.8172, 2.7103, 1.5610], + device='cuda:2'), covar=tensor([0.3687, 0.1756, 0.4816, 0.3032, 0.1648, 0.2565, 0.1424, 0.4632], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0344, 0.0426, 0.0354, 0.0379, 0.0380, 0.0372, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:34:07,019 INFO [finetune.py:976] (2/7) Epoch 13, batch 3750, loss[loss=0.1695, simple_loss=0.2458, pruned_loss=0.04656, over 4774.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.255, pruned_loss=0.05821, over 955727.37 frames. ], batch size: 29, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:34:08,311 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72485.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:34:28,891 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1296, 2.5579, 0.7660, 1.4820, 1.5180, 1.8545, 1.5955, 0.7841], + device='cuda:2'), covar=tensor([0.1468, 0.1305, 0.1836, 0.1325, 0.1090, 0.0923, 0.1553, 0.1682], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0246, 0.0139, 0.0121, 0.0132, 0.0152, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:34:39,242 INFO [finetune.py:976] (2/7) Epoch 13, batch 3800, loss[loss=0.1905, simple_loss=0.2457, pruned_loss=0.06762, over 4800.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2555, pruned_loss=0.05828, over 954945.37 frames. ], batch size: 25, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:34:43,505 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8701, 2.4648, 1.9909, 2.2289, 1.6018, 1.9753, 1.9386, 1.5807], + device='cuda:2'), covar=tensor([0.2040, 0.1292, 0.0909, 0.1255, 0.3720, 0.1145, 0.2084, 0.2746], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0313, 0.0226, 0.0285, 0.0314, 0.0267, 0.0257, 0.0272], + device='cuda:2'), out_proj_covar=tensor([1.1843e-04, 1.2480e-04, 9.0442e-05, 1.1388e-04, 1.2800e-04, 1.0659e-04, + 1.0399e-04, 1.0884e-04], device='cuda:2') +2023-04-27 04:34:46,447 INFO [optim.py:369] (2/7) 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 04:35:05,813 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 3850, loss[loss=0.2, simple_loss=0.2525, pruned_loss=0.07368, over 4806.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.253, pruned_loss=0.05693, over 956801.13 frames. ], batch size: 45, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:35:37,011 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72623.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:35:43,533 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8957, 2.3117, 1.8597, 1.6845, 1.3911, 1.3915, 1.8327, 1.3499], + device='cuda:2'), covar=tensor([0.1747, 0.1394, 0.1459, 0.1836, 0.2427, 0.2037, 0.1073, 0.2048], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0204, 0.0202, 0.0184, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 04:35:44,026 INFO [finetune.py:976] (2/7) Epoch 13, batch 3900, loss[loss=0.1926, simple_loss=0.2505, pruned_loss=0.06732, over 4754.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2502, pruned_loss=0.05607, over 956455.01 frames. ], batch size: 27, lr: 3.59e-03, grad_scale: 32.0 +2023-04-27 04:35:45,124 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.7122, 4.7181, 3.2724, 5.4570, 4.8332, 4.7056, 2.3002, 4.5806], + device='cuda:2'), covar=tensor([0.1468, 0.0943, 0.2823, 0.0975, 0.3181, 0.1792, 0.5579, 0.2170], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0214, 0.0248, 0.0301, 0.0295, 0.0248, 0.0269, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:35:50,924 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6037, 0.9536, 1.6429, 2.0522, 1.6716, 1.5358, 1.5854, 1.5919], + device='cuda:2'), covar=tensor([0.4857, 0.6740, 0.6402, 0.6481, 0.6277, 0.7929, 0.7776, 0.8520], + device='cuda:2'), in_proj_covar=tensor([0.0410, 0.0406, 0.0494, 0.0512, 0.0441, 0.0462, 0.0468, 0.0472], + device='cuda:2'), out_proj_covar=tensor([9.9449e-05, 1.0040e-04, 1.1120e-04, 1.2146e-04, 1.0651e-04, 1.1132e-04, + 1.1180e-04, 1.1284e-04], device='cuda:2') +2023-04-27 04:35:51,961 INFO [optim.py:369] (2/7) 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 04:35:54,991 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0117, 3.9941, 3.0171, 4.6636, 4.0678, 4.0340, 2.1879, 3.9495], + device='cuda:2'), covar=tensor([0.1809, 0.0915, 0.2502, 0.1449, 0.3215, 0.1891, 0.4929, 0.2506], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0214, 0.0249, 0.0302, 0.0295, 0.0248, 0.0270, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:35:56,880 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7919, 2.1401, 1.7467, 1.3840, 1.3358, 1.3147, 1.7508, 1.2700], + device='cuda:2'), covar=tensor([0.1883, 0.1535, 0.1593, 0.2059, 0.2585, 0.2251, 0.1169, 0.2214], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0214, 0.0169, 0.0204, 0.0202, 0.0184, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 04:36:32,469 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 3950, loss[loss=0.1549, simple_loss=0.2185, pruned_loss=0.04564, over 4901.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.248, pruned_loss=0.05515, over 957057.79 frames. ], batch size: 46, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:36:42,842 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2193, 1.6043, 1.6090, 1.8365, 1.6306, 1.9859, 1.3757, 3.4347], + device='cuda:2'), covar=tensor([0.0666, 0.0821, 0.0789, 0.1192, 0.0662, 0.0488, 0.0772, 0.0137], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 04:37:16,106 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9099, 1.6422, 2.1628, 2.3546, 1.9843, 1.8473, 1.9986, 1.9534], + device='cuda:2'), covar=tensor([0.5287, 0.7502, 0.7522, 0.6209, 0.6488, 0.8834, 0.9006, 0.9854], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0407, 0.0495, 0.0513, 0.0441, 0.0463, 0.0469, 0.0473], + device='cuda:2'), out_proj_covar=tensor([9.9745e-05, 1.0070e-04, 1.1151e-04, 1.2175e-04, 1.0658e-04, 1.1166e-04, + 1.1213e-04, 1.1304e-04], device='cuda:2') +2023-04-27 04:37:38,719 INFO [finetune.py:976] (2/7) Epoch 13, batch 4000, loss[loss=0.1535, simple_loss=0.2275, pruned_loss=0.03975, over 4832.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2463, pruned_loss=0.05431, over 956739.34 frames. ], batch size: 30, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:37:57,171 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 13, batch 4050, loss[loss=0.1636, simple_loss=0.2232, pruned_loss=0.05204, over 4719.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2502, pruned_loss=0.05592, over 957023.88 frames. ], batch size: 23, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:38:29,181 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72785.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:38:48,785 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6301, 1.2306, 1.2378, 1.4225, 1.8389, 1.4234, 1.2100, 1.1930], + device='cuda:2'), covar=tensor([0.1442, 0.1535, 0.2111, 0.1338, 0.0941, 0.1953, 0.2133, 0.2110], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0317, 0.0352, 0.0293, 0.0330, 0.0315, 0.0306, 0.0362], + device='cuda:2'), out_proj_covar=tensor([6.3728e-05, 6.6624e-05, 7.5545e-05, 5.9930e-05, 6.8898e-05, 6.6833e-05, + 6.5047e-05, 7.7430e-05], device='cuda:2') +2023-04-27 04:38:53,055 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 04:38:54,879 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 04:38:57,777 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3074, 2.0520, 2.4088, 2.6913, 2.6798, 2.2601, 1.7114, 2.3814], + device='cuda:2'), covar=tensor([0.0897, 0.1118, 0.0531, 0.0598, 0.0666, 0.0892, 0.0902, 0.0573], + device='cuda:2'), in_proj_covar=tensor([0.0190, 0.0202, 0.0181, 0.0172, 0.0177, 0.0182, 0.0155, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:39:01,350 INFO [finetune.py:976] (2/7) Epoch 13, batch 4100, loss[loss=0.1513, simple_loss=0.2277, pruned_loss=0.03741, over 4788.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2525, pruned_loss=0.05664, over 956290.77 frames. ], batch size: 29, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:39:01,406 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 04:39:24,592 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7366, 1.3893, 1.8771, 2.2850, 1.8713, 1.7054, 1.8201, 1.7721], + device='cuda:2'), covar=tensor([0.5079, 0.7468, 0.7214, 0.6252, 0.6691, 0.8697, 0.8818, 0.9089], + device='cuda:2'), in_proj_covar=tensor([0.0412, 0.0407, 0.0496, 0.0513, 0.0442, 0.0462, 0.0469, 0.0473], + device='cuda:2'), out_proj_covar=tensor([9.9905e-05, 1.0079e-04, 1.1168e-04, 1.2173e-04, 1.0678e-04, 1.1143e-04, + 1.1203e-04, 1.1316e-04], device='cuda:2') +2023-04-27 04:39:34,765 INFO [finetune.py:976] (2/7) Epoch 13, batch 4150, loss[loss=0.2343, simple_loss=0.3026, pruned_loss=0.08297, over 4827.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2537, pruned_loss=0.05741, over 955148.74 frames. ], batch size: 47, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:39:42,044 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72894.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:39:49,442 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8127, 2.2764, 1.9474, 2.1867, 1.6097, 1.8949, 1.9228, 1.4770], + device='cuda:2'), covar=tensor([0.2141, 0.1396, 0.0894, 0.1261, 0.3481, 0.1239, 0.2012, 0.2892], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0314, 0.0228, 0.0286, 0.0316, 0.0268, 0.0257, 0.0273], + device='cuda:2'), out_proj_covar=tensor([1.1899e-04, 1.2529e-04, 9.1089e-05, 1.1430e-04, 1.2875e-04, 1.0719e-04, + 1.0411e-04, 1.0926e-04], device='cuda:2') +2023-04-27 04:40:08,512 INFO [finetune.py:976] (2/7) Epoch 13, batch 4200, loss[loss=0.1693, simple_loss=0.2467, pruned_loss=0.04601, over 4758.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2539, pruned_loss=0.05717, over 956118.97 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:40:08,670 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6366, 1.1592, 1.6981, 2.0379, 1.6920, 1.5867, 1.6239, 1.6970], + device='cuda:2'), covar=tensor([0.6580, 0.9045, 0.9273, 1.0239, 0.8315, 1.0428, 1.0531, 1.0333], + device='cuda:2'), in_proj_covar=tensor([0.0411, 0.0407, 0.0495, 0.0512, 0.0441, 0.0462, 0.0469, 0.0473], + device='cuda:2'), out_proj_covar=tensor([9.9753e-05, 1.0084e-04, 1.1153e-04, 1.2156e-04, 1.0665e-04, 1.1129e-04, + 1.1200e-04, 1.1307e-04], device='cuda:2') +2023-04-27 04:40:15,130 INFO [optim.py:369] (2/7) 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 04:40:23,781 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72955.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:40:26,740 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1099, 2.3623, 2.1732, 2.3430, 2.0526, 2.3146, 2.3255, 2.2035], + device='cuda:2'), covar=tensor([0.3821, 0.6553, 0.5753, 0.5449, 0.6421, 0.7941, 0.7399, 0.6288], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0377, 0.0319, 0.0330, 0.0343, 0.0398, 0.0359, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:40:30,253 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72964.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 04:40:41,159 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 4250, loss[loss=0.1943, simple_loss=0.2579, pruned_loss=0.06538, over 4313.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2526, pruned_loss=0.05732, over 956485.80 frames. ], batch size: 65, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:41:10,641 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73025.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 04:41:13,602 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73030.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:41:15,356 INFO [finetune.py:976] (2/7) Epoch 13, batch 4300, loss[loss=0.1999, simple_loss=0.2511, pruned_loss=0.07435, over 4252.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.25, pruned_loss=0.05654, over 955804.30 frames. ], batch size: 18, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:41:22,018 INFO [optim.py:369] (2/7) 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 04:41:51,108 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9037, 1.4423, 5.3571, 5.0131, 4.6841, 5.0893, 4.6547, 4.6902], + device='cuda:2'), covar=tensor([0.6704, 0.6099, 0.0776, 0.1410, 0.0889, 0.1773, 0.1135, 0.1235], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0304, 0.0401, 0.0405, 0.0344, 0.0402, 0.0311, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:41:59,416 INFO [finetune.py:976] (2/7) Epoch 13, batch 4350, loss[loss=0.1411, simple_loss=0.2157, pruned_loss=0.03328, over 4827.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.247, pruned_loss=0.05519, over 955979.15 frames. ], batch size: 30, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:42:38,505 INFO [finetune.py:976] (2/7) Epoch 13, batch 4400, loss[loss=0.2223, simple_loss=0.2963, pruned_loss=0.07416, over 4841.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2487, pruned_loss=0.05609, over 955852.79 frames. ], batch size: 49, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:42:47,738 INFO [optim.py:369] (2/7) 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 04:42:58,541 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 04:43:29,111 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5323, 1.5298, 0.5944, 1.2307, 1.4068, 1.3728, 1.2651, 1.3586], + device='cuda:2'), covar=tensor([0.0521, 0.0392, 0.0398, 0.0600, 0.0291, 0.0570, 0.0516, 0.0612], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 04:43:38,136 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6354, 3.6043, 2.5914, 4.1932, 3.7081, 3.6086, 1.5991, 3.4798], + device='cuda:2'), covar=tensor([0.1718, 0.1129, 0.3248, 0.1826, 0.3275, 0.2101, 0.5900, 0.2595], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0214, 0.0249, 0.0303, 0.0296, 0.0248, 0.0271, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:43:39,918 INFO [finetune.py:976] (2/7) Epoch 13, batch 4450, loss[loss=0.1816, simple_loss=0.2562, pruned_loss=0.05348, over 4759.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2524, pruned_loss=0.05733, over 955785.54 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:44:19,679 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-04-27 04:44:30,379 INFO [finetune.py:976] (2/7) Epoch 13, batch 4500, loss[loss=0.2134, simple_loss=0.2958, pruned_loss=0.06547, over 4801.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2549, pruned_loss=0.05826, over 955055.42 frames. ], batch size: 51, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:44:33,083 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.25 vs. limit=5.0 +2023-04-27 04:44:37,125 INFO [optim.py:369] (2/7) 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 04:44:40,828 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73250.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:44:46,379 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8646, 1.7175, 1.0618, 1.5130, 1.8179, 1.6710, 1.5404, 1.5975], + device='cuda:2'), covar=tensor([0.0479, 0.0361, 0.0336, 0.0538, 0.0249, 0.0515, 0.0519, 0.0560], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0037, 0.0050, 0.0037, 0.0048, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 04:45:04,241 INFO [finetune.py:976] (2/7) Epoch 13, batch 4550, loss[loss=0.2474, simple_loss=0.2958, pruned_loss=0.09946, over 4730.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.256, pruned_loss=0.05892, over 954488.77 frames. ], batch size: 59, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:45:10,384 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:45:28,147 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73320.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 04:45:37,747 INFO [finetune.py:976] (2/7) Epoch 13, batch 4600, loss[loss=0.2019, simple_loss=0.2638, pruned_loss=0.07002, over 4821.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2545, pruned_loss=0.05802, over 954300.07 frames. ], batch size: 33, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:45:44,461 INFO [optim.py:369] (2/7) 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 04:45:50,707 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73354.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:46:11,138 INFO [finetune.py:976] (2/7) Epoch 13, batch 4650, loss[loss=0.1636, simple_loss=0.2351, pruned_loss=0.04603, over 4826.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2525, pruned_loss=0.05786, over 954709.23 frames. ], batch size: 33, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:46:44,503 INFO [finetune.py:976] (2/7) Epoch 13, batch 4700, loss[loss=0.1192, simple_loss=0.1955, pruned_loss=0.02148, over 4940.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2502, pruned_loss=0.05746, over 953511.68 frames. ], batch size: 38, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:46:47,184 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 04:46:51,528 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 13, batch 4750, loss[loss=0.139, simple_loss=0.2157, pruned_loss=0.03116, over 4777.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2473, pruned_loss=0.05619, over 953070.11 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:48:11,945 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6777, 1.5245, 1.9411, 1.9975, 1.4877, 1.3302, 1.5983, 1.1007], + device='cuda:2'), covar=tensor([0.0560, 0.0848, 0.0403, 0.0780, 0.0835, 0.1201, 0.0672, 0.0683], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0067, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:48:26,415 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5098, 1.9732, 1.3800, 1.1077, 1.1307, 1.1221, 1.3027, 1.0182], + device='cuda:2'), covar=tensor([0.2018, 0.1520, 0.1840, 0.2233, 0.2630, 0.2457, 0.1249, 0.2286], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0213, 0.0168, 0.0204, 0.0201, 0.0183, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 04:48:28,608 INFO [finetune.py:976] (2/7) Epoch 13, batch 4800, loss[loss=0.2423, simple_loss=0.3157, pruned_loss=0.08447, over 4810.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2491, pruned_loss=0.05661, over 954394.24 frames. ], batch size: 45, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:48:36,792 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73550.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:48:43,579 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 04:49:02,204 INFO [finetune.py:976] (2/7) Epoch 13, batch 4850, loss[loss=0.1832, simple_loss=0.2543, pruned_loss=0.05604, over 4816.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2523, pruned_loss=0.05754, over 956125.66 frames. ], batch size: 38, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:49:17,655 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73590.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:49:28,462 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73598.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:49:43,582 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3899, 1.3217, 1.3827, 1.0314, 1.3524, 1.1593, 1.7436, 1.1851], + device='cuda:2'), covar=tensor([0.3462, 0.1730, 0.5148, 0.2612, 0.1569, 0.2106, 0.1556, 0.5024], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0350, 0.0429, 0.0357, 0.0385, 0.0385, 0.0375, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:49:50,437 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3099, 1.7846, 1.6558, 2.1234, 1.9225, 2.2076, 1.7213, 4.5005], + device='cuda:2'), covar=tensor([0.0580, 0.0808, 0.0817, 0.1135, 0.0638, 0.0474, 0.0712, 0.0105], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 04:49:53,458 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73620.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 04:50:13,630 INFO [finetune.py:976] (2/7) Epoch 13, batch 4900, loss[loss=0.201, simple_loss=0.2808, pruned_loss=0.06055, over 4892.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2543, pruned_loss=0.05826, over 954887.20 frames. ], batch size: 35, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:50:28,449 INFO [optim.py:369] (2/7) 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 04:50:38,435 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73649.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:50:39,755 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73651.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:50:52,690 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3526, 1.5739, 1.7425, 1.8710, 1.7576, 1.8459, 1.8138, 1.7552], + device='cuda:2'), covar=tensor([0.4098, 0.5725, 0.4987, 0.4721, 0.5684, 0.7451, 0.5838, 0.5306], + device='cuda:2'), in_proj_covar=tensor([0.0327, 0.0372, 0.0314, 0.0327, 0.0338, 0.0394, 0.0354, 0.0321], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:50:55,037 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 4950, loss[loss=0.1792, simple_loss=0.2578, pruned_loss=0.05027, over 4923.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2554, pruned_loss=0.05808, over 955714.75 frames. ], batch size: 38, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:51:24,376 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4551, 1.7510, 1.8462, 1.9716, 1.8228, 1.9462, 1.9294, 1.8871], + device='cuda:2'), covar=tensor([0.4718, 0.5786, 0.5137, 0.4778, 0.5729, 0.7771, 0.5783, 0.5448], + device='cuda:2'), in_proj_covar=tensor([0.0328, 0.0373, 0.0315, 0.0327, 0.0339, 0.0395, 0.0354, 0.0322], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:51:36,574 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 5000, loss[loss=0.1993, simple_loss=0.2628, pruned_loss=0.0679, over 4117.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2533, pruned_loss=0.05728, over 955193.68 frames. ], batch size: 65, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:51:45,208 INFO [optim.py:369] (2/7) 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 04:52:11,152 INFO [finetune.py:976] (2/7) Epoch 13, batch 5050, loss[loss=0.1614, simple_loss=0.2269, pruned_loss=0.04796, over 4902.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2507, pruned_loss=0.05675, over 956507.17 frames. ], batch size: 43, lr: 3.58e-03, grad_scale: 32.0 +2023-04-27 04:52:17,807 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73793.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:52:31,001 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9719, 2.5874, 2.1228, 2.4023, 1.7548, 2.1199, 2.1601, 1.6650], + device='cuda:2'), covar=tensor([0.1824, 0.1122, 0.0818, 0.1101, 0.3305, 0.1115, 0.1826, 0.2515], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0309, 0.0225, 0.0284, 0.0313, 0.0266, 0.0255, 0.0271], + device='cuda:2'), out_proj_covar=tensor([1.1734e-04, 1.2304e-04, 8.9854e-05, 1.1345e-04, 1.2764e-04, 1.0645e-04, + 1.0339e-04, 1.0839e-04], device='cuda:2') +2023-04-27 04:52:56,560 INFO [finetune.py:976] (2/7) Epoch 13, batch 5100, loss[loss=0.1494, simple_loss=0.2152, pruned_loss=0.04183, over 4829.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2483, pruned_loss=0.05601, over 955702.98 frames. ], batch size: 30, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:53:03,638 INFO [optim.py:369] (2/7) 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 04:53:44,072 INFO [finetune.py:976] (2/7) Epoch 13, batch 5150, loss[loss=0.2079, simple_loss=0.2685, pruned_loss=0.07371, over 4904.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2487, pruned_loss=0.05651, over 956717.76 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:54:10,824 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4861, 1.2862, 4.3517, 4.1191, 3.8054, 4.0968, 3.9173, 3.8194], + device='cuda:2'), covar=tensor([0.6628, 0.5846, 0.0945, 0.1422, 0.1000, 0.1513, 0.1647, 0.1363], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0301, 0.0396, 0.0401, 0.0339, 0.0398, 0.0308, 0.0360], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 04:54:18,052 INFO [finetune.py:976] (2/7) Epoch 13, batch 5200, loss[loss=0.2179, simple_loss=0.3006, pruned_loss=0.06765, over 4848.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2525, pruned_loss=0.05766, over 955929.60 frames. ], batch size: 44, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:54:24,749 INFO [optim.py:369] (2/7) 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 04:54:26,033 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73946.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:54:28,360 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73949.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:55:05,874 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9353, 2.2119, 2.0714, 2.2773, 2.0146, 2.1717, 2.1492, 2.0589], + device='cuda:2'), covar=tensor([0.4657, 0.6567, 0.5789, 0.4966, 0.6369, 0.7951, 0.6756, 0.6141], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0376, 0.0316, 0.0329, 0.0341, 0.0397, 0.0356, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:55:06,580 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-04-27 04:55:07,559 INFO [finetune.py:976] (2/7) Epoch 13, batch 5250, loss[loss=0.1828, simple_loss=0.2634, pruned_loss=0.05109, over 4904.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.254, pruned_loss=0.05724, over 955070.71 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 04:55:16,183 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73997.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:55:42,305 INFO [finetune.py:976] (2/7) Epoch 13, batch 5300, loss[loss=0.206, simple_loss=0.2794, pruned_loss=0.06627, over 4773.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2553, pruned_loss=0.05791, over 953941.63 frames. ], batch size: 51, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:55:54,436 INFO [optim.py:369] (2/7) 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 04:56:19,395 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74064.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:56:36,838 INFO [finetune.py:976] (2/7) Epoch 13, batch 5350, loss[loss=0.2043, simple_loss=0.267, pruned_loss=0.07079, over 4887.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2556, pruned_loss=0.05784, over 953899.76 frames. ], batch size: 32, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:56:39,887 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:57:05,261 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 13, batch 5400, loss[loss=0.159, simple_loss=0.232, pruned_loss=0.04295, over 4746.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2523, pruned_loss=0.05672, over 954612.99 frames. ], batch size: 26, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:57:17,119 INFO [optim.py:369] (2/7) 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 04:57:43,730 INFO [finetune.py:976] (2/7) Epoch 13, batch 5450, loss[loss=0.1743, simple_loss=0.2436, pruned_loss=0.05256, over 4907.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2492, pruned_loss=0.05561, over 954781.45 frames. ], batch size: 36, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:57:57,100 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5820, 3.1571, 1.0023, 1.8553, 2.6062, 1.6172, 4.4428, 2.5758], + device='cuda:2'), covar=tensor([0.0597, 0.0578, 0.0789, 0.1244, 0.0461, 0.0950, 0.0235, 0.0508], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0065, 0.0048, 0.0046, 0.0050, 0.0051, 0.0075, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 04:58:07,165 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1673, 1.4425, 1.6045, 1.7362, 1.6577, 1.7945, 1.7451, 1.6540], + device='cuda:2'), covar=tensor([0.4197, 0.5341, 0.4638, 0.4517, 0.5748, 0.7732, 0.5017, 0.5103], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0375, 0.0317, 0.0330, 0.0341, 0.0398, 0.0357, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 04:58:35,028 INFO [finetune.py:976] (2/7) Epoch 13, batch 5500, loss[loss=0.1794, simple_loss=0.2518, pruned_loss=0.05351, over 4842.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2459, pruned_loss=0.05426, over 954990.09 frames. ], batch size: 44, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:58:39,429 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 04:58:43,635 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74246.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:59:08,625 INFO [finetune.py:976] (2/7) Epoch 13, batch 5550, loss[loss=0.1631, simple_loss=0.2353, pruned_loss=0.04546, over 4886.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2482, pruned_loss=0.05578, over 952745.56 frames. ], batch size: 32, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:59:15,401 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:59:20,224 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74301.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 04:59:36,774 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0808, 2.5125, 0.7341, 1.3528, 1.4432, 1.8733, 1.5427, 0.8093], + device='cuda:2'), covar=tensor([0.1586, 0.1252, 0.1976, 0.1472, 0.1220, 0.1024, 0.1657, 0.1828], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0138, 0.0121, 0.0132, 0.0152, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 04:59:39,647 INFO [finetune.py:976] (2/7) Epoch 13, batch 5600, loss[loss=0.1867, simple_loss=0.2628, pruned_loss=0.05533, over 4805.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2516, pruned_loss=0.05655, over 952815.17 frames. ], batch size: 45, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 04:59:46,068 INFO [optim.py:369] (2/7) 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 04:59:57,699 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5227, 1.3803, 0.6636, 1.2866, 1.5642, 1.4292, 1.3560, 1.3879], + device='cuda:2'), covar=tensor([0.0519, 0.0393, 0.0409, 0.0558, 0.0312, 0.0517, 0.0521, 0.0596], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 05:00:03,497 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3541, 1.3080, 1.3662, 0.9317, 1.2576, 1.1746, 1.6931, 1.1307], + device='cuda:2'), covar=tensor([0.3662, 0.1862, 0.5014, 0.2892, 0.1769, 0.2330, 0.1631, 0.5227], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0343, 0.0423, 0.0352, 0.0379, 0.0379, 0.0369, 0.0414], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:00:16,164 INFO [finetune.py:976] (2/7) Epoch 13, batch 5650, loss[loss=0.1616, simple_loss=0.2234, pruned_loss=0.04987, over 4747.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2544, pruned_loss=0.05742, over 952765.08 frames. ], batch size: 23, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 05:00:25,011 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74388.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:00:43,026 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.86 vs. limit=5.0 +2023-04-27 05:00:44,669 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74420.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:00:52,321 INFO [finetune.py:976] (2/7) Epoch 13, batch 5700, loss[loss=0.2086, simple_loss=0.2529, pruned_loss=0.08212, over 4052.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2507, pruned_loss=0.05737, over 933123.85 frames. ], batch size: 17, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 05:00:54,135 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74436.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:00:58,755 INFO [optim.py:369] (2/7) 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 05:01:05,381 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8685, 3.6634, 3.1765, 3.2566, 2.5569, 3.1097, 3.1963, 2.6033], + device='cuda:2'), covar=tensor([0.1929, 0.1104, 0.0574, 0.1250, 0.2799, 0.0978, 0.1699, 0.2243], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0314, 0.0228, 0.0289, 0.0317, 0.0270, 0.0260, 0.0276], + device='cuda:2'), out_proj_covar=tensor([1.1928e-04, 1.2502e-04, 9.1260e-05, 1.1520e-04, 1.2927e-04, 1.0805e-04, + 1.0514e-04, 1.1022e-04], device='cuda:2') +2023-04-27 05:01:23,649 INFO [finetune.py:976] (2/7) Epoch 14, batch 0, loss[loss=0.1984, simple_loss=0.2623, pruned_loss=0.06727, over 4865.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2623, pruned_loss=0.06727, over 4865.00 frames. ], batch size: 31, lr: 3.57e-03, grad_scale: 64.0 +2023-04-27 05:01:23,649 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 05:01:26,967 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2798, 2.6808, 1.0453, 1.3735, 2.0135, 1.3535, 3.0694, 1.7982], + device='cuda:2'), covar=tensor([0.0574, 0.0535, 0.0676, 0.1190, 0.0424, 0.0891, 0.0275, 0.0535], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0066, 0.0048, 0.0046, 0.0050, 0.0051, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 05:01:32,467 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2691, 1.5512, 1.4199, 1.8226, 1.6483, 1.8546, 1.4123, 3.0883], + device='cuda:2'), covar=tensor([0.0603, 0.0825, 0.0798, 0.1257, 0.0639, 0.0428, 0.0731, 0.0208], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 05:01:32,577 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3367, 2.7134, 1.0830, 1.4715, 2.0370, 1.4214, 3.0997, 1.8331], + device='cuda:2'), covar=tensor([0.0597, 0.0571, 0.0650, 0.1147, 0.0412, 0.0846, 0.0285, 0.0542], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0066, 0.0048, 0.0046, 0.0050, 0.0051, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 05:01:45,227 INFO [finetune.py:1010] (2/7) Epoch 14, validation: loss=0.1535, simple_loss=0.226, pruned_loss=0.04054, over 2265189.00 frames. +2023-04-27 05:01:45,228 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 05:02:09,285 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2606, 2.9089, 1.0301, 1.5345, 1.7097, 2.0790, 1.7431, 0.9739], + device='cuda:2'), covar=tensor([0.1673, 0.1675, 0.2002, 0.1542, 0.1294, 0.1296, 0.1738, 0.2161], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0138, 0.0121, 0.0132, 0.0152, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:02:18,139 INFO [finetune.py:976] (2/7) Epoch 14, batch 50, loss[loss=0.1455, simple_loss=0.2126, pruned_loss=0.03916, over 4702.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2537, pruned_loss=0.05731, over 217931.12 frames. ], batch size: 23, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:02:41,160 INFO [optim.py:369] (2/7) 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 05:02:51,799 INFO [finetune.py:976] (2/7) Epoch 14, batch 100, loss[loss=0.2335, simple_loss=0.293, pruned_loss=0.08699, over 4896.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2506, pruned_loss=0.05791, over 380307.10 frames. ], batch size: 36, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:03:09,808 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-04-27 05:03:29,442 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74595.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:03:30,031 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74596.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:03:41,542 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1428, 2.6922, 2.3128, 2.5168, 1.9340, 2.2020, 2.3184, 1.8786], + device='cuda:2'), covar=tensor([0.1892, 0.1242, 0.0757, 0.1325, 0.3032, 0.1234, 0.1971, 0.2448], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0309, 0.0225, 0.0284, 0.0313, 0.0266, 0.0255, 0.0271], + device='cuda:2'), out_proj_covar=tensor([1.1708e-04, 1.2317e-04, 9.0056e-05, 1.1337e-04, 1.2737e-04, 1.0649e-04, + 1.0333e-04, 1.0824e-04], device='cuda:2') +2023-04-27 05:03:51,942 INFO [finetune.py:976] (2/7) Epoch 14, batch 150, loss[loss=0.1717, simple_loss=0.2476, pruned_loss=0.04794, over 4834.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2452, pruned_loss=0.05577, over 509310.83 frames. ], batch size: 33, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:03:56,114 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9972, 2.3785, 1.3580, 1.7157, 2.4143, 1.8373, 1.7763, 1.9656], + device='cuda:2'), covar=tensor([0.0468, 0.0325, 0.0292, 0.0527, 0.0228, 0.0509, 0.0473, 0.0513], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 05:04:13,501 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.60 vs. limit=5.0 +2023-04-27 05:04:20,059 INFO [optim.py:369] (2/7) 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 05:04:22,705 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 05:04:27,535 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 200, loss[loss=0.1873, simple_loss=0.2632, pruned_loss=0.05569, over 4842.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2439, pruned_loss=0.05519, over 609005.18 frames. ], batch size: 47, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:05:05,314 INFO [finetune.py:976] (2/7) Epoch 14, batch 250, loss[loss=0.1323, simple_loss=0.2048, pruned_loss=0.02996, over 4777.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2483, pruned_loss=0.05666, over 687537.01 frames. ], batch size: 26, lr: 3.57e-03, grad_scale: 32.0 +2023-04-27 05:05:11,853 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:05:39,270 INFO [optim.py:369] (2/7) 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 05:05:49,929 INFO [finetune.py:976] (2/7) Epoch 14, batch 300, loss[loss=0.1714, simple_loss=0.2415, pruned_loss=0.05068, over 4813.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2525, pruned_loss=0.05803, over 747594.46 frames. ], batch size: 40, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:05:50,076 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4922, 1.1556, 1.2586, 1.1633, 1.6791, 1.2903, 1.0990, 1.2403], + device='cuda:2'), covar=tensor([0.1883, 0.1591, 0.2206, 0.1664, 0.0885, 0.1591, 0.2103, 0.2261], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0316, 0.0351, 0.0292, 0.0329, 0.0314, 0.0305, 0.0362], + device='cuda:2'), out_proj_covar=tensor([6.3907e-05, 6.6354e-05, 7.5146e-05, 5.9711e-05, 6.8594e-05, 6.6543e-05, + 6.4668e-05, 7.7461e-05], device='cuda:2') +2023-04-27 05:05:54,767 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 350, loss[loss=0.1772, simple_loss=0.2525, pruned_loss=0.051, over 4785.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2546, pruned_loss=0.0588, over 794542.72 frames. ], batch size: 29, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:07:09,534 INFO [optim.py:369] (2/7) 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 05:07:25,450 INFO [finetune.py:976] (2/7) Epoch 14, batch 400, loss[loss=0.2046, simple_loss=0.2729, pruned_loss=0.06817, over 4864.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2567, pruned_loss=0.05945, over 830478.78 frames. ], batch size: 34, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:07:47,534 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-04-27 05:07:49,824 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74896.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:07:57,758 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:07:59,305 INFO [finetune.py:976] (2/7) Epoch 14, batch 450, loss[loss=0.2459, simple_loss=0.2867, pruned_loss=0.1025, over 4831.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2561, pruned_loss=0.05921, over 857857.62 frames. ], batch size: 30, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:08:37,890 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74944.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:08:38,413 INFO [optim.py:369] (2/7) 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 05:08:42,183 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 500, loss[loss=0.2803, simple_loss=0.323, pruned_loss=0.1188, over 4334.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2532, pruned_loss=0.05803, over 878991.78 frames. ], batch size: 66, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:08:56,094 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74964.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:08:56,160 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 05:08:59,769 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74970.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:09:27,831 INFO [finetune.py:976] (2/7) Epoch 14, batch 550, loss[loss=0.1936, simple_loss=0.2603, pruned_loss=0.06344, over 4837.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2503, pruned_loss=0.05705, over 895615.29 frames. ], batch size: 33, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:09:37,447 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75025.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:09:39,814 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9144, 1.5869, 1.5119, 1.6029, 2.1352, 1.7000, 1.3974, 1.3998], + device='cuda:2'), covar=tensor([0.1544, 0.1430, 0.1909, 0.1393, 0.0663, 0.1645, 0.2212, 0.2412], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0318, 0.0353, 0.0295, 0.0332, 0.0316, 0.0308, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.4304e-05, 6.6670e-05, 7.5744e-05, 6.0288e-05, 6.9296e-05, 6.7059e-05, + 6.5198e-05, 7.7714e-05], device='cuda:2') +2023-04-27 05:09:51,450 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 14, batch 600, loss[loss=0.1903, simple_loss=0.2638, pruned_loss=0.05842, over 4854.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2509, pruned_loss=0.05731, over 908214.66 frames. ], batch size: 44, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:10:35,751 INFO [finetune.py:976] (2/7) Epoch 14, batch 650, loss[loss=0.1608, simple_loss=0.2479, pruned_loss=0.03686, over 4793.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2532, pruned_loss=0.05801, over 919338.27 frames. ], batch size: 29, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:10:54,114 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8728, 1.6080, 1.8332, 2.2061, 2.3384, 1.7365, 1.4487, 1.8041], + device='cuda:2'), covar=tensor([0.0809, 0.1199, 0.0701, 0.0606, 0.0558, 0.0864, 0.0824, 0.0628], + device='cuda:2'), in_proj_covar=tensor([0.0191, 0.0203, 0.0183, 0.0173, 0.0179, 0.0184, 0.0156, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:10:59,289 INFO [optim.py:369] (2/7) 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 05:11:09,458 INFO [finetune.py:976] (2/7) Epoch 14, batch 700, loss[loss=0.1946, simple_loss=0.2635, pruned_loss=0.0629, over 4885.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2553, pruned_loss=0.05928, over 927290.08 frames. ], batch size: 32, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:11:10,780 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7186, 4.2410, 0.7087, 2.2523, 2.4108, 2.7962, 2.4752, 0.8585], + device='cuda:2'), covar=tensor([0.1374, 0.1122, 0.2292, 0.1189, 0.1010, 0.1079, 0.1424, 0.2166], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0244, 0.0137, 0.0120, 0.0132, 0.0151, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:11:46,818 INFO [finetune.py:976] (2/7) Epoch 14, batch 750, loss[loss=0.1588, simple_loss=0.2377, pruned_loss=0.03994, over 4766.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2565, pruned_loss=0.05962, over 933574.06 frames. ], batch size: 28, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:12:09,245 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75228.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:13,668 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6379, 1.3458, 1.8105, 2.0718, 1.8153, 1.5907, 1.7174, 1.6697], + device='cuda:2'), covar=tensor([0.4812, 0.6714, 0.6362, 0.6249, 0.5631, 0.8484, 0.8231, 0.8213], + device='cuda:2'), in_proj_covar=tensor([0.0415, 0.0410, 0.0497, 0.0514, 0.0445, 0.0467, 0.0472, 0.0477], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:12:31,822 INFO [optim.py:369] (2/7) 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 05:12:35,601 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75251.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:41,575 INFO [finetune.py:976] (2/7) Epoch 14, batch 800, loss[loss=0.1907, simple_loss=0.2552, pruned_loss=0.06311, over 4831.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2557, pruned_loss=0.05922, over 937417.53 frames. ], batch size: 49, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:12:44,560 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75265.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:47,615 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75270.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:12:57,719 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6686, 2.7702, 2.3907, 2.5917, 2.8457, 2.5753, 3.8822, 2.0922], + device='cuda:2'), covar=tensor([0.4553, 0.2136, 0.4431, 0.3504, 0.2161, 0.2707, 0.1513, 0.4598], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0345, 0.0424, 0.0353, 0.0379, 0.0378, 0.0369, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:13:00,651 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:13:07,658 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 850, loss[loss=0.141, simple_loss=0.2126, pruned_loss=0.03467, over 4823.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2525, pruned_loss=0.05794, over 942618.43 frames. ], batch size: 41, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:13:20,853 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:13:40,185 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75331.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:13:56,948 INFO [optim.py:369] (2/7) 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 05:13:58,407 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 05:14:12,837 INFO [finetune.py:976] (2/7) Epoch 14, batch 900, loss[loss=0.1876, simple_loss=0.263, pruned_loss=0.05608, over 4904.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2504, pruned_loss=0.05767, over 944519.44 frames. ], batch size: 37, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:14:57,513 INFO [finetune.py:976] (2/7) Epoch 14, batch 950, loss[loss=0.1807, simple_loss=0.2386, pruned_loss=0.06142, over 4156.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2497, pruned_loss=0.05809, over 944704.53 frames. ], batch size: 17, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:15:02,568 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6183, 1.6961, 1.8399, 2.3770, 2.5668, 2.0895, 2.0258, 1.9243], + device='cuda:2'), covar=tensor([0.1937, 0.2063, 0.2105, 0.2143, 0.1296, 0.2231, 0.2478, 0.2224], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0317, 0.0354, 0.0293, 0.0332, 0.0315, 0.0307, 0.0363], + device='cuda:2'), out_proj_covar=tensor([6.4060e-05, 6.6676e-05, 7.5968e-05, 5.9990e-05, 6.9202e-05, 6.6675e-05, + 6.4967e-05, 7.7640e-05], device='cuda:2') +2023-04-27 05:15:04,286 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-04-27 05:15:20,049 INFO [optim.py:369] (2/7) 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 05:15:30,322 INFO [finetune.py:976] (2/7) Epoch 14, batch 1000, loss[loss=0.1975, simple_loss=0.2773, pruned_loss=0.05888, over 4206.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2518, pruned_loss=0.05838, over 946514.20 frames. ], batch size: 65, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:15:51,485 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2521, 1.5164, 1.4209, 1.5191, 1.2897, 1.2637, 1.2853, 1.1044], + device='cuda:2'), covar=tensor([0.1925, 0.1378, 0.1067, 0.1305, 0.3673, 0.1383, 0.1875, 0.2570], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0312, 0.0228, 0.0285, 0.0315, 0.0268, 0.0256, 0.0273], + device='cuda:2'), out_proj_covar=tensor([1.1792e-04, 1.2440e-04, 9.0971e-05, 1.1362e-04, 1.2825e-04, 1.0740e-04, + 1.0390e-04, 1.0903e-04], device='cuda:2') +2023-04-27 05:16:03,284 INFO [finetune.py:976] (2/7) Epoch 14, batch 1050, loss[loss=0.1538, simple_loss=0.2419, pruned_loss=0.03279, over 4800.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2537, pruned_loss=0.05885, over 948319.29 frames. ], batch size: 29, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:16:08,845 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7136, 3.5465, 1.0598, 2.1037, 2.2492, 2.5381, 2.1733, 1.1950], + device='cuda:2'), covar=tensor([0.1268, 0.1074, 0.1925, 0.1243, 0.0891, 0.1034, 0.1353, 0.1744], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0137, 0.0121, 0.0133, 0.0152, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:16:13,638 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 05:16:25,347 INFO [optim.py:369] (2/7) 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 05:16:36,963 INFO [finetune.py:976] (2/7) Epoch 14, batch 1100, loss[loss=0.1927, simple_loss=0.26, pruned_loss=0.06268, over 4853.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2556, pruned_loss=0.05945, over 949210.94 frames. ], batch size: 44, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:16:39,530 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75565.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:16:43,186 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3980, 3.4048, 0.8854, 1.8945, 1.9313, 2.2950, 1.9590, 0.8734], + device='cuda:2'), covar=tensor([0.1450, 0.1036, 0.2024, 0.1245, 0.1077, 0.1102, 0.1531, 0.2152], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0137, 0.0121, 0.0133, 0.0152, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:16:57,407 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75584.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:16:58,660 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1202, 2.3525, 0.9696, 1.5282, 1.5094, 1.8319, 1.5409, 0.9514], + device='cuda:2'), covar=tensor([0.1298, 0.0938, 0.1469, 0.1110, 0.1033, 0.0845, 0.1429, 0.1771], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0137, 0.0121, 0.0133, 0.0152, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:17:11,245 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5807, 1.5729, 0.8576, 1.2732, 1.7876, 1.4555, 1.3760, 1.5017], + device='cuda:2'), covar=tensor([0.0503, 0.0367, 0.0345, 0.0543, 0.0272, 0.0499, 0.0484, 0.0578], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 05:17:31,382 INFO [finetune.py:976] (2/7) Epoch 14, batch 1150, loss[loss=0.1793, simple_loss=0.2485, pruned_loss=0.05503, over 4806.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2564, pruned_loss=0.05971, over 948356.59 frames. ], batch size: 51, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:17:32,670 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:17:36,932 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75620.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:17:40,634 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75626.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:17:53,272 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 14, batch 1200, loss[loss=0.2004, simple_loss=0.2552, pruned_loss=0.07278, over 4871.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2555, pruned_loss=0.05943, over 950754.04 frames. ], batch size: 31, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:18:09,670 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75668.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:18:43,841 INFO [finetune.py:976] (2/7) Epoch 14, batch 1250, loss[loss=0.1797, simple_loss=0.2483, pruned_loss=0.05552, over 4816.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2526, pruned_loss=0.05803, over 952673.07 frames. ], batch size: 38, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:19:24,353 INFO [optim.py:369] (2/7) 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 05:19:30,718 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 05:19:45,790 INFO [finetune.py:976] (2/7) Epoch 14, batch 1300, loss[loss=0.1418, simple_loss=0.2057, pruned_loss=0.03894, over 4693.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2478, pruned_loss=0.05612, over 953118.99 frames. ], batch size: 23, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:20:50,514 INFO [finetune.py:976] (2/7) Epoch 14, batch 1350, loss[loss=0.1928, simple_loss=0.2623, pruned_loss=0.06164, over 4901.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2482, pruned_loss=0.05612, over 954834.57 frames. ], batch size: 35, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:21:39,791 INFO [optim.py:369] (2/7) 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 05:21:55,187 INFO [finetune.py:976] (2/7) Epoch 14, batch 1400, loss[loss=0.2166, simple_loss=0.2866, pruned_loss=0.07325, over 4868.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2534, pruned_loss=0.05818, over 957686.19 frames. ], batch size: 31, lr: 3.56e-03, grad_scale: 32.0 +2023-04-27 05:22:34,308 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75884.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:22:50,816 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 05:22:51,835 INFO [finetune.py:976] (2/7) Epoch 14, batch 1450, loss[loss=0.1848, simple_loss=0.2617, pruned_loss=0.0539, over 4915.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2549, pruned_loss=0.05812, over 957893.98 frames. ], batch size: 38, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:23:03,119 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75926.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:23:06,746 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:23:14,850 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 05:23:25,299 INFO [finetune.py:976] (2/7) Epoch 14, batch 1500, loss[loss=0.1858, simple_loss=0.2693, pruned_loss=0.05118, over 4905.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2574, pruned_loss=0.059, over 957361.91 frames. ], batch size: 36, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:23:25,426 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4450, 1.0273, 1.2435, 1.1452, 1.6069, 1.2929, 1.0220, 1.2001], + device='cuda:2'), covar=tensor([0.1529, 0.1373, 0.1644, 0.1327, 0.0824, 0.1301, 0.2071, 0.1927], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0314, 0.0351, 0.0290, 0.0329, 0.0312, 0.0303, 0.0361], + device='cuda:2'), out_proj_covar=tensor([6.3592e-05, 6.5859e-05, 7.5301e-05, 5.9285e-05, 6.8450e-05, 6.6112e-05, + 6.4222e-05, 7.7036e-05], device='cuda:2') +2023-04-27 05:23:34,212 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:24:11,480 INFO [finetune.py:976] (2/7) Epoch 14, batch 1550, loss[loss=0.1637, simple_loss=0.2354, pruned_loss=0.04596, over 4898.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2558, pruned_loss=0.05864, over 956182.09 frames. ], batch size: 37, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:24:40,106 INFO [optim.py:369] (2/7) 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 05:24:42,809 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.04 vs. limit=5.0 +2023-04-27 05:24:54,223 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2446, 1.6570, 2.0900, 2.5511, 2.0327, 1.6261, 1.3891, 1.8150], + device='cuda:2'), covar=tensor([0.3081, 0.3345, 0.1675, 0.2390, 0.2786, 0.2811, 0.4431, 0.2275], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0249, 0.0224, 0.0318, 0.0217, 0.0231, 0.0232, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 05:25:01,381 INFO [finetune.py:976] (2/7) Epoch 14, batch 1600, loss[loss=0.1595, simple_loss=0.2283, pruned_loss=0.0453, over 4832.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2542, pruned_loss=0.05859, over 957564.00 frames. ], batch size: 30, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:25:05,807 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 05:25:34,939 INFO [finetune.py:976] (2/7) Epoch 14, batch 1650, loss[loss=0.1686, simple_loss=0.2387, pruned_loss=0.04925, over 4820.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2504, pruned_loss=0.05728, over 959032.14 frames. ], batch size: 25, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:25:58,459 INFO [optim.py:369] (2/7) 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 05:26:02,288 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9655, 2.1393, 2.1523, 2.3170, 2.1280, 2.2482, 2.2450, 2.1775], + device='cuda:2'), covar=tensor([0.4234, 0.7264, 0.5822, 0.5112, 0.5924, 0.7868, 0.7398, 0.6400], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0376, 0.0317, 0.0329, 0.0342, 0.0399, 0.0355, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 05:26:08,244 INFO [finetune.py:976] (2/7) Epoch 14, batch 1700, loss[loss=0.2231, simple_loss=0.2741, pruned_loss=0.08605, over 4739.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.249, pruned_loss=0.05716, over 958447.00 frames. ], batch size: 59, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:26:08,448 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-27 05:26:33,594 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4309, 1.2103, 4.0722, 3.7759, 3.6200, 3.8660, 3.8560, 3.5988], + device='cuda:2'), covar=tensor([0.7009, 0.6042, 0.0941, 0.1711, 0.1145, 0.1626, 0.1484, 0.1484], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0303, 0.0398, 0.0399, 0.0343, 0.0399, 0.0310, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:26:34,274 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8176, 2.1167, 1.7113, 1.4561, 1.3123, 1.3130, 1.7836, 1.2562], + device='cuda:2'), covar=tensor([0.1657, 0.1366, 0.1479, 0.1824, 0.2484, 0.1961, 0.1066, 0.2190], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0213, 0.0168, 0.0204, 0.0201, 0.0183, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 05:26:42,126 INFO [finetune.py:976] (2/7) Epoch 14, batch 1750, loss[loss=0.2292, simple_loss=0.2986, pruned_loss=0.07988, over 4814.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2507, pruned_loss=0.05723, over 958457.66 frames. ], batch size: 45, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:27:06,454 INFO [optim.py:369] (2/7) 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 05:27:14,109 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 05:27:15,746 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5611, 1.6246, 0.5589, 1.2953, 1.6855, 1.4319, 1.3544, 1.4269], + device='cuda:2'), covar=tensor([0.0533, 0.0402, 0.0419, 0.0601, 0.0292, 0.0537, 0.0532, 0.0612], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 05:27:16,247 INFO [finetune.py:976] (2/7) Epoch 14, batch 1800, loss[loss=0.1927, simple_loss=0.2794, pruned_loss=0.05302, over 4905.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2526, pruned_loss=0.05727, over 957340.61 frames. ], batch size: 36, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:27:27,566 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76269.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 05:28:02,612 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:28:12,935 INFO [finetune.py:976] (2/7) Epoch 14, batch 1850, loss[loss=0.1658, simple_loss=0.2327, pruned_loss=0.04946, over 4753.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.254, pruned_loss=0.05757, over 958037.85 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:28:20,905 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76323.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:28:23,457 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 05:28:25,682 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76330.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 05:28:36,248 INFO [optim.py:369] (2/7) 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 05:28:43,362 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76355.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:28:46,905 INFO [finetune.py:976] (2/7) Epoch 14, batch 1900, loss[loss=0.1871, simple_loss=0.259, pruned_loss=0.05763, over 4810.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2544, pruned_loss=0.05726, over 958132.86 frames. ], batch size: 41, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:29:01,973 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76384.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:29:20,622 INFO [finetune.py:976] (2/7) Epoch 14, batch 1950, loss[loss=0.1702, simple_loss=0.2471, pruned_loss=0.04663, over 4824.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2543, pruned_loss=0.05758, over 958013.89 frames. ], batch size: 38, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:29:42,670 INFO [optim.py:369] (2/7) 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 05:29:55,872 INFO [finetune.py:976] (2/7) Epoch 14, batch 2000, loss[loss=0.1956, simple_loss=0.2659, pruned_loss=0.06264, over 4869.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2513, pruned_loss=0.05654, over 957030.80 frames. ], batch size: 31, lr: 3.55e-03, grad_scale: 32.0 +2023-04-27 05:30:58,162 INFO [finetune.py:976] (2/7) Epoch 14, batch 2050, loss[loss=0.1865, simple_loss=0.2555, pruned_loss=0.05875, over 4816.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2485, pruned_loss=0.05597, over 955176.90 frames. ], batch size: 38, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:31:19,798 INFO [optim.py:369] (2/7) 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 05:31:19,923 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2776, 1.9656, 2.3639, 2.5958, 2.6220, 2.2079, 1.8125, 2.2656], + device='cuda:2'), covar=tensor([0.0850, 0.1049, 0.0602, 0.0602, 0.0586, 0.0799, 0.0766, 0.0590], + device='cuda:2'), in_proj_covar=tensor([0.0191, 0.0203, 0.0182, 0.0173, 0.0178, 0.0182, 0.0154, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:31:32,054 INFO [finetune.py:976] (2/7) Epoch 14, batch 2100, loss[loss=0.1302, simple_loss=0.205, pruned_loss=0.02766, over 4773.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2468, pruned_loss=0.05555, over 955076.68 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:32:05,802 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-04-27 05:32:06,102 INFO [finetune.py:976] (2/7) Epoch 14, batch 2150, loss[loss=0.2035, simple_loss=0.2873, pruned_loss=0.05982, over 4905.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2511, pruned_loss=0.05718, over 955466.21 frames. ], batch size: 37, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:32:14,725 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76625.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:32:27,725 INFO [optim.py:369] (2/7) 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 05:32:30,845 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:32:33,225 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1092, 2.6004, 1.0096, 1.3499, 2.0094, 1.2130, 3.3462, 1.6725], + device='cuda:2'), covar=tensor([0.0643, 0.0734, 0.0815, 0.1216, 0.0485, 0.0992, 0.0228, 0.0627], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 05:32:44,023 INFO [finetune.py:976] (2/7) Epoch 14, batch 2200, loss[loss=0.2034, simple_loss=0.2792, pruned_loss=0.06382, over 4938.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2543, pruned_loss=0.05822, over 955728.48 frames. ], batch size: 38, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:33:06,763 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76679.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 05:33:08,608 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4056, 2.7890, 2.4709, 2.7845, 2.1385, 2.5375, 2.3929, 2.1604], + device='cuda:2'), covar=tensor([0.1590, 0.1166, 0.0767, 0.0795, 0.2527, 0.0932, 0.1674, 0.1890], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0313, 0.0227, 0.0287, 0.0314, 0.0268, 0.0257, 0.0273], + device='cuda:2'), out_proj_covar=tensor([1.1851e-04, 1.2481e-04, 9.0739e-05, 1.1438e-04, 1.2814e-04, 1.0706e-04, + 1.0436e-04, 1.0903e-04], device='cuda:2') +2023-04-27 05:33:41,673 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3640, 1.5970, 1.6527, 2.1851, 2.3356, 1.8853, 1.8590, 1.6870], + device='cuda:2'), covar=tensor([0.1562, 0.1800, 0.1767, 0.1409, 0.1338, 0.2013, 0.2240, 0.2011], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0314, 0.0351, 0.0290, 0.0328, 0.0313, 0.0303, 0.0360], + device='cuda:2'), out_proj_covar=tensor([6.3612e-05, 6.6030e-05, 7.5353e-05, 5.9226e-05, 6.8420e-05, 6.6439e-05, + 6.4167e-05, 7.6829e-05], device='cuda:2') +2023-04-27 05:33:46,580 INFO [finetune.py:976] (2/7) Epoch 14, batch 2250, loss[loss=0.1898, simple_loss=0.263, pruned_loss=0.05826, over 4908.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2542, pruned_loss=0.05844, over 954034.14 frames. ], batch size: 36, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:33:56,843 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2551, 1.6734, 2.1380, 2.7062, 2.0539, 1.6386, 1.2754, 1.8723], + device='cuda:2'), covar=tensor([0.3863, 0.3581, 0.1783, 0.2386, 0.3120, 0.2961, 0.4580, 0.2422], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0248, 0.0224, 0.0317, 0.0217, 0.0230, 0.0231, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 05:34:30,285 INFO [optim.py:369] (2/7) 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 05:34:47,190 INFO [finetune.py:976] (2/7) Epoch 14, batch 2300, loss[loss=0.1792, simple_loss=0.2484, pruned_loss=0.05496, over 4862.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2544, pruned_loss=0.05777, over 955432.22 frames. ], batch size: 34, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:34:58,073 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 05:34:59,754 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.3376, 4.2111, 3.0968, 5.0063, 4.3343, 4.3229, 1.9391, 4.2678], + device='cuda:2'), covar=tensor([0.1678, 0.1048, 0.3532, 0.0949, 0.2967, 0.1448, 0.5394, 0.2169], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0212, 0.0247, 0.0301, 0.0295, 0.0245, 0.0270, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 05:35:26,138 INFO [finetune.py:976] (2/7) Epoch 14, batch 2350, loss[loss=0.137, simple_loss=0.2116, pruned_loss=0.0312, over 4926.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2526, pruned_loss=0.05741, over 956360.78 frames. ], batch size: 33, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:36:05,984 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76839.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:36:09,505 INFO [optim.py:369] (2/7) 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 05:36:20,258 INFO [finetune.py:976] (2/7) Epoch 14, batch 2400, loss[loss=0.1775, simple_loss=0.2309, pruned_loss=0.06208, over 4115.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2496, pruned_loss=0.05655, over 957451.51 frames. ], batch size: 18, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:36:20,992 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8406, 2.3100, 1.7617, 1.7191, 1.3231, 1.3594, 1.8830, 1.2170], + device='cuda:2'), covar=tensor([0.1619, 0.1466, 0.1458, 0.1805, 0.2358, 0.1977, 0.1040, 0.2067], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0213, 0.0168, 0.0204, 0.0201, 0.0183, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 05:36:36,759 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5293, 1.4705, 4.2196, 3.9182, 3.7021, 3.9961, 3.9809, 3.6894], + device='cuda:2'), covar=tensor([0.6942, 0.5840, 0.1000, 0.1885, 0.1195, 0.1831, 0.1225, 0.1566], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0303, 0.0398, 0.0400, 0.0342, 0.0399, 0.0311, 0.0360], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:36:46,484 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76900.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:36:48,935 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2283, 1.8479, 2.1869, 2.5585, 2.5198, 2.2308, 1.7959, 2.3118], + device='cuda:2'), covar=tensor([0.0715, 0.1085, 0.0582, 0.0492, 0.0542, 0.0667, 0.0697, 0.0509], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0200, 0.0180, 0.0170, 0.0176, 0.0179, 0.0152, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:36:54,081 INFO [finetune.py:976] (2/7) Epoch 14, batch 2450, loss[loss=0.1961, simple_loss=0.2517, pruned_loss=0.07029, over 4203.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2471, pruned_loss=0.05589, over 953255.36 frames. ], batch size: 18, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:37:04,096 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76925.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:37:17,025 INFO [optim.py:369] (2/7) 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 05:37:20,147 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:37:27,665 INFO [finetune.py:976] (2/7) Epoch 14, batch 2500, loss[loss=0.2544, simple_loss=0.3317, pruned_loss=0.08849, over 4818.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2494, pruned_loss=0.05678, over 954674.05 frames. ], batch size: 38, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:37:34,474 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-27 05:37:36,503 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76973.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 05:37:40,677 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76979.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:37:43,676 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 05:37:52,651 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:38:07,210 INFO [finetune.py:976] (2/7) Epoch 14, batch 2550, loss[loss=0.1511, simple_loss=0.2228, pruned_loss=0.0397, over 4764.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2518, pruned_loss=0.05702, over 955162.01 frames. ], batch size: 28, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:38:18,429 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77027.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:38:30,334 INFO [optim.py:369] (2/7) 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 05:38:37,102 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4933, 1.4980, 1.7254, 1.7498, 1.4110, 1.1997, 1.4973, 1.0662], + device='cuda:2'), covar=tensor([0.0640, 0.0553, 0.0500, 0.0667, 0.0835, 0.1088, 0.0762, 0.0741], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0075, 0.0096, 0.0076, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:38:40,001 INFO [finetune.py:976] (2/7) Epoch 14, batch 2600, loss[loss=0.1689, simple_loss=0.2439, pruned_loss=0.04697, over 4902.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2524, pruned_loss=0.05673, over 953689.60 frames. ], batch size: 36, lr: 3.55e-03, grad_scale: 64.0 +2023-04-27 05:39:20,526 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77095.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:39:33,969 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-27 05:39:41,113 INFO [finetune.py:976] (2/7) Epoch 14, batch 2650, loss[loss=0.1698, simple_loss=0.2474, pruned_loss=0.04611, over 4920.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2542, pruned_loss=0.0572, over 955543.93 frames. ], batch size: 42, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:40:19,436 INFO [optim.py:369] (2/7) 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 05:40:26,291 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77156.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 05:40:29,225 INFO [finetune.py:976] (2/7) Epoch 14, batch 2700, loss[loss=0.1847, simple_loss=0.2671, pruned_loss=0.05111, over 4904.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2525, pruned_loss=0.05613, over 956484.29 frames. ], batch size: 37, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:40:53,255 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77195.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:03,078 INFO [finetune.py:976] (2/7) Epoch 14, batch 2750, loss[loss=0.1804, simple_loss=0.2411, pruned_loss=0.0598, over 4812.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2496, pruned_loss=0.05546, over 957676.72 frames. ], batch size: 41, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:41:03,196 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77211.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:07,656 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-27 05:41:11,130 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 05:41:13,347 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9215, 1.8491, 2.0885, 2.3471, 1.7929, 1.6452, 1.9017, 1.1390], + device='cuda:2'), covar=tensor([0.0767, 0.0766, 0.0648, 0.0844, 0.0960, 0.1166, 0.0835, 0.0931], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0075, 0.0096, 0.0076, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:41:37,789 INFO [optim.py:369] (2/7) 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 05:41:42,668 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7806, 3.6441, 2.8036, 4.2800, 3.6784, 3.6852, 1.4962, 3.7231], + device='cuda:2'), covar=tensor([0.1693, 0.1190, 0.3145, 0.1580, 0.2959, 0.1832, 0.5696, 0.2246], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0212, 0.0248, 0.0302, 0.0295, 0.0244, 0.0270, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 05:41:48,207 INFO [finetune.py:976] (2/7) Epoch 14, batch 2800, loss[loss=0.1514, simple_loss=0.2232, pruned_loss=0.03982, over 4691.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2459, pruned_loss=0.05412, over 955512.56 frames. ], batch size: 23, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:41:55,144 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:55,751 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:41:58,625 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77277.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:42:00,346 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9680, 2.2790, 0.8610, 1.2221, 1.6044, 1.1985, 2.5230, 1.4565], + device='cuda:2'), covar=tensor([0.0653, 0.0612, 0.0670, 0.1261, 0.0453, 0.1009, 0.0369, 0.0707], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0067, 0.0049, 0.0047, 0.0051, 0.0052, 0.0076, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 05:42:22,270 INFO [finetune.py:976] (2/7) Epoch 14, batch 2850, loss[loss=0.168, simple_loss=0.2345, pruned_loss=0.0507, over 4894.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2454, pruned_loss=0.0541, over 954521.49 frames. ], batch size: 32, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:42:27,873 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1447, 2.4659, 0.8041, 1.4624, 1.4479, 1.8874, 1.6189, 0.8917], + device='cuda:2'), covar=tensor([0.1462, 0.1178, 0.1866, 0.1305, 0.1221, 0.0966, 0.1462, 0.1831], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0247, 0.0140, 0.0122, 0.0133, 0.0154, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:42:29,323 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-27 05:42:37,431 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77334.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:42:39,860 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77338.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:42:44,406 INFO [optim.py:369] (2/7) 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 05:42:52,722 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5584, 2.1138, 2.6197, 2.9541, 3.0832, 2.5144, 2.0210, 2.5665], + device='cuda:2'), covar=tensor([0.0788, 0.1066, 0.0541, 0.0508, 0.0478, 0.0708, 0.0736, 0.0519], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0203, 0.0183, 0.0173, 0.0179, 0.0183, 0.0155, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:42:55,657 INFO [finetune.py:976] (2/7) Epoch 14, batch 2900, loss[loss=0.2508, simple_loss=0.3176, pruned_loss=0.09197, over 4802.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2486, pruned_loss=0.05531, over 954305.25 frames. ], batch size: 41, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:42:58,591 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 05:43:17,709 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9237, 1.5117, 2.0282, 2.3177, 1.9657, 1.8882, 2.0129, 1.9400], + device='cuda:2'), covar=tensor([0.5109, 0.7051, 0.7380, 0.6635, 0.6683, 0.8668, 0.8790, 0.8214], + device='cuda:2'), in_proj_covar=tensor([0.0413, 0.0406, 0.0494, 0.0511, 0.0442, 0.0464, 0.0471, 0.0473], + device='cuda:2'), out_proj_covar=tensor([9.9888e-05, 1.0061e-04, 1.1117e-04, 1.2130e-04, 1.0661e-04, 1.1167e-04, + 1.1233e-04, 1.1273e-04], device='cuda:2') +2023-04-27 05:43:29,374 INFO [finetune.py:976] (2/7) Epoch 14, batch 2950, loss[loss=0.2042, simple_loss=0.2665, pruned_loss=0.07092, over 4827.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.252, pruned_loss=0.05658, over 955128.16 frames. ], batch size: 30, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:43:30,685 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1808, 1.6098, 1.5427, 1.9828, 1.8226, 2.1481, 1.5525, 4.3974], + device='cuda:2'), covar=tensor([0.0606, 0.0834, 0.0813, 0.1209, 0.0649, 0.0554, 0.0767, 0.0125], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 05:43:50,926 INFO [optim.py:369] (2/7) 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 05:43:56,100 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77451.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 05:44:03,067 INFO [finetune.py:976] (2/7) Epoch 14, batch 3000, loss[loss=0.1533, simple_loss=0.2402, pruned_loss=0.03321, over 4791.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2542, pruned_loss=0.0572, over 955287.23 frames. ], batch size: 29, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:44:03,067 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 05:44:12,343 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6857, 1.8166, 1.7372, 1.4179, 1.9283, 1.5172, 2.3311, 1.5414], + device='cuda:2'), covar=tensor([0.3835, 0.1801, 0.5545, 0.2901, 0.1501, 0.2626, 0.1501, 0.5080], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0347, 0.0429, 0.0356, 0.0383, 0.0385, 0.0373, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:44:19,534 INFO [finetune.py:1010] (2/7) Epoch 14, validation: loss=0.1527, simple_loss=0.224, pruned_loss=0.04073, over 2265189.00 frames. +2023-04-27 05:44:19,534 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6355MB +2023-04-27 05:45:02,948 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:45:24,240 INFO [finetune.py:976] (2/7) Epoch 14, batch 3050, loss[loss=0.1465, simple_loss=0.2111, pruned_loss=0.04098, over 4303.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2541, pruned_loss=0.05702, over 954701.92 frames. ], batch size: 18, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:45:45,781 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:45:51,964 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-04-27 05:45:57,727 INFO [finetune.py:976] (2/7) Epoch 14, batch 3100, loss[loss=0.1688, simple_loss=0.2324, pruned_loss=0.05258, over 4924.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2518, pruned_loss=0.05617, over 953619.55 frames. ], batch size: 33, lr: 3.54e-03, grad_scale: 64.0 +2023-04-27 05:46:02,402 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77567.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:46:10,050 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6028, 1.5732, 1.8688, 1.8721, 1.4789, 1.3328, 1.6087, 0.9857], + device='cuda:2'), covar=tensor([0.0622, 0.0683, 0.0484, 0.0827, 0.0807, 0.1215, 0.0697, 0.0820], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0075, 0.0097, 0.0076, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:46:19,155 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0858, 2.4417, 1.0193, 1.3914, 1.8978, 1.1480, 3.2163, 1.7470], + device='cuda:2'), covar=tensor([0.0628, 0.0604, 0.0785, 0.1221, 0.0506, 0.1025, 0.0259, 0.0602], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 05:46:32,672 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6274, 1.4539, 1.7141, 1.9551, 2.1227, 1.6242, 1.2502, 1.7745], + device='cuda:2'), covar=tensor([0.0903, 0.1376, 0.0809, 0.0620, 0.0599, 0.0847, 0.0834, 0.0639], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0204, 0.0183, 0.0173, 0.0180, 0.0184, 0.0155, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:46:36,287 INFO [finetune.py:976] (2/7) Epoch 14, batch 3150, loss[loss=0.1784, simple_loss=0.2484, pruned_loss=0.05419, over 4915.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.249, pruned_loss=0.05529, over 951974.93 frames. ], batch size: 43, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:46:44,734 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:47:05,389 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77629.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:47:07,855 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77633.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:47:20,988 INFO [optim.py:369] (2/7) 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 05:47:41,185 INFO [finetune.py:976] (2/7) Epoch 14, batch 3200, loss[loss=0.1748, simple_loss=0.2266, pruned_loss=0.06157, over 4823.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2454, pruned_loss=0.05441, over 952833.53 frames. ], batch size: 30, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:48:05,089 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:48:05,098 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:48:48,228 INFO [finetune.py:976] (2/7) Epoch 14, batch 3250, loss[loss=0.2174, simple_loss=0.2748, pruned_loss=0.07998, over 4699.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.246, pruned_loss=0.05519, over 950844.29 frames. ], batch size: 23, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:49:28,768 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:49:33,680 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8900, 1.5924, 1.8982, 2.2675, 2.3258, 1.9043, 1.5901, 1.9705], + device='cuda:2'), covar=tensor([0.0955, 0.1235, 0.0740, 0.0589, 0.0646, 0.0904, 0.0773, 0.0612], + device='cuda:2'), in_proj_covar=tensor([0.0190, 0.0201, 0.0181, 0.0172, 0.0177, 0.0181, 0.0153, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:49:36,704 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77751.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:49:42,688 INFO [finetune.py:976] (2/7) Epoch 14, batch 3300, loss[loss=0.2316, simple_loss=0.3002, pruned_loss=0.08148, over 4896.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2508, pruned_loss=0.05691, over 952274.40 frames. ], batch size: 43, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:50:08,403 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77799.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:50:15,852 INFO [finetune.py:976] (2/7) Epoch 14, batch 3350, loss[loss=0.1629, simple_loss=0.2415, pruned_loss=0.04217, over 4747.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2527, pruned_loss=0.05714, over 950877.11 frames. ], batch size: 27, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:50:24,477 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 05:50:37,583 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0955, 2.5271, 1.0420, 1.4097, 2.0736, 1.2063, 3.6099, 1.7097], + device='cuda:2'), covar=tensor([0.0688, 0.0663, 0.0842, 0.1297, 0.0505, 0.1056, 0.0273, 0.0691], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 05:50:39,907 INFO [optim.py:369] (2/7) 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 05:50:49,111 INFO [finetune.py:976] (2/7) Epoch 14, batch 3400, loss[loss=0.1989, simple_loss=0.2597, pruned_loss=0.069, over 4922.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2543, pruned_loss=0.05787, over 949454.16 frames. ], batch size: 29, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:50:52,866 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77867.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:51:22,439 INFO [finetune.py:976] (2/7) Epoch 14, batch 3450, loss[loss=0.1739, simple_loss=0.2316, pruned_loss=0.05807, over 4895.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2545, pruned_loss=0.0579, over 950018.55 frames. ], batch size: 32, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:51:24,869 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77915.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:51:26,761 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2905, 4.4963, 1.0266, 2.2980, 2.4914, 2.7398, 2.4887, 0.9907], + device='cuda:2'), covar=tensor([0.1248, 0.0908, 0.2174, 0.1259, 0.1014, 0.1290, 0.1464, 0.2151], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0246, 0.0139, 0.0121, 0.0132, 0.0154, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:51:33,854 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:51:45,497 INFO [optim.py:369] (2/7) 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 05:51:54,679 INFO [finetune.py:976] (2/7) Epoch 14, batch 3500, loss[loss=0.2607, simple_loss=0.318, pruned_loss=0.1017, over 4815.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2517, pruned_loss=0.05691, over 951479.62 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:52:01,423 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:52:07,432 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:52:20,490 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2681, 3.1669, 0.8546, 1.5975, 1.5715, 2.2065, 1.7420, 0.9125], + device='cuda:2'), covar=tensor([0.1934, 0.1328, 0.2455, 0.1836, 0.1450, 0.1410, 0.1780, 0.2256], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0245, 0.0138, 0.0121, 0.0132, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 05:52:28,917 INFO [finetune.py:976] (2/7) Epoch 14, batch 3550, loss[loss=0.1591, simple_loss=0.2368, pruned_loss=0.04073, over 4827.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2495, pruned_loss=0.05648, over 952405.42 frames. ], batch size: 39, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:52:36,484 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-27 05:52:42,983 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:53:17,515 INFO [finetune.py:976] (2/7) Epoch 14, batch 3600, loss[loss=0.1593, simple_loss=0.2307, pruned_loss=0.04394, over 4801.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2474, pruned_loss=0.05604, over 952317.73 frames. ], batch size: 29, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:54:11,989 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9867, 1.6273, 1.5663, 1.7611, 2.2445, 1.8463, 1.5396, 1.3504], + device='cuda:2'), covar=tensor([0.1861, 0.1366, 0.1948, 0.1201, 0.0929, 0.1502, 0.2574, 0.2357], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0318, 0.0354, 0.0292, 0.0330, 0.0316, 0.0305, 0.0362], + device='cuda:2'), out_proj_covar=tensor([6.3930e-05, 6.6722e-05, 7.5828e-05, 5.9782e-05, 6.8689e-05, 6.7006e-05, + 6.4511e-05, 7.7329e-05], device='cuda:2') +2023-04-27 05:54:18,583 INFO [finetune.py:976] (2/7) Epoch 14, batch 3650, loss[loss=0.199, simple_loss=0.2687, pruned_loss=0.06468, over 4863.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2485, pruned_loss=0.05664, over 951690.22 frames. ], batch size: 31, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:54:34,241 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0668, 2.1355, 2.2514, 2.9346, 2.9470, 2.6737, 2.5246, 2.1675], + device='cuda:2'), covar=tensor([0.1199, 0.1548, 0.1536, 0.1377, 0.1007, 0.1345, 0.1919, 0.1730], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0317, 0.0353, 0.0292, 0.0329, 0.0316, 0.0304, 0.0361], + device='cuda:2'), out_proj_covar=tensor([6.3776e-05, 6.6539e-05, 7.5622e-05, 5.9634e-05, 6.8545e-05, 6.6921e-05, + 6.4386e-05, 7.7167e-05], device='cuda:2') +2023-04-27 05:54:48,998 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1091, 2.2264, 1.8876, 1.8318, 2.3742, 1.8980, 2.8428, 1.5881], + device='cuda:2'), covar=tensor([0.3919, 0.1984, 0.4713, 0.3381, 0.1737, 0.2565, 0.1579, 0.4724], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0351, 0.0431, 0.0360, 0.0386, 0.0387, 0.0378, 0.0425], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:54:56,789 INFO [optim.py:369] (2/7) 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 05:55:18,205 INFO [finetune.py:976] (2/7) Epoch 14, batch 3700, loss[loss=0.1665, simple_loss=0.23, pruned_loss=0.05145, over 4154.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2523, pruned_loss=0.05758, over 953032.94 frames. ], batch size: 18, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:55:22,094 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 05:55:56,843 INFO [finetune.py:976] (2/7) Epoch 14, batch 3750, loss[loss=0.1792, simple_loss=0.2497, pruned_loss=0.05436, over 4760.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2536, pruned_loss=0.05731, over 954130.71 frames. ], batch size: 26, lr: 3.54e-03, grad_scale: 32.0 +2023-04-27 05:55:57,604 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3796, 1.2955, 1.3262, 0.9127, 1.2933, 1.1526, 1.5173, 1.3072], + device='cuda:2'), covar=tensor([0.3350, 0.1601, 0.3897, 0.2266, 0.1394, 0.2016, 0.1642, 0.3836], + device='cuda:2'), in_proj_covar=tensor([0.0348, 0.0351, 0.0431, 0.0360, 0.0386, 0.0387, 0.0378, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:56:06,168 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9050, 2.3143, 1.9936, 2.2597, 1.6374, 1.9985, 1.9505, 1.5891], + device='cuda:2'), covar=tensor([0.1705, 0.1018, 0.0743, 0.0994, 0.2961, 0.0952, 0.1655, 0.2066], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0308, 0.0223, 0.0280, 0.0310, 0.0263, 0.0252, 0.0268], + device='cuda:2'), out_proj_covar=tensor([1.1623e-04, 1.2257e-04, 8.8800e-05, 1.1191e-04, 1.2644e-04, 1.0531e-04, + 1.0214e-04, 1.0716e-04], device='cuda:2') +2023-04-27 05:56:09,775 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:56:18,615 INFO [optim.py:369] (2/7) 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 05:56:30,164 INFO [finetune.py:976] (2/7) Epoch 14, batch 3800, loss[loss=0.201, simple_loss=0.2731, pruned_loss=0.0644, over 4818.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2553, pruned_loss=0.0574, over 955202.41 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:56:37,445 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:56:47,234 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4792, 0.9459, 0.4928, 1.1720, 1.1069, 1.3411, 1.2367, 1.2576], + device='cuda:2'), covar=tensor([0.0536, 0.0430, 0.0404, 0.0594, 0.0308, 0.0564, 0.0524, 0.0613], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 05:56:50,269 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78293.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 05:57:03,399 INFO [finetune.py:976] (2/7) Epoch 14, batch 3850, loss[loss=0.1771, simple_loss=0.2429, pruned_loss=0.05561, over 4922.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2539, pruned_loss=0.05714, over 955361.60 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:57:09,825 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:57:10,575 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-04-27 05:57:17,862 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78333.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:57:22,168 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9743, 1.6954, 1.9581, 2.2981, 2.3731, 1.8659, 1.4561, 2.0337], + device='cuda:2'), covar=tensor([0.0809, 0.1171, 0.0648, 0.0592, 0.0594, 0.0774, 0.0790, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0191, 0.0202, 0.0182, 0.0172, 0.0177, 0.0182, 0.0153, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 05:57:25,784 INFO [optim.py:369] (2/7) 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 05:57:30,745 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 05:57:36,828 INFO [finetune.py:976] (2/7) Epoch 14, batch 3900, loss[loss=0.1439, simple_loss=0.2219, pruned_loss=0.03296, over 4835.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2513, pruned_loss=0.05665, over 952353.98 frames. ], batch size: 47, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:57:50,413 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78381.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:58:09,805 INFO [finetune.py:976] (2/7) Epoch 14, batch 3950, loss[loss=0.1603, simple_loss=0.2269, pruned_loss=0.04686, over 4818.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2497, pruned_loss=0.05671, over 952901.77 frames. ], batch size: 41, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 05:58:25,588 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7838, 1.9053, 0.8305, 1.4339, 1.8946, 1.6167, 1.5171, 1.5891], + device='cuda:2'), covar=tensor([0.0506, 0.0375, 0.0351, 0.0549, 0.0268, 0.0502, 0.0504, 0.0593], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0025, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 05:58:50,265 INFO [optim.py:369] (2/7) 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 05:59:04,624 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78460.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 05:59:05,120 INFO [finetune.py:976] (2/7) Epoch 14, batch 4000, loss[loss=0.1676, simple_loss=0.2437, pruned_loss=0.04574, over 4815.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2469, pruned_loss=0.05527, over 953142.66 frames. ], batch size: 40, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:00:00,331 INFO [finetune.py:976] (2/7) Epoch 14, batch 4050, loss[loss=0.166, simple_loss=0.243, pruned_loss=0.04456, over 4811.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2514, pruned_loss=0.05711, over 955469.37 frames. ], batch size: 40, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:00:19,666 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78521.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:00:29,329 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6790, 1.9060, 0.7384, 1.0186, 1.3636, 1.0176, 2.1253, 1.1465], + device='cuda:2'), covar=tensor([0.0568, 0.0563, 0.0589, 0.1093, 0.0398, 0.0877, 0.0318, 0.0699], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 06:00:51,058 INFO [optim.py:369] (2/7) 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 06:01:03,058 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78556.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:01:06,003 INFO [finetune.py:976] (2/7) Epoch 14, batch 4100, loss[loss=0.1786, simple_loss=0.258, pruned_loss=0.04962, over 4848.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2551, pruned_loss=0.05829, over 954895.86 frames. ], batch size: 49, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:01:12,337 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7835, 2.3945, 0.8299, 1.0294, 1.6522, 1.0804, 3.0855, 1.2670], + device='cuda:2'), covar=tensor([0.0898, 0.0900, 0.1049, 0.1944, 0.0729, 0.1474, 0.0469, 0.1129], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 06:01:17,286 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 06:01:45,289 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78588.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 06:02:05,268 INFO [finetune.py:976] (2/7) Epoch 14, batch 4150, loss[loss=0.2041, simple_loss=0.2717, pruned_loss=0.06824, over 4737.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2569, pruned_loss=0.05903, over 957030.85 frames. ], batch size: 59, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:02:09,540 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:02:17,808 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-27 06:02:29,493 INFO [optim.py:369] (2/7) 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 06:02:38,717 INFO [finetune.py:976] (2/7) Epoch 14, batch 4200, loss[loss=0.1345, simple_loss=0.2023, pruned_loss=0.03336, over 4812.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2573, pruned_loss=0.05885, over 957466.71 frames. ], batch size: 25, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:03:03,628 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3158, 1.9925, 2.3686, 2.6079, 2.7412, 2.1991, 1.9058, 2.2965], + device='cuda:2'), covar=tensor([0.0746, 0.0984, 0.0523, 0.0492, 0.0515, 0.0732, 0.0727, 0.0549], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0203, 0.0183, 0.0173, 0.0178, 0.0184, 0.0154, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:03:12,003 INFO [finetune.py:976] (2/7) Epoch 14, batch 4250, loss[loss=0.1594, simple_loss=0.2341, pruned_loss=0.0424, over 4930.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2538, pruned_loss=0.05768, over 956077.49 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:03:36,184 INFO [optim.py:369] (2/7) 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 06:03:45,440 INFO [finetune.py:976] (2/7) Epoch 14, batch 4300, loss[loss=0.1441, simple_loss=0.2195, pruned_loss=0.03432, over 4817.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2509, pruned_loss=0.0562, over 956569.58 frames. ], batch size: 25, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:04:19,292 INFO [finetune.py:976] (2/7) Epoch 14, batch 4350, loss[loss=0.1281, simple_loss=0.1865, pruned_loss=0.03484, over 4455.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2472, pruned_loss=0.0546, over 957714.04 frames. ], batch size: 20, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:04:22,429 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78816.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:04:23,660 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8652, 4.1260, 0.7831, 2.2340, 2.3223, 2.8215, 2.5130, 0.9680], + device='cuda:2'), covar=tensor([0.1386, 0.1053, 0.2266, 0.1333, 0.1039, 0.1099, 0.1409, 0.2086], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0139, 0.0121, 0.0132, 0.0153, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:04:43,555 INFO [optim.py:369] (2/7) 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 06:05:04,244 INFO [finetune.py:976] (2/7) Epoch 14, batch 4400, loss[loss=0.1744, simple_loss=0.2517, pruned_loss=0.04861, over 4709.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2488, pruned_loss=0.05565, over 956656.56 frames. ], batch size: 59, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:05:13,200 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2421, 1.6082, 1.5489, 2.1218, 2.2776, 1.8797, 1.8148, 1.6186], + device='cuda:2'), covar=tensor([0.1748, 0.1717, 0.1926, 0.1802, 0.1256, 0.1870, 0.2084, 0.2064], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0318, 0.0354, 0.0293, 0.0329, 0.0315, 0.0304, 0.0362], + device='cuda:2'), out_proj_covar=tensor([6.4109e-05, 6.6650e-05, 7.6047e-05, 5.9842e-05, 6.8319e-05, 6.6607e-05, + 6.4424e-05, 7.7371e-05], device='cuda:2') +2023-04-27 06:05:38,858 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:06:11,504 INFO [finetune.py:976] (2/7) Epoch 14, batch 4450, loss[loss=0.2498, simple_loss=0.3077, pruned_loss=0.09601, over 4754.00 frames. ], tot_loss[loss=0.182, simple_loss=0.251, pruned_loss=0.05648, over 953306.01 frames. ], batch size: 59, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:06:12,193 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78912.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:06:44,171 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:07:03,712 INFO [optim.py:369] (2/7) 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 06:07:18,900 INFO [finetune.py:976] (2/7) Epoch 14, batch 4500, loss[loss=0.1765, simple_loss=0.2542, pruned_loss=0.04945, over 4762.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2521, pruned_loss=0.05636, over 952697.82 frames. ], batch size: 54, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:07:20,862 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7566, 1.3097, 1.8780, 2.1851, 1.8373, 1.7142, 1.8080, 1.7912], + device='cuda:2'), covar=tensor([0.4749, 0.6614, 0.6358, 0.6261, 0.6087, 0.8031, 0.8234, 0.7655], + device='cuda:2'), in_proj_covar=tensor([0.0414, 0.0407, 0.0495, 0.0509, 0.0444, 0.0465, 0.0472, 0.0475], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:07:21,398 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78965.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:07:21,976 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1093, 1.3833, 1.3440, 1.6305, 1.4396, 1.6572, 1.2743, 2.9715], + device='cuda:2'), covar=tensor([0.0652, 0.0837, 0.0800, 0.1200, 0.0677, 0.0546, 0.0770, 0.0195], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0039, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 06:07:27,376 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7736, 1.4416, 1.3579, 1.6097, 2.0331, 1.6306, 1.4100, 1.2857], + device='cuda:2'), covar=tensor([0.1418, 0.1511, 0.1865, 0.1359, 0.0786, 0.1432, 0.2125, 0.2013], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0318, 0.0356, 0.0293, 0.0329, 0.0316, 0.0305, 0.0363], + device='cuda:2'), out_proj_covar=tensor([6.4234e-05, 6.6650e-05, 7.6430e-05, 5.9846e-05, 6.8493e-05, 6.6900e-05, + 6.4530e-05, 7.7518e-05], device='cuda:2') +2023-04-27 06:07:35,316 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 06:07:52,194 INFO [finetune.py:976] (2/7) Epoch 14, batch 4550, loss[loss=0.205, simple_loss=0.2637, pruned_loss=0.07319, over 4180.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2543, pruned_loss=0.05712, over 954208.63 frames. ], batch size: 65, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:07:54,207 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.94 vs. limit=5.0 +2023-04-27 06:07:58,393 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1108, 2.4927, 0.8469, 1.4375, 1.4320, 1.8746, 1.6025, 0.7927], + device='cuda:2'), covar=tensor([0.1492, 0.1126, 0.1758, 0.1351, 0.1180, 0.0897, 0.1455, 0.1806], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0246, 0.0139, 0.0122, 0.0133, 0.0153, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:08:01,460 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79026.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:08:14,992 INFO [optim.py:369] (2/7) 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 06:08:21,034 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 06:08:26,108 INFO [finetune.py:976] (2/7) Epoch 14, batch 4600, loss[loss=0.172, simple_loss=0.2436, pruned_loss=0.05024, over 4822.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2534, pruned_loss=0.05638, over 954420.15 frames. ], batch size: 47, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:08:44,761 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1095, 1.6043, 4.4326, 4.1657, 3.9406, 4.0140, 3.8671, 3.9325], + device='cuda:2'), covar=tensor([0.5944, 0.5054, 0.0966, 0.1514, 0.0904, 0.1725, 0.2862, 0.1357], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0305, 0.0401, 0.0406, 0.0347, 0.0406, 0.0312, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:08:59,224 INFO [finetune.py:976] (2/7) Epoch 14, batch 4650, loss[loss=0.1606, simple_loss=0.2281, pruned_loss=0.04656, over 4730.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2515, pruned_loss=0.05636, over 953079.37 frames. ], batch size: 23, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:09:02,362 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:09:21,599 INFO [optim.py:369] (2/7) 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 06:09:32,660 INFO [finetune.py:976] (2/7) Epoch 14, batch 4700, loss[loss=0.1484, simple_loss=0.2243, pruned_loss=0.03628, over 4831.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2488, pruned_loss=0.05538, over 953284.68 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:09:34,546 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 4750, loss[loss=0.2238, simple_loss=0.2675, pruned_loss=0.09005, over 4476.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2472, pruned_loss=0.05518, over 954193.04 frames. ], batch size: 19, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:10:06,534 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79212.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:10:18,589 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6508, 1.8303, 1.8695, 2.4629, 2.7039, 2.1682, 2.1192, 1.9029], + device='cuda:2'), covar=tensor([0.1780, 0.1893, 0.1842, 0.1776, 0.1164, 0.1902, 0.2615, 0.2146], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0319, 0.0356, 0.0292, 0.0330, 0.0317, 0.0305, 0.0363], + device='cuda:2'), out_proj_covar=tensor([6.4393e-05, 6.6865e-05, 7.6476e-05, 5.9717e-05, 6.8763e-05, 6.7096e-05, + 6.4487e-05, 7.7528e-05], device='cuda:2') +2023-04-27 06:10:37,093 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 06:10:37,980 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:10:59,215 INFO [finetune.py:976] (2/7) Epoch 14, batch 4800, loss[loss=0.2074, simple_loss=0.2768, pruned_loss=0.06901, over 4891.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2493, pruned_loss=0.05606, over 955871.08 frames. ], batch size: 43, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:11:21,233 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-27 06:11:25,261 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79285.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:11:32,017 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2800, 1.5427, 1.5421, 2.1838, 2.3949, 1.9400, 1.8634, 1.6268], + device='cuda:2'), covar=tensor([0.1943, 0.2178, 0.2231, 0.1967, 0.1194, 0.2196, 0.2373, 0.2485], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0320, 0.0357, 0.0293, 0.0331, 0.0318, 0.0306, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.4439e-05, 6.7142e-05, 7.6772e-05, 5.9827e-05, 6.8943e-05, 6.7405e-05, + 6.4650e-05, 7.7742e-05], device='cuda:2') +2023-04-27 06:11:58,653 INFO [finetune.py:976] (2/7) Epoch 14, batch 4850, loss[loss=0.1755, simple_loss=0.246, pruned_loss=0.05252, over 4753.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2529, pruned_loss=0.0574, over 956182.83 frames. ], batch size: 27, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:12:10,417 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:12:23,704 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79333.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:12:36,862 INFO [optim.py:369] (2/7) 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 06:12:36,981 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79346.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:12:56,921 INFO [finetune.py:976] (2/7) Epoch 14, batch 4900, loss[loss=0.1722, simple_loss=0.242, pruned_loss=0.05127, over 4267.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2545, pruned_loss=0.05824, over 956129.65 frames. ], batch size: 65, lr: 3.53e-03, grad_scale: 32.0 +2023-04-27 06:13:00,246 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79364.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:13:00,395 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 06:13:07,473 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79375.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:13:19,222 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 4950, loss[loss=0.1403, simple_loss=0.2183, pruned_loss=0.03111, over 4913.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2554, pruned_loss=0.05819, over 956001.24 frames. ], batch size: 38, lr: 3.52e-03, grad_scale: 32.0 +2023-04-27 06:13:40,500 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:13:47,620 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79436.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:13:53,549 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 14, batch 5000, loss[loss=0.1847, simple_loss=0.2587, pruned_loss=0.0554, over 4913.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2542, pruned_loss=0.05762, over 956696.90 frames. ], batch size: 36, lr: 3.52e-03, grad_scale: 32.0 +2023-04-27 06:14:35,550 INFO [finetune.py:976] (2/7) Epoch 14, batch 5050, loss[loss=0.1721, simple_loss=0.2342, pruned_loss=0.05505, over 4799.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2503, pruned_loss=0.05615, over 955767.20 frames. ], batch size: 51, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:14:49,321 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5493, 2.0899, 1.7448, 1.8827, 1.6057, 1.6179, 1.6468, 1.3505], + device='cuda:2'), covar=tensor([0.2067, 0.1293, 0.0893, 0.1251, 0.3319, 0.1316, 0.1876, 0.2431], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0309, 0.0223, 0.0280, 0.0310, 0.0262, 0.0252, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1615e-04, 1.2287e-04, 8.8631e-05, 1.1147e-04, 1.2619e-04, 1.0488e-04, + 1.0209e-04, 1.0661e-04], device='cuda:2') +2023-04-27 06:14:52,903 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8170, 1.7980, 0.8773, 1.4718, 1.8406, 1.6634, 1.5641, 1.5817], + device='cuda:2'), covar=tensor([0.0495, 0.0375, 0.0347, 0.0539, 0.0262, 0.0500, 0.0497, 0.0581], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 06:14:59,515 INFO [optim.py:369] (2/7) 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 06:14:59,686 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7567, 1.3060, 1.8610, 2.2567, 1.8136, 1.6885, 1.7780, 1.8038], + device='cuda:2'), covar=tensor([0.5479, 0.7361, 0.7039, 0.6226, 0.6658, 0.9053, 0.9023, 0.8894], + device='cuda:2'), in_proj_covar=tensor([0.0417, 0.0408, 0.0497, 0.0512, 0.0446, 0.0467, 0.0474, 0.0477], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:15:01,455 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4167, 3.3301, 2.5030, 3.8719, 3.3431, 3.3586, 1.3235, 3.2786], + device='cuda:2'), covar=tensor([0.2052, 0.1433, 0.3201, 0.2194, 0.3292, 0.2100, 0.6268, 0.2671], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0216, 0.0250, 0.0303, 0.0300, 0.0249, 0.0274, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 06:15:05,895 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.17 vs. limit=5.0 +2023-04-27 06:15:08,090 INFO [finetune.py:976] (2/7) Epoch 14, batch 5100, loss[loss=0.1554, simple_loss=0.2199, pruned_loss=0.04547, over 4794.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2484, pruned_loss=0.05585, over 955300.85 frames. ], batch size: 26, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:15:25,158 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 5150, loss[loss=0.1334, simple_loss=0.2083, pruned_loss=0.02926, over 4865.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2479, pruned_loss=0.05577, over 954412.15 frames. ], batch size: 31, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:15:48,490 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79621.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:02,431 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79641.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:16:06,015 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 06:16:12,629 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 5200, loss[loss=0.2547, simple_loss=0.3194, pruned_loss=0.09502, over 4860.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2494, pruned_loss=0.05557, over 956104.01 frames. ], batch size: 44, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:16:19,853 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79669.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:20,546 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7043, 1.3164, 1.5068, 1.7292, 1.5296, 1.2369, 0.7814, 1.3668], + device='cuda:2'), covar=tensor([0.2748, 0.3180, 0.1595, 0.1927, 0.2305, 0.2384, 0.4606, 0.1947], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0247, 0.0222, 0.0315, 0.0214, 0.0229, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 06:16:28,683 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2471, 1.4367, 1.3754, 1.6792, 1.5656, 1.9229, 1.3366, 3.6466], + device='cuda:2'), covar=tensor([0.0661, 0.0810, 0.0864, 0.1292, 0.0709, 0.0555, 0.0772, 0.0124], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 06:16:34,512 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79689.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:16:54,348 INFO [finetune.py:976] (2/7) Epoch 14, batch 5250, loss[loss=0.1593, simple_loss=0.235, pruned_loss=0.04179, over 4693.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2511, pruned_loss=0.05556, over 956573.69 frames. ], batch size: 23, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:17:03,906 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:17:05,052 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:17:19,828 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79731.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:17:36,846 INFO [optim.py:369] (2/7) 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 06:17:42,986 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0392, 2.0613, 1.6351, 1.6074, 1.9948, 1.5515, 2.5622, 1.2776], + device='cuda:2'), covar=tensor([0.3531, 0.1721, 0.4872, 0.3083, 0.1742, 0.2708, 0.1396, 0.5055], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0348, 0.0430, 0.0356, 0.0384, 0.0381, 0.0374, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:17:45,876 INFO [finetune.py:976] (2/7) Epoch 14, batch 5300, loss[loss=0.1615, simple_loss=0.2319, pruned_loss=0.04559, over 4744.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2534, pruned_loss=0.05684, over 956894.95 frames. ], batch size: 59, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:17:46,584 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79762.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:17:55,078 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.5567, 1.4002, 1.3584, 0.9606, 1.3083, 1.1852, 1.6345, 1.3331], + device='cuda:2'), covar=tensor([0.2962, 0.1457, 0.4326, 0.2299, 0.1371, 0.2020, 0.1608, 0.4019], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0349, 0.0432, 0.0357, 0.0385, 0.0382, 0.0375, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:18:18,183 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3188, 3.0227, 2.4277, 2.7312, 2.2113, 2.5549, 2.6180, 2.0011], + device='cuda:2'), covar=tensor([0.2382, 0.1433, 0.0942, 0.1485, 0.3053, 0.1272, 0.2370, 0.3011], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0313, 0.0225, 0.0284, 0.0313, 0.0266, 0.0255, 0.0271], + device='cuda:2'), out_proj_covar=tensor([1.1752e-04, 1.2461e-04, 8.9785e-05, 1.1309e-04, 1.2767e-04, 1.0616e-04, + 1.0329e-04, 1.0802e-04], device='cuda:2') +2023-04-27 06:18:38,543 INFO [finetune.py:976] (2/7) Epoch 14, batch 5350, loss[loss=0.1799, simple_loss=0.2504, pruned_loss=0.05467, over 4773.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.254, pruned_loss=0.05676, over 955830.76 frames. ], batch size: 28, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:18:43,682 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-04-27 06:18:45,990 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79823.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:18:57,775 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0009, 1.3982, 5.2106, 4.8838, 4.5089, 4.9176, 4.5797, 4.6492], + device='cuda:2'), covar=tensor([0.6380, 0.5739, 0.0930, 0.1565, 0.0964, 0.1241, 0.1054, 0.1344], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0303, 0.0400, 0.0401, 0.0346, 0.0401, 0.0309, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:19:02,342 INFO [optim.py:369] (2/7) 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 06:19:04,322 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4792, 2.9203, 1.0108, 1.5940, 2.2225, 1.4670, 4.2134, 2.0832], + device='cuda:2'), covar=tensor([0.0630, 0.0731, 0.0840, 0.1272, 0.0514, 0.1050, 0.0261, 0.0642], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0051, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 06:19:11,410 INFO [finetune.py:976] (2/7) Epoch 14, batch 5400, loss[loss=0.189, simple_loss=0.249, pruned_loss=0.06451, over 4816.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2511, pruned_loss=0.05561, over 955762.99 frames. ], batch size: 40, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:19:45,249 INFO [finetune.py:976] (2/7) Epoch 14, batch 5450, loss[loss=0.1395, simple_loss=0.2157, pruned_loss=0.03164, over 4868.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2479, pruned_loss=0.05512, over 954477.00 frames. ], batch size: 34, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:20:04,589 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:20:04,608 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79941.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:20:09,136 INFO [optim.py:369] (2/7) 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 06:20:18,679 INFO [finetune.py:976] (2/7) Epoch 14, batch 5500, loss[loss=0.1604, simple_loss=0.2398, pruned_loss=0.04052, over 4809.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2452, pruned_loss=0.05412, over 954718.13 frames. ], batch size: 45, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:20:19,441 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6530, 1.7848, 1.4731, 1.1357, 1.2422, 1.2058, 1.4250, 1.1196], + device='cuda:2'), covar=tensor([0.1782, 0.1430, 0.1590, 0.1860, 0.2501, 0.2091, 0.1233, 0.2227], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0213, 0.0169, 0.0204, 0.0200, 0.0184, 0.0156, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 06:20:26,747 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.04 vs. limit=2.0 +2023-04-27 06:20:34,559 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6717, 2.8131, 2.2963, 2.3511, 2.8485, 2.4915, 3.6466, 2.1174], + device='cuda:2'), covar=tensor([0.3955, 0.2166, 0.4312, 0.3646, 0.1731, 0.2718, 0.2039, 0.4230], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0346, 0.0426, 0.0355, 0.0382, 0.0380, 0.0371, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:20:36,320 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79989.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:20:36,349 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 14, batch 5550, loss[loss=0.1994, simple_loss=0.2776, pruned_loss=0.06058, over 4861.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2477, pruned_loss=0.05545, over 954643.05 frames. ], batch size: 31, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:20:53,952 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0612, 0.6533, 0.9165, 0.7489, 1.2086, 0.9497, 0.8412, 0.9475], + device='cuda:2'), covar=tensor([0.1504, 0.1476, 0.1922, 0.1483, 0.0978, 0.1309, 0.1571, 0.2146], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0316, 0.0352, 0.0290, 0.0329, 0.0314, 0.0304, 0.0362], + device='cuda:2'), out_proj_covar=tensor([6.4131e-05, 6.6193e-05, 7.5570e-05, 5.9221e-05, 6.8572e-05, 6.6464e-05, + 6.4245e-05, 7.7344e-05], device='cuda:2') +2023-04-27 06:20:54,499 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80013.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:20:59,249 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80020.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:21:06,028 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80031.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 06:21:09,684 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80037.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:21:16,570 INFO [optim.py:369] (2/7) 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 06:21:24,831 INFO [finetune.py:976] (2/7) Epoch 14, batch 5600, loss[loss=0.1795, simple_loss=0.2577, pruned_loss=0.05063, over 4816.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2514, pruned_loss=0.05654, over 953323.61 frames. ], batch size: 38, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:21:28,971 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80079.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:22:01,084 INFO [finetune.py:976] (2/7) Epoch 14, batch 5650, loss[loss=0.2354, simple_loss=0.2967, pruned_loss=0.08705, over 4728.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2533, pruned_loss=0.05672, over 952450.18 frames. ], batch size: 59, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:22:11,109 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80118.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:22:28,363 INFO [optim.py:369] (2/7) 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 06:22:47,952 INFO [finetune.py:976] (2/7) Epoch 14, batch 5700, loss[loss=0.16, simple_loss=0.2198, pruned_loss=0.05013, over 4460.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2494, pruned_loss=0.05602, over 936418.07 frames. ], batch size: 19, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:23:36,604 INFO [finetune.py:976] (2/7) Epoch 15, batch 0, loss[loss=0.1652, simple_loss=0.2411, pruned_loss=0.04462, over 4824.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2411, pruned_loss=0.04462, over 4824.00 frames. ], batch size: 30, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:23:36,604 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 06:23:38,895 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7185, 1.8794, 1.8531, 1.3144, 1.9974, 1.5480, 2.4218, 1.6419], + device='cuda:2'), covar=tensor([0.3882, 0.1734, 0.5034, 0.2971, 0.1408, 0.2400, 0.1592, 0.4640], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0346, 0.0427, 0.0355, 0.0381, 0.0381, 0.0372, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:23:53,229 INFO [finetune.py:1010] (2/7) Epoch 15, validation: loss=0.1536, simple_loss=0.2258, pruned_loss=0.04063, over 2265189.00 frames. +2023-04-27 06:23:53,230 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 06:23:53,370 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3291, 1.7952, 2.1530, 2.8137, 2.1841, 1.6925, 1.7034, 2.1139], + device='cuda:2'), covar=tensor([0.3363, 0.3721, 0.1809, 0.2732, 0.2985, 0.2858, 0.4156, 0.2511], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0249, 0.0225, 0.0318, 0.0216, 0.0231, 0.0231, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 06:24:48,182 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0667, 2.4636, 0.9373, 1.3211, 1.8305, 1.2071, 3.0539, 1.5626], + device='cuda:2'), covar=tensor([0.0634, 0.0540, 0.0783, 0.1259, 0.0506, 0.1060, 0.0253, 0.0647], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0049, 0.0047, 0.0050, 0.0052, 0.0076, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 06:24:56,493 INFO [finetune.py:976] (2/7) Epoch 15, batch 50, loss[loss=0.1725, simple_loss=0.2477, pruned_loss=0.04866, over 4757.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2538, pruned_loss=0.05619, over 216497.04 frames. ], batch size: 28, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:25:05,176 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80241.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:25:05,227 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0260, 1.2664, 1.7724, 2.1017, 1.8143, 1.4246, 0.9828, 1.4854], + device='cuda:2'), covar=tensor([0.3542, 0.4160, 0.1928, 0.2492, 0.2903, 0.2976, 0.4744, 0.2372], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0249, 0.0225, 0.0317, 0.0216, 0.0230, 0.0231, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 06:25:08,721 INFO [optim.py:369] (2/7) 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 06:25:08,839 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6695, 1.5639, 0.7141, 1.3605, 1.5646, 1.4809, 1.4119, 1.4674], + device='cuda:2'), covar=tensor([0.0535, 0.0393, 0.0387, 0.0592, 0.0300, 0.0530, 0.0505, 0.0580], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 06:25:36,375 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 100, loss[loss=0.1932, simple_loss=0.2526, pruned_loss=0.06692, over 4824.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2499, pruned_loss=0.05585, over 382153.98 frames. ], batch size: 30, lr: 3.52e-03, grad_scale: 16.0 +2023-04-27 06:25:41,453 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:25:57,022 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80313.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:25:57,679 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.2317, 2.7004, 2.5343, 2.7333, 2.4928, 2.8208, 2.5475, 2.5146], + device='cuda:2'), covar=tensor([0.3717, 0.5955, 0.4850, 0.4241, 0.5437, 0.6260, 0.7071, 0.5623], + device='cuda:2'), in_proj_covar=tensor([0.0332, 0.0376, 0.0321, 0.0332, 0.0343, 0.0401, 0.0357, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 06:26:00,024 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6503, 1.2236, 0.5073, 1.2785, 1.3838, 1.4733, 1.3636, 1.3591], + device='cuda:2'), covar=tensor([0.0482, 0.0409, 0.0401, 0.0580, 0.0285, 0.0503, 0.0482, 0.0585], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0048, 0.0050], + device='cuda:2') +2023-04-27 06:26:12,989 INFO [finetune.py:976] (2/7) Epoch 15, batch 150, loss[loss=0.1542, simple_loss=0.2195, pruned_loss=0.04445, over 4821.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2453, pruned_loss=0.05456, over 508507.94 frames. ], batch size: 30, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:26:18,213 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 06:26:28,918 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 200, loss[loss=0.1581, simple_loss=0.2435, pruned_loss=0.03632, over 4917.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2444, pruned_loss=0.05453, over 609182.57 frames. ], batch size: 37, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:26:50,390 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80393.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:27:06,645 INFO [zipformer.py:1188] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.3162, 1.3435, 1.7036, 1.6264, 1.3036, 1.0410, 1.4221, 0.9749], + device='cuda:2'), covar=tensor([0.0723, 0.0773, 0.0481, 0.0921, 0.0971, 0.1470, 0.0701, 0.0738], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0071, 0.0070, 0.0068, 0.0076, 0.0098, 0.0076, 0.0070], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:27:16,984 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 06:27:18,653 INFO [finetune.py:976] (2/7) Epoch 15, batch 250, loss[loss=0.2067, simple_loss=0.2878, pruned_loss=0.06281, over 4834.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.249, pruned_loss=0.05624, over 686827.90 frames. ], batch size: 49, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:27:25,586 INFO [optim.py:369] (2/7) 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 06:27:31,401 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80454.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:27:36,889 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 15, batch 300, loss[loss=0.1643, simple_loss=0.2464, pruned_loss=0.04106, over 4826.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2537, pruned_loss=0.05782, over 745965.40 frames. ], batch size: 30, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:27:55,970 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80494.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:28:13,004 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4083, 1.8279, 1.6931, 2.0327, 1.9350, 1.9313, 1.7015, 4.3291], + device='cuda:2'), covar=tensor([0.0541, 0.0706, 0.0721, 0.1145, 0.0607, 0.0542, 0.0667, 0.0121], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 06:28:13,031 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9269, 1.9213, 2.2610, 2.3227, 1.8705, 1.6028, 1.9837, 1.1422], + device='cuda:2'), covar=tensor([0.0619, 0.0691, 0.0455, 0.0734, 0.0739, 0.1098, 0.0697, 0.0799], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0067, 0.0075, 0.0097, 0.0075, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:28:27,732 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80524.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:28:34,615 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 06:28:36,206 INFO [finetune.py:976] (2/7) Epoch 15, batch 350, loss[loss=0.1963, simple_loss=0.2704, pruned_loss=0.0611, over 4711.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2552, pruned_loss=0.0582, over 791858.24 frames. ], batch size: 59, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:28:42,201 INFO [optim.py:369] (2/7) 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 06:28:48,610 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:29:15,645 INFO [finetune.py:976] (2/7) Epoch 15, batch 400, loss[loss=0.1919, simple_loss=0.2677, pruned_loss=0.05805, over 4790.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2568, pruned_loss=0.05867, over 828158.06 frames. ], batch size: 51, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:30:07,689 INFO [finetune.py:976] (2/7) Epoch 15, batch 450, loss[loss=0.1445, simple_loss=0.207, pruned_loss=0.04103, over 4162.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2549, pruned_loss=0.05828, over 855681.55 frames. ], batch size: 18, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:30:13,732 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80639.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:30:18,556 INFO [optim.py:369] (2/7) 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 06:31:13,676 INFO [finetune.py:976] (2/7) Epoch 15, batch 500, loss[loss=0.207, simple_loss=0.2666, pruned_loss=0.07374, over 4143.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2517, pruned_loss=0.0571, over 876725.76 frames. ], batch size: 18, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:32:06,672 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:32:07,804 INFO [finetune.py:976] (2/7) Epoch 15, batch 550, loss[loss=0.1395, simple_loss=0.2077, pruned_loss=0.03568, over 4865.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2477, pruned_loss=0.05561, over 890743.85 frames. ], batch size: 31, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:32:13,247 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80749.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:32:19,340 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5247, 2.0265, 2.5246, 2.9584, 2.3571, 1.9175, 1.8802, 2.2667], + device='cuda:2'), covar=tensor([0.3702, 0.3644, 0.1764, 0.2873, 0.3115, 0.3027, 0.4205, 0.2452], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0246, 0.0223, 0.0315, 0.0214, 0.0229, 0.0229, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 06:32:23,342 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8663, 2.4930, 1.8388, 1.7586, 1.3846, 1.3823, 1.9056, 1.3423], + device='cuda:2'), covar=tensor([0.1915, 0.1416, 0.1545, 0.1982, 0.2457, 0.2244, 0.1063, 0.2215], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0212, 0.0169, 0.0204, 0.0200, 0.0184, 0.0156, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 06:32:32,595 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80773.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:32:41,567 INFO [finetune.py:976] (2/7) Epoch 15, batch 600, loss[loss=0.214, simple_loss=0.2824, pruned_loss=0.0728, over 4745.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2479, pruned_loss=0.05566, over 905015.66 frames. ], batch size: 54, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:32:47,146 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:33:03,052 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:33:12,683 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 650, loss[loss=0.2275, simple_loss=0.2889, pruned_loss=0.08309, over 4905.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2513, pruned_loss=0.05706, over 917951.69 frames. ], batch size: 43, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:33:20,402 INFO [optim.py:369] (2/7) 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 06:33:22,323 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:33:23,599 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7797, 1.9908, 2.0662, 2.1496, 1.9936, 2.1142, 2.0654, 2.0432], + device='cuda:2'), covar=tensor([0.4414, 0.7557, 0.5409, 0.5235, 0.6273, 0.7915, 0.7489, 0.6550], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0373, 0.0318, 0.0331, 0.0343, 0.0400, 0.0355, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 06:33:39,801 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5536, 1.3758, 1.9311, 1.8949, 1.4615, 1.1648, 1.5691, 1.0894], + device='cuda:2'), covar=tensor([0.0625, 0.1010, 0.0465, 0.0643, 0.0785, 0.1405, 0.0811, 0.0822], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0071, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:33:45,915 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-04-27 06:33:47,152 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.64 vs. limit=5.0 +2023-04-27 06:33:48,174 INFO [finetune.py:976] (2/7) Epoch 15, batch 700, loss[loss=0.2637, simple_loss=0.3351, pruned_loss=0.09612, over 4102.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2531, pruned_loss=0.05789, over 924390.43 frames. ], batch size: 65, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:34:21,753 INFO [finetune.py:976] (2/7) Epoch 15, batch 750, loss[loss=0.1741, simple_loss=0.2446, pruned_loss=0.05183, over 4742.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2552, pruned_loss=0.05868, over 930808.97 frames. ], batch size: 59, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:34:22,472 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80939.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:34:27,257 INFO [optim.py:369] (2/7) 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 06:34:54,975 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:34:55,526 INFO [finetune.py:976] (2/7) Epoch 15, batch 800, loss[loss=0.1438, simple_loss=0.2176, pruned_loss=0.03497, over 4784.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2549, pruned_loss=0.05787, over 936270.63 frames. ], batch size: 26, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:35:01,298 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-04-27 06:35:35,657 INFO [finetune.py:976] (2/7) Epoch 15, batch 850, loss[loss=0.2058, simple_loss=0.2751, pruned_loss=0.06825, over 4892.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2521, pruned_loss=0.05661, over 941890.60 frames. ], batch size: 35, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:35:46,470 INFO [optim.py:369] (2/7) 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 06:35:47,795 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 900, loss[loss=0.1619, simple_loss=0.2291, pruned_loss=0.04732, over 4928.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.249, pruned_loss=0.05533, over 945398.68 frames. ], batch size: 38, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:36:49,898 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:10,980 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3619, 1.2920, 1.6465, 1.5230, 1.2961, 1.1499, 1.3612, 0.8993], + device='cuda:2'), covar=tensor([0.0510, 0.0667, 0.0387, 0.0629, 0.0699, 0.1095, 0.0555, 0.0556], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0075, 0.0096, 0.0075, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:37:15,240 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:22,520 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81119.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:35,000 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:37:48,249 INFO [finetune.py:976] (2/7) Epoch 15, batch 950, loss[loss=0.1703, simple_loss=0.2476, pruned_loss=0.0465, over 4894.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2479, pruned_loss=0.05512, over 949470.55 frames. ], batch size: 35, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:38:04,672 INFO [optim.py:369] (2/7) 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 06:38:06,680 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81167.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:38:23,606 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81177.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:38:32,219 INFO [finetune.py:976] (2/7) Epoch 15, batch 1000, loss[loss=0.2011, simple_loss=0.265, pruned_loss=0.06863, over 4061.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2502, pruned_loss=0.05634, over 950806.25 frames. ], batch size: 65, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:38:38,868 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81198.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:38:50,016 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4898, 2.6350, 2.1228, 2.2349, 2.4879, 1.9944, 3.3155, 1.6659], + device='cuda:2'), covar=tensor([0.4207, 0.2170, 0.4765, 0.3243, 0.2308, 0.3122, 0.1598, 0.5334], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0350, 0.0430, 0.0357, 0.0385, 0.0385, 0.0375, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:39:05,805 INFO [finetune.py:976] (2/7) Epoch 15, batch 1050, loss[loss=0.1662, simple_loss=0.2468, pruned_loss=0.04278, over 4916.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2524, pruned_loss=0.05602, over 952291.10 frames. ], batch size: 42, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:39:11,771 INFO [optim.py:369] (2/7) 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 06:39:38,634 INFO [finetune.py:976] (2/7) Epoch 15, batch 1100, loss[loss=0.2078, simple_loss=0.2834, pruned_loss=0.06614, over 4819.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2525, pruned_loss=0.056, over 952870.63 frames. ], batch size: 38, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:39:46,760 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:39:47,948 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81301.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:40:11,234 INFO [finetune.py:976] (2/7) Epoch 15, batch 1150, loss[loss=0.1723, simple_loss=0.2477, pruned_loss=0.04846, over 4929.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2535, pruned_loss=0.05625, over 952113.12 frames. ], batch size: 41, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:40:13,590 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.39 vs. limit=5.0 +2023-04-27 06:40:18,551 INFO [optim.py:369] (2/7) 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 06:40:27,086 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81360.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:40:28,330 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81362.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:40:44,964 INFO [finetune.py:976] (2/7) Epoch 15, batch 1200, loss[loss=0.1607, simple_loss=0.2385, pruned_loss=0.04146, over 4904.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2524, pruned_loss=0.05609, over 951713.37 frames. ], batch size: 46, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:40:48,497 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81392.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:41:12,545 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 1250, loss[loss=0.2158, simple_loss=0.272, pruned_loss=0.07985, over 4759.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2515, pruned_loss=0.05679, over 953567.50 frames. ], batch size: 28, lr: 3.51e-03, grad_scale: 16.0 +2023-04-27 06:41:20,061 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81440.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:41:24,357 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 06:41:31,407 INFO [optim.py:369] (2/7) 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 06:42:05,041 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81472.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:42:08,092 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81477.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:42:08,875 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-04-27 06:42:20,542 INFO [finetune.py:976] (2/7) Epoch 15, batch 1300, loss[loss=0.1593, simple_loss=0.2365, pruned_loss=0.04108, over 4937.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2481, pruned_loss=0.05518, over 953399.69 frames. ], batch size: 42, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:42:41,577 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.3661, 1.2226, 1.3160, 0.9373, 1.3067, 1.1034, 1.5433, 1.3183], + device='cuda:2'), covar=tensor([0.3453, 0.1731, 0.4643, 0.2474, 0.1354, 0.2059, 0.1876, 0.4549], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0348, 0.0428, 0.0356, 0.0382, 0.0383, 0.0374, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:43:12,244 INFO [finetune.py:976] (2/7) Epoch 15, batch 1350, loss[loss=0.1699, simple_loss=0.2317, pruned_loss=0.05403, over 4713.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2476, pruned_loss=0.05515, over 955707.03 frames. ], batch size: 23, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:43:21,915 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9492, 1.4874, 1.4435, 1.6584, 2.1413, 1.7342, 1.4356, 1.4082], + device='cuda:2'), covar=tensor([0.1466, 0.1534, 0.1840, 0.1393, 0.0916, 0.1638, 0.2217, 0.1993], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0317, 0.0353, 0.0292, 0.0330, 0.0314, 0.0304, 0.0361], + device='cuda:2'), out_proj_covar=tensor([6.4260e-05, 6.6360e-05, 7.5630e-05, 5.9751e-05, 6.8722e-05, 6.6471e-05, + 6.4340e-05, 7.7210e-05], device='cuda:2') +2023-04-27 06:43:24,613 INFO [optim.py:369] (2/7) 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 06:43:54,406 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4614, 1.3563, 1.6586, 1.6433, 1.3437, 1.2703, 1.3793, 0.8393], + device='cuda:2'), covar=tensor([0.0558, 0.0660, 0.0467, 0.0627, 0.0792, 0.1214, 0.0616, 0.0684], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0070, 0.0069, 0.0067, 0.0075, 0.0096, 0.0074, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:43:56,190 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7781, 1.2716, 1.3209, 1.5328, 1.9798, 1.5687, 1.3441, 1.2648], + device='cuda:2'), covar=tensor([0.1561, 0.1596, 0.2019, 0.1374, 0.0877, 0.1665, 0.1962, 0.2069], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0317, 0.0353, 0.0292, 0.0330, 0.0314, 0.0304, 0.0362], + device='cuda:2'), out_proj_covar=tensor([6.4275e-05, 6.6279e-05, 7.5661e-05, 5.9657e-05, 6.8703e-05, 6.6402e-05, + 6.4306e-05, 7.7247e-05], device='cuda:2') +2023-04-27 06:44:07,693 INFO [finetune.py:976] (2/7) Epoch 15, batch 1400, loss[loss=0.1433, simple_loss=0.2146, pruned_loss=0.03605, over 4188.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2514, pruned_loss=0.05629, over 955981.61 frames. ], batch size: 18, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:44:21,309 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8969, 2.9268, 2.2350, 3.2907, 2.8662, 2.8957, 1.1797, 2.7208], + device='cuda:2'), covar=tensor([0.2289, 0.1626, 0.3264, 0.2717, 0.4136, 0.2351, 0.6074, 0.3151], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0213, 0.0249, 0.0301, 0.0296, 0.0246, 0.0269, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 06:44:21,947 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4004, 1.2422, 4.0534, 3.7582, 3.5865, 3.8998, 3.7895, 3.5727], + device='cuda:2'), covar=tensor([0.7323, 0.6211, 0.1147, 0.1896, 0.1303, 0.1743, 0.1835, 0.1539], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0306, 0.0401, 0.0403, 0.0348, 0.0404, 0.0313, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:44:41,266 INFO [finetune.py:976] (2/7) Epoch 15, batch 1450, loss[loss=0.1738, simple_loss=0.2506, pruned_loss=0.04856, over 4827.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.253, pruned_loss=0.05677, over 954685.66 frames. ], batch size: 47, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:44:46,722 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6602, 1.5428, 0.8782, 1.3276, 1.8326, 1.4964, 1.4019, 1.3789], + device='cuda:2'), covar=tensor([0.0487, 0.0358, 0.0331, 0.0514, 0.0269, 0.0488, 0.0468, 0.0545], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0037, 0.0050, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 06:44:47,199 INFO [optim.py:369] (2/7) 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 06:44:53,607 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81655.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:44:55,289 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81657.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:44:56,025 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-04-27 06:45:00,731 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6805, 2.2540, 1.6989, 1.6000, 1.2839, 1.2719, 1.8313, 1.2357], + device='cuda:2'), covar=tensor([0.1506, 0.1289, 0.1445, 0.1720, 0.2315, 0.1838, 0.0901, 0.2002], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0212, 0.0168, 0.0203, 0.0200, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 06:45:04,234 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81670.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:45:15,082 INFO [finetune.py:976] (2/7) Epoch 15, batch 1500, loss[loss=0.1795, simple_loss=0.2566, pruned_loss=0.05122, over 4879.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2537, pruned_loss=0.05683, over 954412.43 frames. ], batch size: 32, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:45:31,512 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6723, 1.0331, 1.6604, 2.1284, 1.7323, 1.5953, 1.6431, 1.6359], + device='cuda:2'), covar=tensor([0.4747, 0.6849, 0.6798, 0.5903, 0.6279, 0.8149, 0.8372, 0.8311], + device='cuda:2'), in_proj_covar=tensor([0.0412, 0.0404, 0.0492, 0.0504, 0.0441, 0.0463, 0.0470, 0.0471], + device='cuda:2'), out_proj_covar=tensor([9.9482e-05, 9.9947e-05, 1.1059e-04, 1.1994e-04, 1.0611e-04, 1.1161e-04, + 1.1207e-04, 1.1215e-04], device='cuda:2') +2023-04-27 06:45:44,083 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81731.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 06:45:48,175 INFO [finetune.py:976] (2/7) Epoch 15, batch 1550, loss[loss=0.1715, simple_loss=0.2438, pruned_loss=0.04961, over 4839.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2531, pruned_loss=0.05616, over 956733.62 frames. ], batch size: 44, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:45:48,347 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 06:45:53,673 INFO [optim.py:369] (2/7) 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 06:46:11,814 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 1600, loss[loss=0.1841, simple_loss=0.2503, pruned_loss=0.05898, over 4911.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2505, pruned_loss=0.05556, over 958409.26 frames. ], batch size: 46, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:46:42,456 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:46:53,851 INFO [finetune.py:976] (2/7) Epoch 15, batch 1650, loss[loss=0.1449, simple_loss=0.2113, pruned_loss=0.03928, over 4824.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2469, pruned_loss=0.05448, over 957201.25 frames. ], batch size: 25, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:47:04,628 INFO [optim.py:369] (2/7) 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 06:47:59,312 INFO [finetune.py:976] (2/7) Epoch 15, batch 1700, loss[loss=0.1837, simple_loss=0.2617, pruned_loss=0.05283, over 4751.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2452, pruned_loss=0.05396, over 954962.45 frames. ], batch size: 54, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:48:02,580 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 06:49:06,082 INFO [finetune.py:976] (2/7) Epoch 15, batch 1750, loss[loss=0.2325, simple_loss=0.3119, pruned_loss=0.07656, over 4811.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2487, pruned_loss=0.056, over 954667.62 frames. ], batch size: 38, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:49:16,873 INFO [optim.py:369] (2/7) 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 06:49:17,019 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5599, 2.1652, 1.7786, 1.5537, 1.1603, 1.1728, 1.9490, 1.2173], + device='cuda:2'), covar=tensor([0.1930, 0.1600, 0.1473, 0.1934, 0.2620, 0.2219, 0.0973, 0.2295], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0210, 0.0167, 0.0202, 0.0199, 0.0182, 0.0154, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 06:49:19,528 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-27 06:49:26,309 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81955.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:49:27,517 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81957.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:49:48,416 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:49:48,923 INFO [finetune.py:976] (2/7) Epoch 15, batch 1800, loss[loss=0.1841, simple_loss=0.2619, pruned_loss=0.05312, over 4800.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.252, pruned_loss=0.05621, over 953960.60 frames. ], batch size: 51, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:49:59,487 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82003.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:00,739 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82005.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:15,244 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3290, 1.7111, 1.7942, 1.8305, 2.3512, 1.9930, 1.7300, 1.6933], + device='cuda:2'), covar=tensor([0.1354, 0.1403, 0.1943, 0.1404, 0.0886, 0.1416, 0.1923, 0.2198], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0317, 0.0352, 0.0292, 0.0330, 0.0314, 0.0304, 0.0363], + device='cuda:2'), out_proj_covar=tensor([6.4340e-05, 6.6279e-05, 7.5447e-05, 5.9589e-05, 6.8724e-05, 6.6377e-05, + 6.4348e-05, 7.7523e-05], device='cuda:2') +2023-04-27 06:50:16,418 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82026.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 06:50:24,044 INFO [finetune.py:976] (2/7) Epoch 15, batch 1850, loss[loss=0.2339, simple_loss=0.2981, pruned_loss=0.08485, over 4807.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2527, pruned_loss=0.05684, over 954643.10 frames. ], batch size: 45, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:50:29,487 INFO [optim.py:369] (2/7) 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 06:50:30,245 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82048.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:34,527 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:50:56,620 INFO [finetune.py:976] (2/7) Epoch 15, batch 1900, loss[loss=0.2269, simple_loss=0.2917, pruned_loss=0.08108, over 4924.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.253, pruned_loss=0.05663, over 952759.93 frames. ], batch size: 41, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:51:14,303 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:51:30,519 INFO [finetune.py:976] (2/7) Epoch 15, batch 1950, loss[loss=0.1987, simple_loss=0.2632, pruned_loss=0.06713, over 4865.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2509, pruned_loss=0.05604, over 953049.40 frames. ], batch size: 31, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:51:36,465 INFO [optim.py:369] (2/7) 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 06:52:04,205 INFO [finetune.py:976] (2/7) Epoch 15, batch 2000, loss[loss=0.1854, simple_loss=0.2497, pruned_loss=0.06056, over 4826.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2493, pruned_loss=0.05584, over 954023.11 frames. ], batch size: 30, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:52:11,643 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3918, 1.7432, 1.8847, 1.9696, 1.8179, 1.8765, 1.9702, 1.9594], + device='cuda:2'), covar=tensor([0.4142, 0.6050, 0.4970, 0.5145, 0.5962, 0.7701, 0.5399, 0.5246], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0374, 0.0318, 0.0334, 0.0343, 0.0400, 0.0353, 0.0327], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 06:52:37,841 INFO [finetune.py:976] (2/7) Epoch 15, batch 2050, loss[loss=0.2036, simple_loss=0.2812, pruned_loss=0.06298, over 4871.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2459, pruned_loss=0.05456, over 955178.51 frames. ], batch size: 31, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:52:43,309 INFO [optim.py:369] (2/7) 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 06:52:49,554 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-04-27 06:52:55,571 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3414, 1.8870, 2.2477, 2.6237, 2.2909, 1.7730, 1.4810, 1.9919], + device='cuda:2'), covar=tensor([0.3387, 0.3222, 0.1751, 0.2207, 0.2393, 0.2682, 0.3971, 0.1999], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0246, 0.0223, 0.0315, 0.0215, 0.0229, 0.0230, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 06:53:21,146 INFO [finetune.py:976] (2/7) Epoch 15, batch 2100, loss[loss=0.2049, simple_loss=0.2773, pruned_loss=0.06625, over 4815.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.245, pruned_loss=0.05402, over 955459.09 frames. ], batch size: 51, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:54:04,956 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82326.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:54:24,279 INFO [finetune.py:976] (2/7) Epoch 15, batch 2150, loss[loss=0.1791, simple_loss=0.2428, pruned_loss=0.05771, over 4816.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2483, pruned_loss=0.05488, over 954892.06 frames. ], batch size: 38, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:54:33,139 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82346.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:54:35,530 INFO [optim.py:369] (2/7) 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 06:54:36,414 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-04-27 06:54:49,974 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 06:54:54,082 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5442, 2.4122, 2.6386, 3.0527, 2.9239, 2.3858, 2.0432, 2.5682], + device='cuda:2'), covar=tensor([0.0826, 0.0864, 0.0530, 0.0498, 0.0570, 0.0752, 0.0717, 0.0518], + device='cuda:2'), in_proj_covar=tensor([0.0190, 0.0202, 0.0182, 0.0173, 0.0177, 0.0182, 0.0154, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 06:54:58,342 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:55:13,107 INFO [finetune.py:976] (2/7) Epoch 15, batch 2200, loss[loss=0.2124, simple_loss=0.2952, pruned_loss=0.06482, over 4804.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2511, pruned_loss=0.05563, over 956518.34 frames. ], batch size: 41, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:55:42,632 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82407.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:55:45,062 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82411.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:55:45,821 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 06:56:12,256 INFO [finetune.py:976] (2/7) Epoch 15, batch 2250, loss[loss=0.2122, simple_loss=0.2768, pruned_loss=0.07383, over 4824.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2527, pruned_loss=0.05688, over 954776.67 frames. ], batch size: 33, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:56:12,400 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4643, 1.7807, 1.7628, 1.9173, 1.7358, 1.8507, 1.9047, 1.8136], + device='cuda:2'), covar=tensor([0.3757, 0.5746, 0.5083, 0.4532, 0.5701, 0.7562, 0.5749, 0.5404], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0375, 0.0319, 0.0335, 0.0343, 0.0401, 0.0355, 0.0327], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 06:56:19,194 INFO [optim.py:369] (2/7) 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 06:56:42,002 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7027, 2.7155, 2.1286, 3.0804, 2.6825, 2.7048, 1.1828, 2.6282], + device='cuda:2'), covar=tensor([0.2012, 0.1495, 0.2821, 0.2381, 0.3040, 0.1895, 0.5243, 0.2917], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0215, 0.0251, 0.0304, 0.0299, 0.0248, 0.0272, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 06:56:45,464 INFO [finetune.py:976] (2/7) Epoch 15, batch 2300, loss[loss=0.1926, simple_loss=0.2551, pruned_loss=0.0651, over 4870.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2526, pruned_loss=0.05654, over 955111.59 frames. ], batch size: 31, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:56:49,047 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82493.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:56:50,237 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-27 06:57:18,143 INFO [finetune.py:976] (2/7) Epoch 15, batch 2350, loss[loss=0.1641, simple_loss=0.2271, pruned_loss=0.05057, over 4903.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2501, pruned_loss=0.05562, over 951841.51 frames. ], batch size: 36, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:57:25,042 INFO [optim.py:369] (2/7) 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 06:57:30,335 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82554.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 06:57:40,475 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 2400, loss[loss=0.1839, simple_loss=0.2526, pruned_loss=0.05762, over 4828.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2467, pruned_loss=0.05445, over 954251.84 frames. ], batch size: 39, lr: 3.50e-03, grad_scale: 32.0 +2023-04-27 06:58:13,419 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-27 06:58:20,996 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82630.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:58:25,777 INFO [finetune.py:976] (2/7) Epoch 15, batch 2450, loss[loss=0.2326, simple_loss=0.2897, pruned_loss=0.08774, over 4902.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2452, pruned_loss=0.05402, over 954409.97 frames. ], batch size: 32, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 06:58:28,845 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:58:31,161 INFO [optim.py:369] (2/7) 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 06:58:40,474 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7683, 1.6319, 1.9514, 2.0711, 1.5766, 1.3905, 1.6652, 1.0618], + device='cuda:2'), covar=tensor([0.0470, 0.0820, 0.0410, 0.0601, 0.0745, 0.1055, 0.0581, 0.0707], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0069, 0.0069, 0.0067, 0.0074, 0.0095, 0.0074, 0.0068], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 06:59:04,412 INFO [finetune.py:976] (2/7) Epoch 15, batch 2500, loss[loss=0.1992, simple_loss=0.2728, pruned_loss=0.0628, over 4826.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2479, pruned_loss=0.05505, over 956643.34 frames. ], batch size: 38, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 06:59:11,878 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82702.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 06:59:37,991 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82711.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:00:09,107 INFO [finetune.py:976] (2/7) Epoch 15, batch 2550, loss[loss=0.1632, simple_loss=0.2364, pruned_loss=0.04497, over 4831.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2517, pruned_loss=0.05565, over 956740.78 frames. ], batch size: 33, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:00:09,789 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8011, 3.6834, 2.7626, 4.3988, 3.7810, 3.8155, 1.6713, 3.8093], + device='cuda:2'), covar=tensor([0.1763, 0.1267, 0.3100, 0.1547, 0.3394, 0.1717, 0.5662, 0.2320], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0213, 0.0249, 0.0301, 0.0297, 0.0246, 0.0269, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 07:00:14,537 INFO [optim.py:369] (2/7) 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 07:00:24,400 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82759.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:00:45,085 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7906, 1.1835, 1.7790, 2.2802, 1.9076, 1.7279, 1.7738, 1.7558], + device='cuda:2'), covar=tensor([0.4744, 0.6509, 0.6521, 0.6038, 0.6511, 0.7873, 0.8023, 0.7133], + device='cuda:2'), in_proj_covar=tensor([0.0417, 0.0408, 0.0497, 0.0510, 0.0446, 0.0469, 0.0476, 0.0479], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:00:47,932 INFO [finetune.py:976] (2/7) Epoch 15, batch 2600, loss[loss=0.2134, simple_loss=0.284, pruned_loss=0.07141, over 4822.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2533, pruned_loss=0.05649, over 955067.45 frames. ], batch size: 30, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:01:05,804 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9730, 2.6262, 1.0148, 1.3048, 1.9026, 1.2410, 3.4083, 1.7200], + device='cuda:2'), covar=tensor([0.0688, 0.0643, 0.0840, 0.1360, 0.0549, 0.1051, 0.0221, 0.0653], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 07:01:43,239 INFO [finetune.py:976] (2/7) Epoch 15, batch 2650, loss[loss=0.2108, simple_loss=0.2776, pruned_loss=0.07198, over 4739.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2527, pruned_loss=0.05681, over 954751.51 frames. ], batch size: 59, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:01:48,662 INFO [optim.py:369] (2/7) 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 07:01:49,965 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82849.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:02:33,354 INFO [finetune.py:976] (2/7) Epoch 15, batch 2700, loss[loss=0.1898, simple_loss=0.2514, pruned_loss=0.06411, over 4920.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2517, pruned_loss=0.05591, over 956207.00 frames. ], batch size: 38, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:03:04,630 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:03:08,339 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4581, 3.4154, 0.9552, 2.0187, 1.9099, 2.6287, 1.9371, 0.9381], + device='cuda:2'), covar=tensor([0.1298, 0.0837, 0.1904, 0.1100, 0.0983, 0.0838, 0.1452, 0.2098], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0244, 0.0136, 0.0120, 0.0130, 0.0151, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:03:11,575 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.74 vs. limit=2.0 +2023-04-27 07:03:12,571 INFO [finetune.py:976] (2/7) Epoch 15, batch 2750, loss[loss=0.159, simple_loss=0.2261, pruned_loss=0.04598, over 4755.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.249, pruned_loss=0.05522, over 954827.79 frames. ], batch size: 27, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:03:18,528 INFO [optim.py:369] (2/7) 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 07:03:28,516 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82963.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:03:46,114 INFO [finetune.py:976] (2/7) Epoch 15, batch 2800, loss[loss=0.1776, simple_loss=0.239, pruned_loss=0.05807, over 4931.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2453, pruned_loss=0.05386, over 954986.88 frames. ], batch size: 38, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:03:55,389 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83002.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:04:09,789 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83024.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:04:19,799 INFO [finetune.py:976] (2/7) Epoch 15, batch 2850, loss[loss=0.2135, simple_loss=0.2876, pruned_loss=0.06975, over 4841.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2449, pruned_loss=0.05424, over 955570.73 frames. ], batch size: 49, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:04:24,178 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4487, 3.4103, 1.0351, 1.8896, 2.0268, 2.6273, 1.9507, 1.0326], + device='cuda:2'), covar=tensor([0.1465, 0.1099, 0.1991, 0.1287, 0.1010, 0.0897, 0.1615, 0.1970], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0246, 0.0137, 0.0121, 0.0131, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:04:25,721 INFO [optim.py:369] (2/7) 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 07:04:27,622 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83050.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:04:33,196 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2100, 1.5460, 1.3593, 1.4489, 1.2822, 1.2511, 1.3247, 1.0205], + device='cuda:2'), covar=tensor([0.1840, 0.1349, 0.0999, 0.1244, 0.3531, 0.1378, 0.1778, 0.2397], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0312, 0.0224, 0.0283, 0.0315, 0.0268, 0.0256, 0.0271], + device='cuda:2'), out_proj_covar=tensor([1.1706e-04, 1.2434e-04, 8.9343e-05, 1.1279e-04, 1.2849e-04, 1.0700e-04, + 1.0347e-04, 1.0812e-04], device='cuda:2') +2023-04-27 07:05:03,759 INFO [finetune.py:976] (2/7) Epoch 15, batch 2900, loss[loss=0.1541, simple_loss=0.2386, pruned_loss=0.0348, over 4810.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2486, pruned_loss=0.05516, over 955359.30 frames. ], batch size: 39, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:06:07,046 INFO [finetune.py:976] (2/7) Epoch 15, batch 2950, loss[loss=0.1689, simple_loss=0.2425, pruned_loss=0.04764, over 4861.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2533, pruned_loss=0.0568, over 953863.80 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:06:18,125 INFO [optim.py:369] (2/7) 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 07:06:19,467 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83149.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:06:56,294 INFO [finetune.py:976] (2/7) Epoch 15, batch 3000, loss[loss=0.1829, simple_loss=0.2709, pruned_loss=0.04746, over 4815.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2549, pruned_loss=0.05771, over 954052.20 frames. ], batch size: 39, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:06:56,294 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 07:07:06,875 INFO [finetune.py:1010] (2/7) Epoch 15, validation: loss=0.1516, simple_loss=0.2237, pruned_loss=0.03975, over 2265189.00 frames. +2023-04-27 07:07:06,876 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 07:07:09,683 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-27 07:07:12,339 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83197.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:07:13,544 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5065, 1.5879, 0.7330, 1.2323, 1.5346, 1.3975, 1.2877, 1.3667], + device='cuda:2'), covar=tensor([0.0520, 0.0374, 0.0388, 0.0572, 0.0299, 0.0519, 0.0490, 0.0590], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 07:07:30,742 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:07:32,597 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9992, 2.7959, 1.9584, 2.0940, 1.3712, 1.3662, 2.1292, 1.3796], + device='cuda:2'), covar=tensor([0.1587, 0.1448, 0.1371, 0.1671, 0.2346, 0.1997, 0.0969, 0.2069], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0215, 0.0169, 0.0206, 0.0202, 0.0185, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 07:07:38,611 INFO [finetune.py:976] (2/7) Epoch 15, batch 3050, loss[loss=0.1985, simple_loss=0.2834, pruned_loss=0.05675, over 4770.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2551, pruned_loss=0.05726, over 955335.04 frames. ], batch size: 51, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:07:50,803 INFO [optim.py:369] (2/7) 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 07:08:02,516 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:08:36,038 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 3100, loss[loss=0.1835, simple_loss=0.2491, pruned_loss=0.05899, over 4842.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2519, pruned_loss=0.05571, over 954521.96 frames. ], batch size: 47, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:09:10,938 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9745, 1.2141, 4.7305, 4.4623, 4.1361, 4.3984, 4.2335, 4.1970], + device='cuda:2'), covar=tensor([0.6608, 0.6037, 0.0923, 0.1723, 0.0999, 0.1125, 0.1697, 0.1354], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0308, 0.0402, 0.0407, 0.0348, 0.0405, 0.0314, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:09:15,039 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:09:26,582 INFO [finetune.py:976] (2/7) Epoch 15, batch 3150, loss[loss=0.1631, simple_loss=0.24, pruned_loss=0.04308, over 4744.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2486, pruned_loss=0.05488, over 955125.50 frames. ], batch size: 27, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:09:31,012 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83345.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:09:32,641 INFO [optim.py:369] (2/7) 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 07:09:37,914 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4489, 1.4073, 1.4588, 1.0383, 1.3851, 1.2138, 1.8162, 1.4188], + device='cuda:2'), covar=tensor([0.3378, 0.1640, 0.4413, 0.2387, 0.1392, 0.1929, 0.1456, 0.4063], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0349, 0.0429, 0.0357, 0.0386, 0.0384, 0.0374, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:09:54,602 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-04-27 07:09:59,876 INFO [finetune.py:976] (2/7) Epoch 15, batch 3200, loss[loss=0.1586, simple_loss=0.2359, pruned_loss=0.04067, over 4935.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2453, pruned_loss=0.05373, over 954565.49 frames. ], batch size: 38, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:10:45,247 INFO [finetune.py:976] (2/7) Epoch 15, batch 3250, loss[loss=0.1548, simple_loss=0.2279, pruned_loss=0.04089, over 4768.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2467, pruned_loss=0.0548, over 953554.37 frames. ], batch size: 26, lr: 3.49e-03, grad_scale: 32.0 +2023-04-27 07:10:54,481 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1428, 1.6439, 1.9687, 2.2645, 1.9927, 1.5654, 0.9879, 1.7083], + device='cuda:2'), covar=tensor([0.3023, 0.3103, 0.1575, 0.2099, 0.2424, 0.2658, 0.4605, 0.2057], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0246, 0.0224, 0.0317, 0.0216, 0.0229, 0.0230, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 07:10:56,132 INFO [optim.py:369] (2/7) 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 07:11:16,625 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-04-27 07:11:47,159 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9019, 1.5476, 1.7374, 2.2567, 2.1794, 1.8846, 1.5131, 2.0944], + device='cuda:2'), covar=tensor([0.0715, 0.1114, 0.0684, 0.0426, 0.0546, 0.0673, 0.0777, 0.0424], + device='cuda:2'), in_proj_covar=tensor([0.0189, 0.0201, 0.0181, 0.0172, 0.0177, 0.0182, 0.0154, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:11:50,144 INFO [finetune.py:976] (2/7) Epoch 15, batch 3300, loss[loss=0.1782, simple_loss=0.2532, pruned_loss=0.05161, over 4799.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2504, pruned_loss=0.05628, over 953419.30 frames. ], batch size: 45, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:12:14,192 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9682, 2.1945, 1.8131, 1.6537, 2.0030, 1.5315, 2.6824, 1.3916], + device='cuda:2'), covar=tensor([0.3828, 0.1619, 0.4300, 0.2913, 0.1818, 0.2920, 0.1246, 0.4833], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0347, 0.0427, 0.0357, 0.0384, 0.0383, 0.0372, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:12:23,029 INFO [finetune.py:976] (2/7) Epoch 15, batch 3350, loss[loss=0.1547, simple_loss=0.2256, pruned_loss=0.04188, over 4836.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2521, pruned_loss=0.05656, over 954244.49 frames. ], batch size: 30, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:12:27,401 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6768, 2.1127, 1.6087, 1.5009, 1.2550, 1.2965, 1.6530, 1.1958], + device='cuda:2'), covar=tensor([0.1641, 0.1402, 0.1459, 0.1776, 0.2423, 0.1928, 0.1047, 0.2054], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0215, 0.0169, 0.0205, 0.0202, 0.0185, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 07:12:28,445 INFO [optim.py:369] (2/7) 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 07:12:31,577 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83552.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:12:33,674 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 07:12:56,723 INFO [finetune.py:976] (2/7) Epoch 15, batch 3400, loss[loss=0.1822, simple_loss=0.2536, pruned_loss=0.05533, over 4858.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2533, pruned_loss=0.05704, over 956199.17 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:13:18,180 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83619.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:13:21,840 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6134, 1.6648, 0.9013, 1.2969, 1.7959, 1.4868, 1.3490, 1.4110], + device='cuda:2'), covar=tensor([0.0468, 0.0344, 0.0343, 0.0529, 0.0275, 0.0464, 0.0463, 0.0522], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0029, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0050, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 07:13:30,156 INFO [finetune.py:976] (2/7) Epoch 15, batch 3450, loss[loss=0.1608, simple_loss=0.2295, pruned_loss=0.04607, over 4877.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.253, pruned_loss=0.0567, over 954194.45 frames. ], batch size: 32, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:13:31,411 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 07:13:49,257 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83667.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:14:13,518 INFO [finetune.py:976] (2/7) Epoch 15, batch 3500, loss[loss=0.1856, simple_loss=0.2465, pruned_loss=0.06238, over 4823.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2513, pruned_loss=0.05661, over 953042.45 frames. ], batch size: 38, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:15:08,363 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4430, 2.1344, 2.3518, 3.0588, 2.4594, 2.0161, 1.8144, 2.3147], + device='cuda:2'), covar=tensor([0.3322, 0.3163, 0.1541, 0.1863, 0.2581, 0.2461, 0.3818, 0.2037], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0247, 0.0225, 0.0318, 0.0216, 0.0229, 0.0231, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 07:15:13,168 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4989, 3.6131, 0.9532, 1.9206, 2.0762, 2.4438, 2.0273, 1.0502], + device='cuda:2'), covar=tensor([0.1384, 0.0888, 0.2047, 0.1260, 0.0955, 0.1076, 0.1500, 0.1944], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0247, 0.0138, 0.0122, 0.0132, 0.0154, 0.0119, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:15:15,482 INFO [finetune.py:976] (2/7) Epoch 15, batch 3550, loss[loss=0.1449, simple_loss=0.226, pruned_loss=0.03194, over 4830.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2478, pruned_loss=0.05569, over 954090.47 frames. ], batch size: 41, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:15:21,399 INFO [optim.py:369] (2/7) 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 07:15:26,889 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0139, 1.3852, 5.3015, 5.0114, 4.6251, 5.0882, 4.6931, 4.6958], + device='cuda:2'), covar=tensor([0.6880, 0.6482, 0.1029, 0.1614, 0.0999, 0.1375, 0.1011, 0.1488], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0306, 0.0400, 0.0404, 0.0347, 0.0404, 0.0313, 0.0359], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:15:29,358 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83760.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:15:49,271 INFO [finetune.py:976] (2/7) Epoch 15, batch 3600, loss[loss=0.1572, simple_loss=0.2361, pruned_loss=0.03913, over 4761.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.246, pruned_loss=0.05542, over 956163.81 frames. ], batch size: 28, lr: 3.49e-03, grad_scale: 64.0 +2023-04-27 07:16:21,331 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.4461, 4.3504, 3.0892, 5.1186, 4.4067, 4.4119, 1.8184, 4.3414], + device='cuda:2'), covar=tensor([0.1555, 0.0966, 0.3327, 0.0875, 0.3445, 0.1479, 0.5790, 0.2018], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0215, 0.0251, 0.0303, 0.0298, 0.0248, 0.0270, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 07:16:32,029 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:16:35,003 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2071, 1.2219, 1.2934, 1.5728, 1.5615, 1.2756, 0.8517, 1.4623], + device='cuda:2'), covar=tensor([0.0881, 0.1334, 0.0884, 0.0593, 0.0691, 0.0762, 0.0930, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0199, 0.0179, 0.0169, 0.0175, 0.0180, 0.0152, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:16:55,917 INFO [finetune.py:976] (2/7) Epoch 15, batch 3650, loss[loss=0.1315, simple_loss=0.2054, pruned_loss=0.02885, over 4773.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2477, pruned_loss=0.05548, over 956078.17 frames. ], batch size: 28, lr: 3.48e-03, grad_scale: 64.0 +2023-04-27 07:17:01,432 INFO [optim.py:369] (2/7) 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 07:17:05,012 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 3700, loss[loss=0.1889, simple_loss=0.2622, pruned_loss=0.05781, over 4936.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2518, pruned_loss=0.05616, over 955973.90 frames. ], batch size: 33, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:17:37,202 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83900.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:17:43,732 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3810, 1.5494, 1.4792, 1.7787, 1.6487, 1.8324, 1.4308, 3.4422], + device='cuda:2'), covar=tensor([0.0576, 0.0782, 0.0808, 0.1173, 0.0638, 0.0501, 0.0708, 0.0161], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 07:17:48,039 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6747, 1.3051, 4.3847, 4.1196, 3.8227, 4.1294, 4.0952, 3.8790], + device='cuda:2'), covar=tensor([0.6451, 0.5882, 0.1076, 0.1623, 0.1038, 0.1288, 0.1255, 0.1508], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0307, 0.0403, 0.0405, 0.0350, 0.0407, 0.0315, 0.0362], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:17:57,615 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 07:18:03,516 INFO [finetune.py:976] (2/7) Epoch 15, batch 3750, loss[loss=0.2165, simple_loss=0.278, pruned_loss=0.07747, over 4891.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2531, pruned_loss=0.05664, over 955095.40 frames. ], batch size: 35, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:18:04,863 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 07:18:10,300 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83949.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:18:34,096 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 3800, loss[loss=0.1519, simple_loss=0.2214, pruned_loss=0.04122, over 4727.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2538, pruned_loss=0.05681, over 953112.70 frames. ], batch size: 54, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:18:36,354 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83988.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:18:39,341 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5366, 1.4539, 1.7763, 1.7098, 1.3412, 1.2873, 1.4663, 1.0405], + device='cuda:2'), covar=tensor([0.0587, 0.0696, 0.0437, 0.0692, 0.0858, 0.1209, 0.0625, 0.0670], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:18:39,348 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7545, 1.2512, 1.4359, 1.4465, 1.9403, 1.5544, 1.2172, 1.3718], + device='cuda:2'), covar=tensor([0.1664, 0.1532, 0.2055, 0.1502, 0.0917, 0.1515, 0.2439, 0.2274], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0318, 0.0353, 0.0293, 0.0332, 0.0315, 0.0305, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.4213e-05, 6.6490e-05, 7.5490e-05, 5.9762e-05, 6.9097e-05, 6.6576e-05, + 6.4671e-05, 7.7777e-05], device='cuda:2') +2023-04-27 07:18:51,954 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84010.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:19:10,672 INFO [finetune.py:976] (2/7) Epoch 15, batch 3850, loss[loss=0.1836, simple_loss=0.2598, pruned_loss=0.05367, over 4850.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2528, pruned_loss=0.05656, over 954033.42 frames. ], batch size: 44, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:19:21,659 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84046.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:19:22,756 INFO [optim.py:369] (2/7) 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 07:19:31,116 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3360, 1.9706, 2.4127, 2.7408, 2.7555, 2.3185, 1.8741, 2.3526], + device='cuda:2'), covar=tensor([0.0852, 0.1071, 0.0612, 0.0588, 0.0688, 0.0780, 0.0802, 0.0586], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0199, 0.0178, 0.0170, 0.0175, 0.0180, 0.0151, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:20:15,362 INFO [finetune.py:976] (2/7) Epoch 15, batch 3900, loss[loss=0.1718, simple_loss=0.233, pruned_loss=0.0553, over 4751.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2495, pruned_loss=0.05541, over 953353.50 frames. ], batch size: 27, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:20:27,909 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:20:50,575 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:21:20,860 INFO [finetune.py:976] (2/7) Epoch 15, batch 3950, loss[loss=0.1783, simple_loss=0.2405, pruned_loss=0.05805, over 4822.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2465, pruned_loss=0.05435, over 954388.20 frames. ], batch size: 33, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:21:34,734 INFO [optim.py:369] (2/7) 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 07:21:37,192 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9136, 2.0808, 1.8122, 1.4682, 1.4875, 1.4883, 1.8259, 1.4266], + device='cuda:2'), covar=tensor([0.1505, 0.1537, 0.1365, 0.1780, 0.2034, 0.1805, 0.0909, 0.1917], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0216, 0.0170, 0.0206, 0.0203, 0.0186, 0.0157, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 07:21:41,489 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:21:42,700 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3390, 3.0955, 2.4945, 2.7945, 2.1294, 2.5647, 2.6236, 1.9065], + device='cuda:2'), covar=tensor([0.2141, 0.1136, 0.0754, 0.1262, 0.3289, 0.1133, 0.2081, 0.2864], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0311, 0.0221, 0.0284, 0.0313, 0.0267, 0.0254, 0.0268], + device='cuda:2'), out_proj_covar=tensor([1.1673e-04, 1.2365e-04, 8.8030e-05, 1.1282e-04, 1.2769e-04, 1.0643e-04, + 1.0280e-04, 1.0694e-04], device='cuda:2') +2023-04-27 07:21:48,129 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4384, 1.6558, 1.6131, 2.1748, 2.4681, 2.0126, 1.8638, 1.6272], + device='cuda:2'), covar=tensor([0.1635, 0.1567, 0.1675, 0.1478, 0.1052, 0.1659, 0.2360, 0.2086], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0319, 0.0354, 0.0295, 0.0332, 0.0316, 0.0306, 0.0365], + device='cuda:2'), out_proj_covar=tensor([6.4509e-05, 6.6706e-05, 7.5821e-05, 6.0239e-05, 6.9169e-05, 6.6725e-05, + 6.4800e-05, 7.7964e-05], device='cuda:2') +2023-04-27 07:22:12,113 INFO [finetune.py:976] (2/7) Epoch 15, batch 4000, loss[loss=0.2083, simple_loss=0.2711, pruned_loss=0.07272, over 4903.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2462, pruned_loss=0.05495, over 954622.60 frames. ], batch size: 43, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:22:31,732 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84199.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:22:41,824 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3087, 2.8109, 0.9530, 1.5508, 2.2142, 1.4463, 4.0679, 1.7725], + device='cuda:2'), covar=tensor([0.0650, 0.0755, 0.0774, 0.1256, 0.0478, 0.0992, 0.0185, 0.0647], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 07:22:46,252 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-04-27 07:23:00,609 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 4050, loss[loss=0.1876, simple_loss=0.266, pruned_loss=0.05464, over 4901.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.25, pruned_loss=0.05613, over 954116.92 frames. ], batch size: 43, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:23:08,754 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1397, 1.4430, 1.3824, 1.5821, 1.5460, 1.6684, 1.3591, 3.0176], + device='cuda:2'), covar=tensor([0.0653, 0.0869, 0.0821, 0.1287, 0.0671, 0.0545, 0.0771, 0.0175], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 07:23:09,758 INFO [optim.py:369] (2/7) 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 07:23:18,167 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:23:35,685 INFO [finetune.py:976] (2/7) Epoch 15, batch 4100, loss[loss=0.1993, simple_loss=0.2639, pruned_loss=0.06738, over 4836.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2507, pruned_loss=0.05615, over 953226.51 frames. ], batch size: 30, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:23:42,180 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84297.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:23:48,049 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84305.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:24:08,375 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:24:09,500 INFO [finetune.py:976] (2/7) Epoch 15, batch 4150, loss[loss=0.1764, simple_loss=0.2464, pruned_loss=0.05319, over 4906.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2518, pruned_loss=0.0565, over 954077.61 frames. ], batch size: 46, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:24:09,582 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3370, 3.3565, 2.5788, 3.9096, 3.4333, 3.3506, 1.4127, 3.2771], + device='cuda:2'), covar=tensor([0.2017, 0.1278, 0.2849, 0.1949, 0.2724, 0.1868, 0.5924, 0.2463], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0212, 0.0248, 0.0301, 0.0296, 0.0246, 0.0268, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 07:24:11,380 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 07:24:43,282 INFO [finetune.py:976] (2/7) Epoch 15, batch 4200, loss[loss=0.1788, simple_loss=0.2459, pruned_loss=0.05586, over 4900.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.252, pruned_loss=0.05654, over 954202.12 frames. ], batch size: 37, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:24:49,269 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84397.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:24:52,527 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 07:25:01,618 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-27 07:25:03,287 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84416.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:25:22,955 INFO [finetune.py:976] (2/7) Epoch 15, batch 4250, loss[loss=0.1574, simple_loss=0.2396, pruned_loss=0.03764, over 4787.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2514, pruned_loss=0.05677, over 954930.00 frames. ], batch size: 29, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:25:33,111 INFO [optim.py:369] (2/7) 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 07:25:42,001 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84464.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:26:06,975 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8983, 1.5004, 1.4121, 1.6067, 2.0326, 1.7181, 1.4383, 1.2762], + device='cuda:2'), covar=tensor([0.1470, 0.1335, 0.1755, 0.1398, 0.0808, 0.1508, 0.1962, 0.2228], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0318, 0.0354, 0.0293, 0.0331, 0.0314, 0.0304, 0.0365], + device='cuda:2'), out_proj_covar=tensor([6.4008e-05, 6.6531e-05, 7.5732e-05, 5.9705e-05, 6.9027e-05, 6.6405e-05, + 6.4344e-05, 7.7955e-05], device='cuda:2') +2023-04-27 07:26:16,820 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1414, 2.8798, 2.2537, 2.6038, 1.9861, 2.2987, 2.5283, 1.7902], + device='cuda:2'), covar=tensor([0.2105, 0.1108, 0.0850, 0.1285, 0.2962, 0.1235, 0.1973, 0.2827], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0309, 0.0220, 0.0282, 0.0312, 0.0265, 0.0252, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1589e-04, 1.2301e-04, 8.7657e-05, 1.1214e-04, 1.2698e-04, 1.0564e-04, + 1.0187e-04, 1.0635e-04], device='cuda:2') +2023-04-27 07:26:28,013 INFO [finetune.py:976] (2/7) Epoch 15, batch 4300, loss[loss=0.162, simple_loss=0.2245, pruned_loss=0.04975, over 4823.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2481, pruned_loss=0.05527, over 956348.45 frames. ], batch size: 30, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:26:51,898 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 07:27:19,031 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84525.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:27:32,534 INFO [finetune.py:976] (2/7) Epoch 15, batch 4350, loss[loss=0.1598, simple_loss=0.2219, pruned_loss=0.04885, over 4734.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2443, pruned_loss=0.05413, over 956630.94 frames. ], batch size: 23, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:27:44,743 INFO [optim.py:369] (2/7) 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 07:27:54,629 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:26,832 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84586.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:27,949 INFO [finetune.py:976] (2/7) Epoch 15, batch 4400, loss[loss=0.1606, simple_loss=0.2258, pruned_loss=0.04765, over 4758.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2447, pruned_loss=0.05411, over 956755.90 frames. ], batch size: 26, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:28:30,487 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84592.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:39,051 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84605.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:28:57,942 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3908, 3.3872, 1.0006, 1.8309, 1.9335, 2.4356, 1.9449, 1.0075], + device='cuda:2'), covar=tensor([0.1433, 0.0909, 0.1945, 0.1248, 0.1017, 0.0986, 0.1441, 0.2107], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0243, 0.0137, 0.0120, 0.0130, 0.0151, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:28:59,169 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84633.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:02,148 INFO [finetune.py:976] (2/7) Epoch 15, batch 4450, loss[loss=0.1948, simple_loss=0.2722, pruned_loss=0.05868, over 4827.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2491, pruned_loss=0.05533, over 956499.86 frames. ], batch size: 40, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:29:04,125 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 07:29:11,995 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84653.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:22,431 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-27 07:29:36,301 INFO [finetune.py:976] (2/7) Epoch 15, batch 4500, loss[loss=0.1962, simple_loss=0.2621, pruned_loss=0.06514, over 4910.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2512, pruned_loss=0.05606, over 955626.96 frames. ], batch size: 36, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:29:37,003 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84689.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:38,816 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:29:40,105 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84694.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:30:09,976 INFO [finetune.py:976] (2/7) Epoch 15, batch 4550, loss[loss=0.1936, simple_loss=0.2724, pruned_loss=0.0574, over 4900.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2531, pruned_loss=0.05711, over 954788.96 frames. ], batch size: 37, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:30:16,091 INFO [optim.py:369] (2/7) 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 07:30:19,280 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84753.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:30:37,989 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.3108, 4.2810, 3.0739, 4.9222, 4.3109, 4.2708, 1.6622, 4.3120], + device='cuda:2'), covar=tensor([0.1481, 0.0863, 0.3257, 0.0851, 0.2799, 0.1503, 0.5729, 0.2032], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0214, 0.0251, 0.0303, 0.0298, 0.0247, 0.0270, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 07:30:43,750 INFO [finetune.py:976] (2/7) Epoch 15, batch 4600, loss[loss=0.1429, simple_loss=0.212, pruned_loss=0.03688, over 4806.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2527, pruned_loss=0.05692, over 955437.51 frames. ], batch size: 25, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:30:51,289 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84800.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:30:51,853 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84801.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:31:23,397 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-27 07:31:33,985 INFO [finetune.py:976] (2/7) Epoch 15, batch 4650, loss[loss=0.1905, simple_loss=0.2661, pruned_loss=0.05741, over 4810.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2501, pruned_loss=0.05624, over 955678.68 frames. ], batch size: 39, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:31:35,434 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-04-27 07:31:45,944 INFO [optim.py:369] (2/7) 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 07:31:46,820 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-04-27 07:31:55,731 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84855.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:32:05,695 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84861.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:32:29,510 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1227, 0.6624, 0.9727, 0.8151, 1.2152, 1.0310, 0.8417, 1.0148], + device='cuda:2'), covar=tensor([0.1665, 0.1487, 0.2235, 0.1571, 0.1063, 0.1363, 0.1613, 0.2194], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0318, 0.0355, 0.0294, 0.0333, 0.0313, 0.0304, 0.0366], + device='cuda:2'), out_proj_covar=tensor([6.4193e-05, 6.6392e-05, 7.6036e-05, 6.0037e-05, 6.9397e-05, 6.6213e-05, + 6.4260e-05, 7.8105e-05], device='cuda:2') +2023-04-27 07:32:30,055 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84881.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:32:39,700 INFO [finetune.py:976] (2/7) Epoch 15, batch 4700, loss[loss=0.1266, simple_loss=0.1949, pruned_loss=0.02917, over 4802.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2475, pruned_loss=0.05536, over 954699.40 frames. ], batch size: 25, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:32:42,705 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84892.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:33:00,659 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84903.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:33:03,167 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5824, 1.4810, 1.8102, 1.8634, 1.3911, 1.2494, 1.5909, 1.1334], + device='cuda:2'), covar=tensor([0.0618, 0.0651, 0.0437, 0.0569, 0.0767, 0.1164, 0.0612, 0.0634], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:33:46,042 INFO [finetune.py:976] (2/7) Epoch 15, batch 4750, loss[loss=0.1561, simple_loss=0.2282, pruned_loss=0.04203, over 4821.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2457, pruned_loss=0.05461, over 954634.56 frames. ], batch size: 51, lr: 3.48e-03, grad_scale: 32.0 +2023-04-27 07:33:47,830 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 07:33:55,660 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84951.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:34:19,536 INFO [finetune.py:976] (2/7) Epoch 15, batch 4800, loss[loss=0.2153, simple_loss=0.2853, pruned_loss=0.07264, over 4915.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2482, pruned_loss=0.05578, over 951936.61 frames. ], batch size: 36, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:34:20,211 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:34:23,030 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84992.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:34:35,169 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6077, 2.2785, 1.4935, 1.6553, 1.2488, 1.2220, 1.5435, 1.1753], + device='cuda:2'), covar=tensor([0.1741, 0.1305, 0.1600, 0.1707, 0.2444, 0.2100, 0.1021, 0.2003], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0213, 0.0169, 0.0204, 0.0201, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 07:34:36,373 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 4850, loss[loss=0.1796, simple_loss=0.2526, pruned_loss=0.05335, over 4848.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2526, pruned_loss=0.05696, over 952298.14 frames. ], batch size: 49, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:34:54,926 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 07:35:27,218 INFO [finetune.py:976] (2/7) Epoch 15, batch 4900, loss[loss=0.2077, simple_loss=0.278, pruned_loss=0.0687, over 4896.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2532, pruned_loss=0.057, over 951094.22 frames. ], batch size: 36, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:35:29,715 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7225, 3.9803, 0.5790, 2.0806, 2.2483, 2.4946, 2.2563, 0.8723], + device='cuda:2'), covar=tensor([0.1332, 0.0734, 0.2185, 0.1198, 0.0975, 0.1106, 0.1409, 0.2119], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0241, 0.0136, 0.0120, 0.0130, 0.0150, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:35:40,338 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85107.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:35:45,214 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 4950, loss[loss=0.1901, simple_loss=0.2637, pruned_loss=0.05822, over 4776.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2542, pruned_loss=0.05683, over 953381.31 frames. ], batch size: 51, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:36:07,085 INFO [optim.py:369] (2/7) 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 07:36:12,982 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85156.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:36:20,328 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85168.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:36:25,194 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:36:28,207 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85181.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:36:33,339 INFO [finetune.py:976] (2/7) Epoch 15, batch 5000, loss[loss=0.1743, simple_loss=0.2402, pruned_loss=0.05424, over 4725.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2522, pruned_loss=0.05651, over 954118.40 frames. ], batch size: 23, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:36:43,929 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-04-27 07:37:00,636 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:37:10,354 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2737, 2.0209, 2.2985, 2.5801, 2.5736, 2.2855, 1.7877, 2.5376], + device='cuda:2'), covar=tensor([0.0713, 0.0986, 0.0576, 0.0499, 0.0550, 0.0731, 0.0765, 0.0453], + device='cuda:2'), in_proj_covar=tensor([0.0193, 0.0206, 0.0185, 0.0176, 0.0181, 0.0186, 0.0157, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:37:12,535 INFO [finetune.py:976] (2/7) Epoch 15, batch 5050, loss[loss=0.1821, simple_loss=0.251, pruned_loss=0.05661, over 4788.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2493, pruned_loss=0.05575, over 955912.57 frames. ], batch size: 59, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:37:16,179 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6181, 1.2866, 4.4502, 4.2089, 3.8639, 4.2092, 4.0926, 3.9286], + device='cuda:2'), covar=tensor([0.6647, 0.6261, 0.1061, 0.1611, 0.1079, 0.1272, 0.1272, 0.1542], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0306, 0.0403, 0.0408, 0.0351, 0.0405, 0.0312, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:37:25,142 INFO [optim.py:369] (2/7) 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 07:37:44,721 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8769, 1.2863, 1.6012, 1.4813, 1.9353, 1.7132, 1.3516, 1.4864], + device='cuda:2'), covar=tensor([0.1609, 0.1405, 0.1658, 0.1626, 0.1078, 0.1247, 0.2006, 0.2022], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0317, 0.0354, 0.0291, 0.0331, 0.0312, 0.0302, 0.0365], + device='cuda:2'), out_proj_covar=tensor([6.3872e-05, 6.6359e-05, 7.5663e-05, 5.9370e-05, 6.8849e-05, 6.5953e-05, + 6.3821e-05, 7.7920e-05], device='cuda:2') +2023-04-27 07:37:56,142 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-27 07:38:17,996 INFO [finetune.py:976] (2/7) Epoch 15, batch 5100, loss[loss=0.1958, simple_loss=0.2602, pruned_loss=0.06568, over 4834.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2466, pruned_loss=0.05505, over 952071.46 frames. ], batch size: 33, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:38:18,663 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) 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] (2/7) Epoch 15, batch 5150, loss[loss=0.1812, simple_loss=0.2599, pruned_loss=0.05121, over 4920.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2475, pruned_loss=0.05548, over 951664.69 frames. ], batch size: 42, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:39:20,282 INFO [optim.py:369] (2/7) 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 07:39:24,505 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4919, 1.3588, 1.3465, 1.0377, 1.3377, 1.1499, 1.6860, 1.1053], + device='cuda:2'), covar=tensor([0.2715, 0.1353, 0.3926, 0.2005, 0.1294, 0.1718, 0.1289, 0.4197], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0349, 0.0426, 0.0355, 0.0384, 0.0380, 0.0372, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:39:25,694 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9490, 2.3478, 1.8981, 2.2486, 1.6214, 1.9105, 2.0008, 1.5845], + device='cuda:2'), covar=tensor([0.1778, 0.1254, 0.0860, 0.1086, 0.3026, 0.1240, 0.1837, 0.2465], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0308, 0.0219, 0.0280, 0.0311, 0.0263, 0.0249, 0.0266], + device='cuda:2'), out_proj_covar=tensor([1.1488e-04, 1.2258e-04, 8.7447e-05, 1.1151e-04, 1.2656e-04, 1.0507e-04, + 1.0092e-04, 1.0618e-04], device='cuda:2') +2023-04-27 07:39:28,589 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3251, 1.2243, 1.3343, 1.5861, 1.6317, 1.3776, 0.9997, 1.5153], + device='cuda:2'), covar=tensor([0.0776, 0.1207, 0.0820, 0.0557, 0.0603, 0.0661, 0.0835, 0.0533], + device='cuda:2'), in_proj_covar=tensor([0.0194, 0.0207, 0.0185, 0.0177, 0.0181, 0.0187, 0.0158, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:39:33,635 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6904, 1.4102, 4.3034, 4.0683, 3.7399, 3.9765, 3.9643, 3.8195], + device='cuda:2'), covar=tensor([0.6651, 0.5749, 0.1073, 0.1568, 0.1144, 0.1471, 0.1607, 0.1416], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0304, 0.0399, 0.0405, 0.0347, 0.0402, 0.0309, 0.0362], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:39:47,580 INFO [finetune.py:976] (2/7) Epoch 15, batch 5200, loss[loss=0.1853, simple_loss=0.2571, pruned_loss=0.05678, over 4938.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2508, pruned_loss=0.05653, over 952384.89 frames. ], batch size: 33, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:39:52,715 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 07:39:58,925 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-27 07:40:00,670 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7673, 2.1040, 0.9882, 1.4867, 2.1857, 1.5552, 1.5356, 1.6639], + device='cuda:2'), covar=tensor([0.0502, 0.0340, 0.0335, 0.0575, 0.0259, 0.0526, 0.0530, 0.0565], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0023, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 07:40:15,428 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85428.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:19,134 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2726, 2.7370, 2.2274, 2.2330, 1.7750, 1.7768, 2.2813, 1.7229], + device='cuda:2'), covar=tensor([0.1625, 0.1406, 0.1392, 0.1567, 0.2288, 0.1812, 0.0998, 0.1821], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0212, 0.0168, 0.0203, 0.0201, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 07:40:21,369 INFO [finetune.py:976] (2/7) Epoch 15, batch 5250, loss[loss=0.1399, simple_loss=0.2171, pruned_loss=0.03137, over 4763.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2515, pruned_loss=0.05624, over 951794.47 frames. ], batch size: 28, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:40:27,425 INFO [optim.py:369] (2/7) 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 07:40:33,311 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85456.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:38,412 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85463.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:40:44,659 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:48,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5455, 1.5606, 3.6403, 3.4188, 3.2549, 3.3490, 3.3527, 3.2479], + device='cuda:2'), covar=tensor([0.6633, 0.4733, 0.1304, 0.1734, 0.1162, 0.1797, 0.2833, 0.1518], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0304, 0.0400, 0.0405, 0.0347, 0.0403, 0.0310, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:40:51,978 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:40:54,961 INFO [finetune.py:976] (2/7) Epoch 15, batch 5300, loss[loss=0.1676, simple_loss=0.2455, pruned_loss=0.04485, over 4767.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2538, pruned_loss=0.05709, over 951683.78 frames. ], batch size: 28, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:40:55,671 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85489.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:01,378 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 07:41:04,898 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85504.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:23,643 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85530.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:28,368 INFO [finetune.py:976] (2/7) Epoch 15, batch 5350, loss[loss=0.1755, simple_loss=0.245, pruned_loss=0.053, over 4879.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2546, pruned_loss=0.05715, over 950955.71 frames. ], batch size: 32, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:41:32,155 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85544.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:41:32,755 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3737, 1.7384, 5.4373, 5.0997, 4.7369, 5.1841, 4.6644, 4.8181], + device='cuda:2'), covar=tensor([0.6029, 0.5525, 0.0860, 0.1454, 0.1165, 0.1478, 0.1067, 0.1395], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0303, 0.0400, 0.0406, 0.0347, 0.0403, 0.0311, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:41:34,434 INFO [optim.py:369] (2/7) 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 07:42:02,182 INFO [finetune.py:976] (2/7) Epoch 15, batch 5400, loss[loss=0.2142, simple_loss=0.2697, pruned_loss=0.07936, over 4864.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.25, pruned_loss=0.05586, over 950484.27 frames. ], batch size: 31, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:42:04,179 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85591.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:42:13,946 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85607.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:42:18,031 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6996, 1.1472, 1.8026, 2.2264, 1.7970, 1.6836, 1.7802, 1.7259], + device='cuda:2'), covar=tensor([0.4752, 0.6982, 0.6618, 0.6176, 0.6434, 0.8519, 0.8481, 0.8224], + device='cuda:2'), in_proj_covar=tensor([0.0417, 0.0407, 0.0493, 0.0505, 0.0445, 0.0468, 0.0474, 0.0477], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:42:27,865 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:42:30,878 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-04-27 07:42:34,432 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9988, 2.3430, 2.1979, 2.3647, 2.0830, 2.2908, 2.2915, 2.2515], + device='cuda:2'), covar=tensor([0.3906, 0.6227, 0.5195, 0.5024, 0.6439, 0.7545, 0.6141, 0.6089], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0371, 0.0317, 0.0331, 0.0341, 0.0396, 0.0352, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 07:42:41,671 INFO [finetune.py:976] (2/7) Epoch 15, batch 5450, loss[loss=0.1369, simple_loss=0.212, pruned_loss=0.03086, over 4825.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2473, pruned_loss=0.05477, over 950864.60 frames. ], batch size: 30, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:42:53,054 INFO [optim.py:369] (2/7) 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 07:43:02,878 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85655.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:43:45,366 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85686.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:43:46,488 INFO [finetune.py:976] (2/7) Epoch 15, batch 5500, loss[loss=0.1984, simple_loss=0.2489, pruned_loss=0.07396, over 4430.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2442, pruned_loss=0.05363, over 953386.98 frames. ], batch size: 19, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:44:00,099 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8758, 2.3811, 1.9483, 2.2031, 1.7428, 1.8885, 1.9917, 1.4240], + device='cuda:2'), covar=tensor([0.1802, 0.1179, 0.0869, 0.1054, 0.3031, 0.1284, 0.1761, 0.2539], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0308, 0.0219, 0.0280, 0.0310, 0.0263, 0.0250, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1459e-04, 1.2252e-04, 8.7466e-05, 1.1135e-04, 1.2618e-04, 1.0487e-04, + 1.0108e-04, 1.0634e-04], device='cuda:2') +2023-04-27 07:44:40,856 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4464, 2.8977, 0.9186, 1.7240, 2.2537, 1.5732, 3.8932, 2.0743], + device='cuda:2'), covar=tensor([0.0614, 0.0847, 0.0947, 0.1185, 0.0510, 0.0943, 0.0198, 0.0582], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 07:44:51,824 INFO [finetune.py:976] (2/7) Epoch 15, batch 5550, loss[loss=0.1836, simple_loss=0.2604, pruned_loss=0.0534, over 4835.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2482, pruned_loss=0.05532, over 953542.13 frames. ], batch size: 33, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:45:02,123 INFO [optim.py:369] (2/7) 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 07:45:11,450 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85771.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:24,800 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 15, batch 5600, loss[loss=0.151, simple_loss=0.2142, pruned_loss=0.0439, over 4032.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2501, pruned_loss=0.05476, over 954085.73 frames. ], batch size: 17, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:45:40,675 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85811.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:45,399 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:45:57,429 INFO [finetune.py:976] (2/7) Epoch 15, batch 5650, loss[loss=0.2285, simple_loss=0.2931, pruned_loss=0.08198, over 4192.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2527, pruned_loss=0.05577, over 951330.22 frames. ], batch size: 65, lr: 3.47e-03, grad_scale: 32.0 +2023-04-27 07:45:58,093 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85839.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:46:00,479 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2244, 2.9354, 2.1830, 2.5437, 1.8293, 2.3442, 2.5161, 1.8684], + device='cuda:2'), covar=tensor([0.2055, 0.1336, 0.0945, 0.1342, 0.3427, 0.1152, 0.2099, 0.2710], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0308, 0.0220, 0.0280, 0.0311, 0.0263, 0.0250, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1464e-04, 1.2251e-04, 8.7616e-05, 1.1132e-04, 1.2643e-04, 1.0481e-04, + 1.0125e-04, 1.0624e-04], device='cuda:2') +2023-04-27 07:46:03,465 INFO [optim.py:369] (2/7) 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 07:46:19,869 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85875.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:46:26,480 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85886.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:46:27,650 INFO [finetune.py:976] (2/7) Epoch 15, batch 5700, loss[loss=0.1253, simple_loss=0.1908, pruned_loss=0.02991, over 3991.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2486, pruned_loss=0.05464, over 938858.61 frames. ], batch size: 17, lr: 3.47e-03, grad_scale: 64.0 +2023-04-27 07:46:38,160 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-04-27 07:46:59,231 INFO [finetune.py:976] (2/7) Epoch 16, batch 0, loss[loss=0.1751, simple_loss=0.2534, pruned_loss=0.0484, over 4888.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2534, pruned_loss=0.0484, over 4888.00 frames. ], batch size: 35, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:46:59,231 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 07:47:06,143 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1170, 2.4823, 0.9868, 1.4098, 1.9281, 1.3240, 3.0104, 1.7040], + device='cuda:2'), covar=tensor([0.0578, 0.0533, 0.0740, 0.1213, 0.0431, 0.0908, 0.0236, 0.0583], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 07:47:15,737 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 07:47:26,823 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5138, 1.3818, 1.7417, 1.7524, 1.3400, 1.2640, 1.4460, 0.8954], + device='cuda:2'), covar=tensor([0.0534, 0.0755, 0.0453, 0.0707, 0.0800, 0.1174, 0.0677, 0.0733], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:47:30,033 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 07:47:30,499 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85930.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:47:32,919 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8436, 2.1281, 1.1478, 1.5037, 2.2104, 1.6923, 1.6381, 1.6698], + device='cuda:2'), covar=tensor([0.0497, 0.0349, 0.0299, 0.0553, 0.0244, 0.0515, 0.0502, 0.0557], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0037, 0.0050, 0.0037, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 07:47:34,159 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:47:41,355 INFO [optim.py:369] (2/7) 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 07:47:46,977 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 50, loss[loss=0.1921, simple_loss=0.254, pruned_loss=0.06513, over 4882.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2534, pruned_loss=0.0564, over 215914.51 frames. ], batch size: 35, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:48:04,120 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85981.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:48:10,260 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85991.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 07:48:23,359 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 100, loss[loss=0.1225, simple_loss=0.1886, pruned_loss=0.02823, over 4236.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2414, pruned_loss=0.05043, over 378533.37 frames. ], batch size: 18, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:48:45,854 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:49:27,105 INFO [optim.py:369] (2/7) 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 07:49:34,855 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86052.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:49:45,051 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4541, 1.6094, 1.5643, 1.8474, 1.7400, 2.0357, 1.4419, 3.7042], + device='cuda:2'), covar=tensor([0.0598, 0.0826, 0.0784, 0.1215, 0.0671, 0.0454, 0.0755, 0.0137], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 07:49:46,197 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2573, 2.2775, 1.9245, 1.9155, 2.3803, 1.8784, 2.9438, 1.7379], + device='cuda:2'), covar=tensor([0.3628, 0.1958, 0.4395, 0.3224, 0.1624, 0.2692, 0.1117, 0.4432], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0345, 0.0422, 0.0352, 0.0381, 0.0376, 0.0367, 0.0415], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:49:46,840 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0555, 1.7657, 2.1961, 2.4597, 2.1370, 1.9125, 2.0609, 1.9972], + device='cuda:2'), covar=tensor([0.5095, 0.7206, 0.7917, 0.5913, 0.6217, 0.9488, 0.9430, 1.0340], + device='cuda:2'), in_proj_covar=tensor([0.0418, 0.0408, 0.0496, 0.0506, 0.0446, 0.0469, 0.0477, 0.0481], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:49:47,422 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86063.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:49:49,103 INFO [finetune.py:976] (2/7) Epoch 16, batch 150, loss[loss=0.1455, simple_loss=0.214, pruned_loss=0.0385, over 4768.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2388, pruned_loss=0.05083, over 507087.90 frames. ], batch size: 26, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:49:52,049 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86069.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:50:02,657 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86113.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:50:45,307 INFO [finetune.py:976] (2/7) Epoch 16, batch 200, loss[loss=0.2003, simple_loss=0.2654, pruned_loss=0.0676, over 4825.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2382, pruned_loss=0.05141, over 605650.64 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:50:49,527 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2328, 1.7782, 2.2862, 2.4996, 2.2803, 1.7253, 1.4247, 2.0243], + device='cuda:2'), covar=tensor([0.3524, 0.3089, 0.1526, 0.2483, 0.2372, 0.2748, 0.4038, 0.2020], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0247, 0.0224, 0.0316, 0.0217, 0.0230, 0.0230, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 07:51:06,752 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86132.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:51:16,787 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86139.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:51:22,243 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 16, batch 250, loss[loss=0.1664, simple_loss=0.2486, pruned_loss=0.04212, over 4909.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2433, pruned_loss=0.0533, over 684187.07 frames. ], batch size: 37, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:51:48,384 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86187.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:06,459 INFO [finetune.py:976] (2/7) Epoch 16, batch 300, loss[loss=0.2085, simple_loss=0.2688, pruned_loss=0.07411, over 4909.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2475, pruned_loss=0.05433, over 742488.04 frames. ], batch size: 37, lr: 3.46e-03, grad_scale: 64.0 +2023-04-27 07:52:18,008 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86231.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:19,819 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86234.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:27,191 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86246.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:52:28,286 INFO [optim.py:369] (2/7) 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 07:52:30,245 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2693, 1.4403, 1.3179, 1.6865, 1.5612, 1.6493, 1.3136, 3.0831], + device='cuda:2'), covar=tensor([0.0723, 0.1104, 0.1070, 0.1364, 0.0852, 0.0601, 0.0983, 0.0298], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 07:52:39,294 INFO [finetune.py:976] (2/7) Epoch 16, batch 350, loss[loss=0.2196, simple_loss=0.2896, pruned_loss=0.07478, over 4792.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2494, pruned_loss=0.05451, over 790945.98 frames. ], batch size: 51, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:52:50,049 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86286.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 07:52:54,389 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6748, 1.5336, 1.8808, 2.0253, 1.4821, 1.2995, 1.6670, 1.0711], + device='cuda:2'), covar=tensor([0.0507, 0.0766, 0.0464, 0.0676, 0.0801, 0.1255, 0.0648, 0.0711], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0075, 0.0097, 0.0075, 0.0068], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 07:53:07,232 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86307.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:53:10,810 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 400, loss[loss=0.1805, simple_loss=0.2504, pruned_loss=0.05529, over 4854.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2511, pruned_loss=0.05505, over 828626.70 frames. ], batch size: 49, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:53:21,091 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:53:35,566 INFO [optim.py:369] (2/7) 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 07:53:41,145 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86358.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:53:45,898 INFO [finetune.py:976] (2/7) Epoch 16, batch 450, loss[loss=0.1754, simple_loss=0.2513, pruned_loss=0.0497, over 4904.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2496, pruned_loss=0.05455, over 857038.61 frames. ], batch size: 37, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:53:52,346 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1101, 1.5712, 1.9530, 2.4917, 2.1298, 1.5922, 1.6299, 1.8832], + device='cuda:2'), covar=tensor([0.3005, 0.3467, 0.1772, 0.2360, 0.2575, 0.2492, 0.4231, 0.2143], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0246, 0.0223, 0.0315, 0.0217, 0.0229, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 07:54:08,697 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7030, 2.2540, 1.5362, 1.7008, 1.3080, 1.3153, 1.6775, 1.2110], + device='cuda:2'), covar=tensor([0.1826, 0.1394, 0.1741, 0.1734, 0.2481, 0.2311, 0.1082, 0.2154], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0214, 0.0170, 0.0206, 0.0202, 0.0185, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 07:54:14,884 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 500, loss[loss=0.2014, simple_loss=0.2559, pruned_loss=0.07349, over 4308.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2483, pruned_loss=0.05506, over 877906.11 frames. ], batch size: 19, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:54:25,198 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:54:34,599 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2824, 1.7772, 2.1885, 2.6854, 2.1790, 1.6859, 1.4936, 2.0781], + device='cuda:2'), covar=tensor([0.3309, 0.3338, 0.1622, 0.2209, 0.2576, 0.2645, 0.4259, 0.2063], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0245, 0.0223, 0.0315, 0.0216, 0.0229, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 07:54:42,230 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86440.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:54:45,735 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86445.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:54:52,712 INFO [optim.py:369] (2/7) 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 07:54:57,858 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 07:55:08,755 INFO [finetune.py:976] (2/7) Epoch 16, batch 550, loss[loss=0.1442, simple_loss=0.2167, pruned_loss=0.0359, over 4936.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2458, pruned_loss=0.05475, over 897157.27 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:55:27,511 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 07:55:29,720 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 07:56:00,654 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86501.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:56:09,080 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 07:56:09,556 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86506.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:56:20,908 INFO [finetune.py:976] (2/7) Epoch 16, batch 600, loss[loss=0.2122, simple_loss=0.2676, pruned_loss=0.07841, over 4871.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.246, pruned_loss=0.05426, over 910882.75 frames. ], batch size: 31, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:56:32,808 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7511, 1.4072, 1.3072, 1.5681, 1.9615, 1.6156, 1.3648, 1.2319], + device='cuda:2'), covar=tensor([0.1623, 0.1399, 0.1803, 0.1307, 0.0937, 0.1505, 0.2288, 0.2326], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0312, 0.0350, 0.0289, 0.0329, 0.0311, 0.0299, 0.0360], + device='cuda:2'), out_proj_covar=tensor([6.3267e-05, 6.5197e-05, 7.4795e-05, 5.9126e-05, 6.8614e-05, 6.5673e-05, + 6.3203e-05, 7.6881e-05], device='cuda:2') +2023-04-27 07:56:35,207 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86531.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:56:45,556 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9554, 2.3405, 1.8521, 2.1542, 1.4852, 2.0652, 2.1138, 1.5237], + device='cuda:2'), covar=tensor([0.1935, 0.1593, 0.1226, 0.1404, 0.3473, 0.1348, 0.1699, 0.2520], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0311, 0.0223, 0.0283, 0.0316, 0.0266, 0.0253, 0.0269], + device='cuda:2'), out_proj_covar=tensor([1.1672e-04, 1.2389e-04, 8.8724e-05, 1.1252e-04, 1.2845e-04, 1.0597e-04, + 1.0236e-04, 1.0699e-04], device='cuda:2') +2023-04-27 07:56:54,677 INFO [optim.py:369] (2/7) 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 07:57:05,048 INFO [finetune.py:976] (2/7) Epoch 16, batch 650, loss[loss=0.1452, simple_loss=0.2124, pruned_loss=0.03899, over 4757.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2502, pruned_loss=0.05531, over 919954.84 frames. ], batch size: 28, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:57:12,976 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86579.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:13,626 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86580.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:18,205 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86586.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 07:57:20,572 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0157, 2.5510, 2.0550, 2.3626, 1.7073, 2.3081, 2.0829, 1.5120], + device='cuda:2'), covar=tensor([0.1990, 0.0943, 0.0793, 0.1222, 0.3261, 0.0930, 0.1889, 0.2731], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0310, 0.0222, 0.0282, 0.0315, 0.0265, 0.0252, 0.0268], + device='cuda:2'), out_proj_covar=tensor([1.1635e-04, 1.2343e-04, 8.8465e-05, 1.1218e-04, 1.2824e-04, 1.0553e-04, + 1.0205e-04, 1.0663e-04], device='cuda:2') +2023-04-27 07:57:29,903 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86602.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:36,593 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:38,332 INFO [finetune.py:976] (2/7) Epoch 16, batch 700, loss[loss=0.19, simple_loss=0.2623, pruned_loss=0.05889, over 4829.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2516, pruned_loss=0.05541, over 929026.98 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:57:49,357 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86634.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:57:54,604 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86641.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:57:55,229 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4446, 1.7058, 1.8739, 1.9578, 1.8372, 1.8823, 1.9813, 1.9451], + device='cuda:2'), covar=tensor([0.4334, 0.6365, 0.4788, 0.4740, 0.5567, 0.7662, 0.5430, 0.5093], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0371, 0.0318, 0.0332, 0.0344, 0.0397, 0.0352, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 07:58:00,292 INFO [optim.py:369] (2/7) 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 07:58:06,397 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86658.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:08,178 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 750, loss[loss=0.205, simple_loss=0.2621, pruned_loss=0.07394, over 4753.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2532, pruned_loss=0.05599, over 935404.79 frames. ], batch size: 26, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:58:38,827 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86706.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:40,071 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86708.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:58:43,809 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 07:58:44,305 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9551, 1.9251, 1.6734, 1.5323, 2.0223, 1.6106, 2.3942, 1.4461], + device='cuda:2'), covar=tensor([0.3777, 0.1718, 0.4456, 0.2853, 0.1598, 0.2437, 0.1376, 0.4616], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0349, 0.0428, 0.0359, 0.0385, 0.0383, 0.0373, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:58:44,816 INFO [finetune.py:976] (2/7) Epoch 16, batch 800, loss[loss=0.1778, simple_loss=0.2525, pruned_loss=0.05156, over 4742.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2531, pruned_loss=0.05604, over 941218.28 frames. ], batch size: 54, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:58:50,325 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86725.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:01,797 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 07:59:05,803 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86756.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:14,116 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5513, 0.7119, 1.3704, 1.8197, 1.5923, 1.4275, 1.4170, 1.4616], + device='cuda:2'), covar=tensor([0.4414, 0.5946, 0.5562, 0.5730, 0.5931, 0.7075, 0.6902, 0.6501], + device='cuda:2'), in_proj_covar=tensor([0.0418, 0.0409, 0.0496, 0.0505, 0.0448, 0.0470, 0.0477, 0.0480], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 07:59:17,588 INFO [finetune.py:976] (2/7) Epoch 16, batch 850, loss[loss=0.1705, simple_loss=0.244, pruned_loss=0.0485, over 4839.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2501, pruned_loss=0.05489, over 943623.73 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 07:59:21,923 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86773.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 07:59:36,380 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86801.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 07:59:49,821 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 900, loss[loss=0.1162, simple_loss=0.1926, pruned_loss=0.01995, over 4837.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2457, pruned_loss=0.05266, over 946305.59 frames. ], batch size: 30, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:00:22,180 INFO [optim.py:369] (2/7) 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 08:00:52,313 INFO [finetune.py:976] (2/7) Epoch 16, batch 950, loss[loss=0.1545, simple_loss=0.2304, pruned_loss=0.03927, over 4798.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2449, pruned_loss=0.05237, over 949934.84 frames. ], batch size: 51, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:01:03,393 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86875.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:01:05,177 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86878.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:01:37,240 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86902.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:01:58,441 INFO [finetune.py:976] (2/7) Epoch 16, batch 1000, loss[loss=0.215, simple_loss=0.2596, pruned_loss=0.08515, over 4249.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.248, pruned_loss=0.0542, over 952664.07 frames. ], batch size: 18, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:02:18,018 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:02:19,922 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86939.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:02:29,745 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6102, 1.3476, 1.7887, 1.8055, 1.3988, 1.2784, 1.4785, 0.9317], + device='cuda:2'), covar=tensor([0.0623, 0.0818, 0.0439, 0.0644, 0.0830, 0.1392, 0.0737, 0.0719], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0075, 0.0097, 0.0075, 0.0068], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:02:31,464 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:02:59,857 INFO [finetune.py:976] (2/7) Epoch 16, batch 1050, loss[loss=0.1886, simple_loss=0.268, pruned_loss=0.05461, over 4812.00 frames. ], tot_loss[loss=0.18, simple_loss=0.251, pruned_loss=0.05449, over 954556.40 frames. ], batch size: 38, lr: 3.46e-03, grad_scale: 32.0 +2023-04-27 08:03:28,670 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87002.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:03:41,409 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 08:03:43,716 INFO [finetune.py:976] (2/7) Epoch 16, batch 1100, loss[loss=0.1521, simple_loss=0.2371, pruned_loss=0.03357, over 4917.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2535, pruned_loss=0.05574, over 955666.89 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:04:06,134 INFO [optim.py:369] (2/7) 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 08:04:08,080 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6927, 2.1699, 1.6604, 1.4482, 1.2753, 1.2672, 1.7790, 1.2787], + device='cuda:2'), covar=tensor([0.1772, 0.1274, 0.1532, 0.1839, 0.2410, 0.2010, 0.1038, 0.2090], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0211, 0.0167, 0.0203, 0.0199, 0.0182, 0.0154, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 08:04:15,692 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87063.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:04:17,376 INFO [finetune.py:976] (2/7) Epoch 16, batch 1150, loss[loss=0.1443, simple_loss=0.2129, pruned_loss=0.03783, over 4776.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2543, pruned_loss=0.05639, over 956779.99 frames. ], batch size: 29, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:04:37,605 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87096.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:04:40,658 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87101.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:04:50,577 INFO [finetune.py:976] (2/7) Epoch 16, batch 1200, loss[loss=0.1621, simple_loss=0.2364, pruned_loss=0.04392, over 4766.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2521, pruned_loss=0.05571, over 954949.98 frames. ], batch size: 51, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:04:51,382 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 08:05:10,106 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87144.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:05:13,048 INFO [optim.py:369] (2/7) 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 08:05:13,121 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87149.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:05:24,513 INFO [finetune.py:976] (2/7) Epoch 16, batch 1250, loss[loss=0.1312, simple_loss=0.2072, pruned_loss=0.02763, over 4794.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2494, pruned_loss=0.05501, over 953234.70 frames. ], batch size: 29, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:05:27,052 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87170.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:05:39,947 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 08:05:52,780 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6511, 2.6250, 2.0389, 2.9966, 2.5889, 2.6150, 1.1528, 2.6038], + device='cuda:2'), covar=tensor([0.1926, 0.1309, 0.2828, 0.2336, 0.2878, 0.2007, 0.4988, 0.2707], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0212, 0.0250, 0.0302, 0.0297, 0.0247, 0.0269, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 08:05:58,111 INFO [finetune.py:976] (2/7) Epoch 16, batch 1300, loss[loss=0.1646, simple_loss=0.2328, pruned_loss=0.0482, over 4810.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2453, pruned_loss=0.05332, over 952352.41 frames. ], batch size: 51, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:06:09,207 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87230.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:12,072 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87234.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:13,313 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87236.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:15,890 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 08:06:21,119 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 16, batch 1350, loss[loss=0.147, simple_loss=0.2135, pruned_loss=0.04024, over 4824.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.245, pruned_loss=0.05346, over 951711.36 frames. ], batch size: 30, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:06:45,594 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87284.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:06:50,351 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87291.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:07:13,995 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-04-27 08:07:16,734 INFO [finetune.py:976] (2/7) Epoch 16, batch 1400, loss[loss=0.178, simple_loss=0.247, pruned_loss=0.0545, over 4889.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.25, pruned_loss=0.05533, over 952143.62 frames. ], batch size: 32, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:07:58,927 INFO [optim.py:369] (2/7) 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 08:08:15,346 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87358.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:08:17,779 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87362.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:08:20,155 INFO [finetune.py:976] (2/7) Epoch 16, batch 1450, loss[loss=0.2286, simple_loss=0.3031, pruned_loss=0.0771, over 4838.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2523, pruned_loss=0.05563, over 954314.16 frames. ], batch size: 47, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:08:25,359 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 08:09:12,953 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1499, 1.4660, 1.3922, 1.7591, 1.6347, 1.7783, 1.4132, 3.3817], + device='cuda:2'), covar=tensor([0.0678, 0.0827, 0.0812, 0.1229, 0.0654, 0.0541, 0.0779, 0.0156], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 08:09:14,071 INFO [finetune.py:976] (2/7) Epoch 16, batch 1500, loss[loss=0.2149, simple_loss=0.2831, pruned_loss=0.07338, over 4720.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2535, pruned_loss=0.05609, over 954661.31 frames. ], batch size: 59, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:09:17,333 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.12 vs. limit=5.0 +2023-04-27 08:09:18,464 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87423.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:09:36,462 INFO [optim.py:369] (2/7) 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 08:09:43,841 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 1550, loss[loss=0.1287, simple_loss=0.2081, pruned_loss=0.02469, over 4759.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2527, pruned_loss=0.05583, over 956156.06 frames. ], batch size: 28, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:09:47,520 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8692, 1.3933, 1.4428, 1.6662, 2.0916, 1.6564, 1.4168, 1.3643], + device='cuda:2'), covar=tensor([0.1402, 0.1511, 0.1716, 0.1103, 0.0812, 0.1874, 0.2126, 0.2122], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0311, 0.0351, 0.0289, 0.0328, 0.0311, 0.0300, 0.0360], + device='cuda:2'), 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:2') +2023-04-27 08:09:49,312 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87470.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:30,180 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 1600, loss[loss=0.1539, simple_loss=0.2251, pruned_loss=0.04139, over 4907.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2505, pruned_loss=0.05532, over 954267.11 frames. ], batch size: 46, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:10:43,626 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87518.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:46,105 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87522.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:54,390 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87534.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:10:54,429 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4914, 1.3765, 1.4724, 1.0231, 1.4942, 1.2671, 1.7838, 1.3094], + device='cuda:2'), covar=tensor([0.3148, 0.1763, 0.4300, 0.2405, 0.1368, 0.1930, 0.1562, 0.4359], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0348, 0.0426, 0.0357, 0.0384, 0.0381, 0.0372, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:10:56,834 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5030, 1.3045, 1.5617, 1.6437, 1.3965, 1.2638, 1.3155, 0.8846], + device='cuda:2'), covar=tensor([0.0516, 0.0776, 0.0583, 0.0497, 0.0655, 0.1213, 0.0615, 0.0777], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0070, 0.0068, 0.0076, 0.0097, 0.0076, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:11:05,886 INFO [optim.py:369] (2/7) 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 08:11:06,012 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9342, 2.3760, 2.0188, 2.2416, 1.6237, 2.0540, 2.0522, 1.5490], + device='cuda:2'), covar=tensor([0.1663, 0.0967, 0.0734, 0.1118, 0.2992, 0.1010, 0.1544, 0.2190], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0310, 0.0221, 0.0282, 0.0314, 0.0263, 0.0252, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1555e-04, 1.2331e-04, 8.8069e-05, 1.1210e-04, 1.2786e-04, 1.0504e-04, + 1.0188e-04, 1.0620e-04], device='cuda:2') +2023-04-27 08:11:16,205 INFO [finetune.py:976] (2/7) Epoch 16, batch 1650, loss[loss=0.1556, simple_loss=0.2329, pruned_loss=0.03917, over 4751.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2487, pruned_loss=0.05518, over 955981.49 frames. ], batch size: 27, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:11:16,333 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87566.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:11:17,547 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9941, 2.6138, 2.1300, 2.4241, 1.8378, 2.2294, 2.1365, 1.5632], + device='cuda:2'), covar=tensor([0.2050, 0.1114, 0.0823, 0.1160, 0.3288, 0.1172, 0.1870, 0.2797], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0311, 0.0222, 0.0283, 0.0315, 0.0264, 0.0253, 0.0268], + device='cuda:2'), out_proj_covar=tensor([1.1593e-04, 1.2376e-04, 8.8361e-05, 1.1240e-04, 1.2824e-04, 1.0529e-04, + 1.0222e-04, 1.0651e-04], device='cuda:2') +2023-04-27 08:11:26,484 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87582.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:28,251 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87584.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:29,407 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87586.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:29,484 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6034, 3.1346, 2.4180, 2.8119, 1.8578, 1.7357, 2.7240, 1.9094], + device='cuda:2'), covar=tensor([0.1451, 0.1370, 0.1376, 0.1373, 0.2215, 0.1815, 0.0901, 0.1840], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0213, 0.0168, 0.0205, 0.0200, 0.0184, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 08:11:31,296 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87605.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:11:49,708 INFO [finetune.py:976] (2/7) Epoch 16, batch 1700, loss[loss=0.1913, simple_loss=0.2669, pruned_loss=0.05787, over 4850.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.247, pruned_loss=0.05495, over 954408.26 frames. ], batch size: 44, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:12:08,452 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87645.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:12:11,746 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5723, 3.4386, 0.7096, 1.7925, 1.8635, 2.4963, 1.9006, 1.0552], + device='cuda:2'), covar=tensor([0.1354, 0.0841, 0.2225, 0.1279, 0.1069, 0.0962, 0.1517, 0.1850], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0137, 0.0121, 0.0131, 0.0154, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:12:12,242 INFO [optim.py:369] (2/7) 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 08:12:13,458 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:12:18,815 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87658.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:12:23,600 INFO [finetune.py:976] (2/7) Epoch 16, batch 1750, loss[loss=0.2056, simple_loss=0.286, pruned_loss=0.06266, over 4932.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2494, pruned_loss=0.05594, over 956228.39 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:12:23,727 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87666.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:12:32,270 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3207, 1.5590, 1.4100, 1.5208, 1.4106, 1.3315, 1.3731, 1.0472], + device='cuda:2'), covar=tensor([0.1735, 0.1483, 0.1013, 0.1252, 0.3406, 0.1341, 0.1957, 0.2603], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0310, 0.0221, 0.0282, 0.0313, 0.0263, 0.0252, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1523e-04, 1.2337e-04, 8.7991e-05, 1.1201e-04, 1.2762e-04, 1.0467e-04, + 1.0190e-04, 1.0606e-04], device='cuda:2') +2023-04-27 08:12:43,058 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 08:12:55,813 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87706.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:13:07,610 INFO [finetune.py:976] (2/7) Epoch 16, batch 1800, loss[loss=0.1772, simple_loss=0.2502, pruned_loss=0.05205, over 4737.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.251, pruned_loss=0.05578, over 956108.10 frames. ], batch size: 27, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:13:14,249 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:13:49,465 INFO [optim.py:369] (2/7) 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 08:14:12,969 INFO [finetune.py:976] (2/7) Epoch 16, batch 1850, loss[loss=0.1794, simple_loss=0.2537, pruned_loss=0.05259, over 4874.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2527, pruned_loss=0.0567, over 956080.13 frames. ], batch size: 35, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:15:07,316 INFO [finetune.py:976] (2/7) Epoch 16, batch 1900, loss[loss=0.1876, simple_loss=0.2599, pruned_loss=0.05764, over 4878.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2536, pruned_loss=0.05666, over 957293.05 frames. ], batch size: 32, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:15:08,021 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87817.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:15:20,774 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3673, 2.0113, 2.2417, 2.7089, 2.5585, 2.1141, 1.8055, 2.4147], + device='cuda:2'), covar=tensor([0.0816, 0.1001, 0.0588, 0.0500, 0.0578, 0.0823, 0.0758, 0.0506], + device='cuda:2'), in_proj_covar=tensor([0.0191, 0.0204, 0.0183, 0.0174, 0.0177, 0.0183, 0.0154, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:15:28,297 INFO [optim.py:369] (2/7) 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 08:15:37,124 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 1950, loss[loss=0.1455, simple_loss=0.2238, pruned_loss=0.03362, over 4254.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.251, pruned_loss=0.05538, over 955548.53 frames. ], batch size: 66, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:16:10,453 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87886.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:16:11,174 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-27 08:16:47,481 INFO [finetune.py:976] (2/7) Epoch 16, batch 2000, loss[loss=0.1771, simple_loss=0.2475, pruned_loss=0.05341, over 4810.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2486, pruned_loss=0.05454, over 955662.98 frames. ], batch size: 41, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:16:47,614 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6410, 1.5206, 1.9429, 2.0531, 1.5075, 1.3074, 1.6345, 1.0307], + device='cuda:2'), covar=tensor([0.0601, 0.0836, 0.0485, 0.0595, 0.0858, 0.1285, 0.0764, 0.0845], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0076, 0.0096, 0.0075, 0.0068], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:16:58,998 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87934.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:17:02,680 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87940.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 08:17:05,748 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:17:08,488 INFO [optim.py:369] (2/7) 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 08:17:17,317 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87961.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:17:21,295 INFO [finetune.py:976] (2/7) Epoch 16, batch 2050, loss[loss=0.1478, simple_loss=0.2174, pruned_loss=0.0391, over 4825.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2446, pruned_loss=0.05287, over 958021.44 frames. ], batch size: 41, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:17:55,289 INFO [finetune.py:976] (2/7) Epoch 16, batch 2100, loss[loss=0.1657, simple_loss=0.2382, pruned_loss=0.04663, over 4895.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2457, pruned_loss=0.05422, over 958280.14 frames. ], batch size: 32, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:17:57,085 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:18:16,282 INFO [optim.py:369] (2/7) 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 08:18:28,507 INFO [finetune.py:976] (2/7) Epoch 16, batch 2150, loss[loss=0.1735, simple_loss=0.2587, pruned_loss=0.04417, over 4913.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2499, pruned_loss=0.05543, over 957845.09 frames. ], batch size: 36, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:18:28,581 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88066.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:18:50,550 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88100.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:01,083 INFO [finetune.py:976] (2/7) Epoch 16, batch 2200, loss[loss=0.1821, simple_loss=0.2506, pruned_loss=0.05676, over 4797.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2526, pruned_loss=0.05622, over 957482.60 frames. ], batch size: 29, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:19:02,307 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88117.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:11,122 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6687, 1.8986, 1.9400, 2.5748, 2.7148, 2.1673, 2.1207, 1.9342], + device='cuda:2'), covar=tensor([0.1986, 0.1998, 0.2313, 0.2274, 0.1516, 0.2152, 0.2466, 0.2599], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0314, 0.0354, 0.0292, 0.0332, 0.0313, 0.0303, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.3787e-05, 6.5453e-05, 7.5584e-05, 5.9601e-05, 6.9119e-05, 6.6156e-05, + 6.3967e-05, 7.7718e-05], device='cuda:2') +2023-04-27 08:19:34,551 INFO [optim.py:369] (2/7) 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 08:19:47,615 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88161.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:19:47,632 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88161.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:50,024 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:19:56,115 INFO [finetune.py:976] (2/7) Epoch 16, batch 2250, loss[loss=0.1862, simple_loss=0.2711, pruned_loss=0.05065, over 4827.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2551, pruned_loss=0.05697, over 957678.45 frames. ], batch size: 49, lr: 3.45e-03, grad_scale: 32.0 +2023-04-27 08:20:21,115 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7349, 3.9625, 0.8600, 2.1208, 2.2382, 2.7192, 2.3017, 0.9965], + device='cuda:2'), covar=tensor([0.1372, 0.0972, 0.2083, 0.1226, 0.1031, 0.1025, 0.1531, 0.2187], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0246, 0.0138, 0.0121, 0.0132, 0.0154, 0.0119, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:20:30,758 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5165, 2.5197, 2.1052, 2.2387, 2.5815, 2.2192, 3.3275, 1.8432], + device='cuda:2'), covar=tensor([0.4016, 0.2619, 0.4872, 0.3449, 0.1890, 0.2701, 0.1572, 0.4592], + device='cuda:2'), in_proj_covar=tensor([0.0345, 0.0352, 0.0432, 0.0361, 0.0388, 0.0387, 0.0376, 0.0426], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:20:50,810 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88209.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:21:01,645 INFO [finetune.py:976] (2/7) Epoch 16, batch 2300, loss[loss=0.2164, simple_loss=0.2796, pruned_loss=0.07663, over 4252.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2537, pruned_loss=0.05622, over 958731.69 frames. ], batch size: 66, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:21:02,984 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3420, 1.9632, 2.1081, 2.5606, 2.5188, 2.0986, 1.6639, 2.3472], + device='cuda:2'), covar=tensor([0.0708, 0.0971, 0.0635, 0.0484, 0.0517, 0.0760, 0.0739, 0.0468], + device='cuda:2'), in_proj_covar=tensor([0.0191, 0.0204, 0.0185, 0.0175, 0.0178, 0.0184, 0.0155, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:21:07,153 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 08:21:18,879 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88240.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:21:21,926 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88245.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:21:24,249 INFO [optim.py:369] (2/7) 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 08:21:28,072 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 08:21:32,154 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88261.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:21:41,584 INFO [finetune.py:976] (2/7) Epoch 16, batch 2350, loss[loss=0.2004, simple_loss=0.2637, pruned_loss=0.06858, over 4865.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2526, pruned_loss=0.05627, over 958073.76 frames. ], batch size: 34, lr: 3.44e-03, grad_scale: 64.0 +2023-04-27 08:22:14,472 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88288.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:22:16,382 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88291.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:22:17,547 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:22:19,427 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0860, 1.6659, 1.8276, 2.2817, 2.2702, 1.8044, 1.4306, 2.1262], + device='cuda:2'), covar=tensor([0.0762, 0.1158, 0.0799, 0.0596, 0.0583, 0.0844, 0.0844, 0.0524], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0204, 0.0185, 0.0175, 0.0179, 0.0184, 0.0156, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:22:38,196 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88309.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:22:48,482 INFO [finetune.py:976] (2/7) Epoch 16, batch 2400, loss[loss=0.2032, simple_loss=0.2692, pruned_loss=0.06865, over 4911.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2491, pruned_loss=0.05519, over 958601.27 frames. ], batch size: 37, lr: 3.44e-03, grad_scale: 64.0 +2023-04-27 08:23:23,049 INFO [optim.py:369] (2/7) 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 08:23:25,037 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88352.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:23:33,497 INFO [finetune.py:976] (2/7) Epoch 16, batch 2450, loss[loss=0.1632, simple_loss=0.2296, pruned_loss=0.04839, over 4875.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2466, pruned_loss=0.05428, over 959193.39 frames. ], batch size: 31, lr: 3.44e-03, grad_scale: 64.0 +2023-04-27 08:24:30,564 INFO [finetune.py:976] (2/7) Epoch 16, batch 2500, loss[loss=0.1944, simple_loss=0.2588, pruned_loss=0.06501, over 4891.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.248, pruned_loss=0.05533, over 958905.79 frames. ], batch size: 32, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:24:31,919 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5177, 1.0627, 1.2744, 1.1029, 1.6348, 1.3115, 1.0844, 1.2059], + device='cuda:2'), covar=tensor([0.1490, 0.1456, 0.2138, 0.1363, 0.0848, 0.1568, 0.1854, 0.2224], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0313, 0.0352, 0.0291, 0.0331, 0.0314, 0.0303, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.3894e-05, 6.5388e-05, 7.5196e-05, 5.9429e-05, 6.8998e-05, 6.6344e-05, + 6.4056e-05, 7.7555e-05], device='cuda:2') +2023-04-27 08:24:54,797 INFO [optim.py:369] (2/7) 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 08:24:58,642 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 2550, loss[loss=0.1838, simple_loss=0.2616, pruned_loss=0.05303, over 4753.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2514, pruned_loss=0.05603, over 957910.84 frames. ], batch size: 54, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:25:38,575 INFO [finetune.py:976] (2/7) Epoch 16, batch 2600, loss[loss=0.1504, simple_loss=0.2186, pruned_loss=0.04106, over 4779.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2529, pruned_loss=0.05653, over 955615.07 frames. ], batch size: 25, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:26:16,941 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4135, 2.0421, 1.8491, 2.2763, 1.9854, 2.1053, 1.8054, 4.6829], + device='cuda:2'), covar=tensor([0.0581, 0.0781, 0.0779, 0.1134, 0.0680, 0.0563, 0.0729, 0.0108], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0039, 0.0057], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 08:26:18,050 INFO [optim.py:369] (2/7) 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 08:26:39,229 INFO [finetune.py:976] (2/7) Epoch 16, batch 2650, loss[loss=0.1722, simple_loss=0.2452, pruned_loss=0.0496, over 4904.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.253, pruned_loss=0.05636, over 956181.26 frames. ], batch size: 38, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:27:01,650 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8772, 1.0368, 1.4996, 1.6146, 1.5668, 1.6475, 1.5361, 1.4949], + device='cuda:2'), covar=tensor([0.3877, 0.4850, 0.4135, 0.3941, 0.5047, 0.6985, 0.4329, 0.4380], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0372, 0.0318, 0.0333, 0.0345, 0.0397, 0.0353, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 08:27:12,949 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5753, 1.2687, 4.2109, 3.9114, 3.6839, 3.9162, 3.7701, 3.6991], + device='cuda:2'), covar=tensor([0.7441, 0.5926, 0.1056, 0.1732, 0.1161, 0.1726, 0.2266, 0.1499], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0299, 0.0398, 0.0402, 0.0344, 0.0402, 0.0306, 0.0359], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:27:24,307 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4396, 1.3043, 1.6367, 1.6822, 1.3019, 1.2163, 1.4027, 0.9483], + device='cuda:2'), covar=tensor([0.0627, 0.0673, 0.0436, 0.0578, 0.0839, 0.1154, 0.0618, 0.0690], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0076, 0.0096, 0.0076, 0.0068], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:27:29,045 INFO [finetune.py:976] (2/7) Epoch 16, batch 2700, loss[loss=0.1426, simple_loss=0.2146, pruned_loss=0.03526, over 4840.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2512, pruned_loss=0.05508, over 956451.17 frames. ], batch size: 44, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:27:50,885 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2646, 1.5694, 1.7203, 1.8400, 1.6877, 1.8161, 1.8409, 1.7605], + device='cuda:2'), covar=tensor([0.4528, 0.5436, 0.4764, 0.4336, 0.5704, 0.7213, 0.4900, 0.4948], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0371, 0.0317, 0.0332, 0.0343, 0.0395, 0.0352, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 08:28:11,105 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88647.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:28:13,349 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 16, batch 2750, loss[loss=0.1951, simple_loss=0.2695, pruned_loss=0.06034, over 4934.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2483, pruned_loss=0.05424, over 956566.19 frames. ], batch size: 33, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:29:13,598 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-04-27 08:29:17,049 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2540, 1.4857, 1.3647, 1.7818, 1.6251, 1.7425, 1.4651, 3.0882], + device='cuda:2'), covar=tensor([0.0650, 0.0837, 0.0872, 0.1205, 0.0654, 0.0488, 0.0726, 0.0170], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 08:29:21,110 INFO [finetune.py:976] (2/7) Epoch 16, batch 2800, loss[loss=0.1531, simple_loss=0.2274, pruned_loss=0.0394, over 4817.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2454, pruned_loss=0.05319, over 956104.10 frames. ], batch size: 38, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:29:33,317 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7287, 2.3425, 2.5630, 3.1714, 2.9378, 2.6276, 2.0976, 2.6523], + device='cuda:2'), covar=tensor([0.0748, 0.0965, 0.0624, 0.0541, 0.0570, 0.0749, 0.0821, 0.0568], + device='cuda:2'), in_proj_covar=tensor([0.0192, 0.0205, 0.0185, 0.0176, 0.0179, 0.0185, 0.0156, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:29:42,735 INFO [optim.py:369] (2/7) 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 08:29:47,475 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88756.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:29:54,457 INFO [finetune.py:976] (2/7) Epoch 16, batch 2850, loss[loss=0.1662, simple_loss=0.2354, pruned_loss=0.04845, over 4772.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2439, pruned_loss=0.05247, over 955473.87 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:30:14,337 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8008, 1.0507, 1.6691, 2.2236, 1.8851, 1.7064, 1.7505, 1.7418], + device='cuda:2'), covar=tensor([0.4877, 0.6605, 0.6593, 0.6121, 0.5866, 0.8019, 0.7851, 0.8454], + device='cuda:2'), in_proj_covar=tensor([0.0420, 0.0411, 0.0497, 0.0511, 0.0450, 0.0476, 0.0482, 0.0483], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:30:20,170 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 2900, loss[loss=0.1298, simple_loss=0.2036, pruned_loss=0.02801, over 4778.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2466, pruned_loss=0.05314, over 954631.91 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:30:49,607 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9558, 2.3424, 2.0579, 2.2444, 1.5816, 2.0495, 2.1648, 1.5347], + device='cuda:2'), covar=tensor([0.1949, 0.1181, 0.0775, 0.1143, 0.3404, 0.1120, 0.1963, 0.2728], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0311, 0.0221, 0.0281, 0.0314, 0.0263, 0.0252, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1546e-04, 1.2392e-04, 8.8027e-05, 1.1185e-04, 1.2792e-04, 1.0474e-04, + 1.0195e-04, 1.0643e-04], device='cuda:2') +2023-04-27 08:30:50,707 INFO [optim.py:369] (2/7) 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 08:30:50,791 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0804, 3.1476, 2.4769, 3.6525, 3.0091, 3.1635, 1.5209, 3.0923], + device='cuda:2'), covar=tensor([0.1932, 0.1488, 0.4410, 0.2342, 0.2926, 0.1858, 0.5115, 0.2509], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0212, 0.0251, 0.0302, 0.0296, 0.0246, 0.0269, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 08:31:02,491 INFO [finetune.py:976] (2/7) Epoch 16, batch 2950, loss[loss=0.1968, simple_loss=0.2775, pruned_loss=0.05809, over 4787.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2494, pruned_loss=0.05368, over 955756.60 frames. ], batch size: 51, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:31:03,180 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8388, 3.9848, 2.7978, 4.5310, 3.9195, 3.8730, 1.6898, 3.9207], + device='cuda:2'), covar=tensor([0.1656, 0.1166, 0.3425, 0.1497, 0.3619, 0.1691, 0.5710, 0.2147], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0212, 0.0251, 0.0302, 0.0296, 0.0246, 0.0269, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 08:31:58,218 INFO [finetune.py:976] (2/7) Epoch 16, batch 3000, loss[loss=0.1717, simple_loss=0.2498, pruned_loss=0.04683, over 4870.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2515, pruned_loss=0.05479, over 954455.50 frames. ], batch size: 34, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:31:58,219 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 08:32:01,819 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1756, 1.6327, 2.0186, 2.2062, 2.0134, 1.6203, 1.2156, 1.6905], + device='cuda:2'), covar=tensor([0.3592, 0.3431, 0.1756, 0.2479, 0.2876, 0.2783, 0.4110, 0.2184], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0245, 0.0223, 0.0313, 0.0215, 0.0229, 0.0226, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 08:32:05,512 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1915, 1.6347, 2.0389, 2.2311, 1.9991, 1.6427, 1.1631, 1.6715], + device='cuda:2'), covar=tensor([0.3373, 0.3359, 0.1736, 0.2337, 0.2913, 0.2821, 0.4107, 0.2118], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0245, 0.0223, 0.0313, 0.0215, 0.0229, 0.0226, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 08:32:06,700 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6233, 1.2185, 1.4328, 1.3819, 1.7939, 1.4980, 1.2711, 1.3823], + device='cuda:2'), covar=tensor([0.1772, 0.1843, 0.2471, 0.1584, 0.1370, 0.1587, 0.2170, 0.2717], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0311, 0.0351, 0.0289, 0.0329, 0.0311, 0.0300, 0.0361], + device='cuda:2'), out_proj_covar=tensor([6.3310e-05, 6.4810e-05, 7.4858e-05, 5.8918e-05, 6.8460e-05, 6.5607e-05, + 6.3491e-05, 7.7042e-05], device='cuda:2') +2023-04-27 08:32:14,556 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 08:32:47,286 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88947.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:32:49,014 INFO [optim.py:369] (2/7) 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 08:32:59,792 INFO [finetune.py:976] (2/7) Epoch 16, batch 3050, loss[loss=0.2398, simple_loss=0.3082, pruned_loss=0.08574, over 4838.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2536, pruned_loss=0.0558, over 953045.32 frames. ], batch size: 49, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:33:30,456 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88995.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:33:55,682 INFO [finetune.py:976] (2/7) Epoch 16, batch 3100, loss[loss=0.1842, simple_loss=0.252, pruned_loss=0.05819, over 4897.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2521, pruned_loss=0.05522, over 953194.47 frames. ], batch size: 43, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:34:13,212 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:34:38,477 INFO [optim.py:369] (2/7) 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 08:34:59,018 INFO [finetune.py:976] (2/7) Epoch 16, batch 3150, loss[loss=0.1851, simple_loss=0.2489, pruned_loss=0.06067, over 4909.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2497, pruned_loss=0.05425, over 954533.18 frames. ], batch size: 32, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:35:14,717 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 3200, loss[loss=0.141, simple_loss=0.2093, pruned_loss=0.03636, over 4809.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2458, pruned_loss=0.05294, over 955512.63 frames. ], batch size: 25, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:36:23,962 INFO [optim.py:369] (2/7) 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 08:36:44,941 INFO [finetune.py:976] (2/7) Epoch 16, batch 3250, loss[loss=0.2145, simple_loss=0.2837, pruned_loss=0.07265, over 4863.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2462, pruned_loss=0.05391, over 953314.89 frames. ], batch size: 34, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:36:46,886 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89169.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:37:46,005 INFO [finetune.py:976] (2/7) Epoch 16, batch 3300, loss[loss=0.148, simple_loss=0.222, pruned_loss=0.03699, over 4752.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2486, pruned_loss=0.05429, over 954193.48 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:38:07,656 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89230.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:38:21,744 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 16, batch 3350, loss[loss=0.2244, simple_loss=0.2876, pruned_loss=0.08061, over 4812.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2524, pruned_loss=0.05605, over 954075.95 frames. ], batch size: 38, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:38:54,311 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.11 vs. limit=5.0 +2023-04-27 08:39:17,175 INFO [finetune.py:976] (2/7) Epoch 16, batch 3400, loss[loss=0.1681, simple_loss=0.2451, pruned_loss=0.04552, over 4922.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2535, pruned_loss=0.05678, over 953959.85 frames. ], batch size: 38, lr: 3.44e-03, grad_scale: 32.0 +2023-04-27 08:39:31,142 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-27 08:39:40,163 INFO [optim.py:369] (2/7) 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 08:39:53,102 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9418, 1.7693, 4.7290, 4.4109, 4.2274, 4.4245, 4.3740, 4.2840], + device='cuda:2'), covar=tensor([0.5847, 0.4752, 0.1005, 0.1789, 0.0930, 0.1374, 0.0846, 0.1168], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0300, 0.0397, 0.0399, 0.0342, 0.0402, 0.0305, 0.0357], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:39:53,662 INFO [finetune.py:976] (2/7) Epoch 16, batch 3450, loss[loss=0.1949, simple_loss=0.2532, pruned_loss=0.06829, over 4425.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2527, pruned_loss=0.05616, over 953612.07 frames. ], batch size: 19, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:40:04,283 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:40:16,189 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89383.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:40:54,851 INFO [finetune.py:976] (2/7) Epoch 16, batch 3500, loss[loss=0.1314, simple_loss=0.2046, pruned_loss=0.0291, over 4928.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2502, pruned_loss=0.05523, over 955340.81 frames. ], batch size: 33, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:41:00,467 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:41:07,512 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89435.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:41:09,966 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89439.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:41:18,077 INFO [optim.py:369] (2/7) 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 08:41:18,313 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.01 vs. limit=5.0 +2023-04-27 08:41:22,500 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6379, 1.4529, 1.8327, 1.9372, 1.4864, 1.2705, 1.6169, 1.0376], + device='cuda:2'), covar=tensor([0.0603, 0.0725, 0.0433, 0.0678, 0.0800, 0.1319, 0.0664, 0.0730], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0071, 0.0070, 0.0069, 0.0077, 0.0098, 0.0076, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:41:34,098 INFO [finetune.py:976] (2/7) Epoch 16, batch 3550, loss[loss=0.1515, simple_loss=0.2148, pruned_loss=0.04406, over 4822.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2473, pruned_loss=0.05412, over 956724.55 frames. ], batch size: 25, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:41:58,281 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89486.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:42:06,479 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5525, 1.9923, 2.4000, 3.0715, 2.3162, 1.8885, 1.9264, 2.2951], + device='cuda:2'), covar=tensor([0.3388, 0.3385, 0.1685, 0.2481, 0.3241, 0.2847, 0.3779, 0.2196], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0245, 0.0223, 0.0312, 0.0216, 0.0229, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 08:42:08,153 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89500.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:42:20,187 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89509.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:42:29,609 INFO [finetune.py:976] (2/7) Epoch 16, batch 3600, loss[loss=0.1643, simple_loss=0.2346, pruned_loss=0.04695, over 4826.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2449, pruned_loss=0.05341, over 956619.76 frames. ], batch size: 33, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:42:41,035 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89525.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:43:13,813 INFO [optim.py:369] (2/7) 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 08:43:36,162 INFO [finetune.py:976] (2/7) Epoch 16, batch 3650, loss[loss=0.1693, simple_loss=0.2477, pruned_loss=0.04543, over 4758.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2473, pruned_loss=0.05451, over 955372.62 frames. ], batch size: 28, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:43:38,726 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89570.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:43:48,397 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6438, 2.6233, 2.2627, 3.0486, 2.6432, 2.6124, 1.3984, 2.6622], + device='cuda:2'), covar=tensor([0.2322, 0.1803, 0.3921, 0.3305, 0.2801, 0.2508, 0.5173, 0.2888], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0214, 0.0254, 0.0306, 0.0300, 0.0250, 0.0274, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 08:43:48,450 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9338, 0.9015, 1.0824, 1.0476, 0.9275, 0.8083, 0.9492, 0.7003], + device='cuda:2'), covar=tensor([0.0531, 0.0481, 0.0483, 0.0493, 0.0638, 0.1010, 0.0409, 0.0672], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0071, 0.0070, 0.0068, 0.0077, 0.0098, 0.0076, 0.0069], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:44:42,831 INFO [finetune.py:976] (2/7) Epoch 16, batch 3700, loss[loss=0.1622, simple_loss=0.2319, pruned_loss=0.04632, over 4833.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2497, pruned_loss=0.0549, over 954985.15 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:45:26,134 INFO [optim.py:369] (2/7) 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 08:45:48,272 INFO [finetune.py:976] (2/7) Epoch 16, batch 3750, loss[loss=0.1762, simple_loss=0.2574, pruned_loss=0.04752, over 4789.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2496, pruned_loss=0.05512, over 953207.14 frames. ], batch size: 45, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:46:02,534 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:46:10,281 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89683.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:46:13,396 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6024, 2.7854, 2.3293, 2.4484, 2.8815, 2.4163, 3.7415, 2.0138], + device='cuda:2'), covar=tensor([0.3909, 0.2120, 0.4659, 0.3363, 0.1655, 0.2746, 0.1174, 0.4368], + device='cuda:2'), in_proj_covar=tensor([0.0347, 0.0350, 0.0432, 0.0358, 0.0386, 0.0386, 0.0374, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:46:30,732 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6163, 1.7787, 0.7902, 1.3285, 1.9019, 1.5264, 1.3548, 1.4214], + device='cuda:2'), covar=tensor([0.0491, 0.0338, 0.0363, 0.0528, 0.0273, 0.0504, 0.0488, 0.0556], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 08:46:49,772 INFO [finetune.py:976] (2/7) Epoch 16, batch 3800, loss[loss=0.1734, simple_loss=0.2511, pruned_loss=0.04787, over 4838.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2521, pruned_loss=0.05553, over 954305.89 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:47:09,442 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89730.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:47:10,032 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89731.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:47:21,764 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89741.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:47:31,967 INFO [optim.py:369] (2/7) 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 08:47:43,195 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7818, 1.3502, 1.9185, 2.3150, 1.8923, 1.7690, 1.8833, 1.8110], + device='cuda:2'), covar=tensor([0.4601, 0.6364, 0.5937, 0.5820, 0.5966, 0.7520, 0.7483, 0.8141], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0408, 0.0494, 0.0507, 0.0447, 0.0473, 0.0477, 0.0481], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:47:43,647 INFO [finetune.py:976] (2/7) Epoch 16, batch 3850, loss[loss=0.1567, simple_loss=0.2248, pruned_loss=0.04437, over 4767.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2507, pruned_loss=0.05487, over 954217.04 frames. ], batch size: 28, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:47:43,867 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 08:47:58,305 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89781.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:48:12,383 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89795.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:48:26,725 INFO [finetune.py:976] (2/7) Epoch 16, batch 3900, loss[loss=0.192, simple_loss=0.2575, pruned_loss=0.06327, over 4903.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.249, pruned_loss=0.05475, over 954406.70 frames. ], batch size: 35, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:48:33,337 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 08:48:59,089 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:48:59,146 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:48:59,621 INFO [finetune.py:976] (2/7) Epoch 16, batch 3950, loss[loss=0.149, simple_loss=0.2238, pruned_loss=0.03712, over 4696.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.246, pruned_loss=0.05341, over 951474.09 frames. ], batch size: 23, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:49:05,392 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89873.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:49:33,476 INFO [finetune.py:976] (2/7) Epoch 16, batch 4000, loss[loss=0.273, simple_loss=0.3184, pruned_loss=0.1138, over 4816.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2459, pruned_loss=0.05377, over 952691.88 frames. ], batch size: 51, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:49:47,309 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89926.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:50:11,835 INFO [optim.py:369] (2/7) 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 08:50:33,231 INFO [finetune.py:976] (2/7) Epoch 16, batch 4050, loss[loss=0.1783, simple_loss=0.2525, pruned_loss=0.05211, over 4865.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2501, pruned_loss=0.05573, over 953882.26 frames. ], batch size: 31, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:51:40,301 INFO [finetune.py:976] (2/7) Epoch 16, batch 4100, loss[loss=0.2079, simple_loss=0.2679, pruned_loss=0.07397, over 4170.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2513, pruned_loss=0.05543, over 952314.86 frames. ], batch size: 65, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:52:02,666 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90030.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:52:12,523 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:52:27,090 INFO [optim.py:369] (2/7) 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 08:52:48,659 INFO [finetune.py:976] (2/7) Epoch 16, batch 4150, loss[loss=0.1743, simple_loss=0.2362, pruned_loss=0.05621, over 4714.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2527, pruned_loss=0.05569, over 953883.42 frames. ], batch size: 23, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:52:58,200 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 08:53:06,918 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:53:09,295 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90081.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:53:13,132 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0186, 4.1267, 0.7978, 2.2281, 2.4399, 2.7366, 2.2580, 0.9554], + device='cuda:2'), covar=tensor([0.1239, 0.0999, 0.2242, 0.1181, 0.0928, 0.1041, 0.1431, 0.2246], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0247, 0.0139, 0.0122, 0.0133, 0.0154, 0.0119, 0.0122], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 08:53:26,044 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90095.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 08:53:50,593 INFO [finetune.py:976] (2/7) Epoch 16, batch 4200, loss[loss=0.1477, simple_loss=0.2164, pruned_loss=0.03949, over 4158.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2534, pruned_loss=0.05561, over 953089.83 frames. ], batch size: 65, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:53:59,892 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90129.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:54:00,557 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90130.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:54:14,873 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90143.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 08:54:24,333 INFO [optim.py:369] (2/7) 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 08:54:27,604 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-27 08:54:34,096 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:54:34,611 INFO [finetune.py:976] (2/7) Epoch 16, batch 4250, loss[loss=0.2377, simple_loss=0.2895, pruned_loss=0.09301, over 4738.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2511, pruned_loss=0.05544, over 951872.39 frames. ], batch size: 23, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:54:52,905 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90191.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:55:02,879 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 08:55:06,207 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90213.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:55:08,444 INFO [finetune.py:976] (2/7) Epoch 16, batch 4300, loss[loss=0.1845, simple_loss=0.2422, pruned_loss=0.06341, over 4870.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2487, pruned_loss=0.05518, over 954354.33 frames. ], batch size: 34, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:55:11,526 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90221.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 08:55:19,263 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:55:21,665 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6140, 0.8280, 1.4932, 1.9739, 1.7208, 1.5607, 1.5507, 1.5705], + device='cuda:2'), covar=tensor([0.4230, 0.5644, 0.5559, 0.5505, 0.5217, 0.6767, 0.6817, 0.6402], + device='cuda:2'), in_proj_covar=tensor([0.0418, 0.0405, 0.0493, 0.0505, 0.0447, 0.0471, 0.0475, 0.0479], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 08:55:32,136 INFO [optim.py:369] (2/7) 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 08:55:41,846 INFO [finetune.py:976] (2/7) Epoch 16, batch 4350, loss[loss=0.1883, simple_loss=0.2482, pruned_loss=0.06417, over 4851.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2458, pruned_loss=0.05418, over 953436.69 frames. ], batch size: 44, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:55:48,117 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 08:55:57,666 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1533, 1.5798, 1.3944, 1.7398, 1.7166, 1.8981, 1.4074, 3.6330], + device='cuda:2'), covar=tensor([0.0611, 0.0767, 0.0787, 0.1212, 0.0621, 0.0514, 0.0727, 0.0141], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 08:56:00,601 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90293.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:56:15,530 INFO [finetune.py:976] (2/7) Epoch 16, batch 4400, loss[loss=0.1612, simple_loss=0.2215, pruned_loss=0.05043, over 4055.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2466, pruned_loss=0.05441, over 954032.22 frames. ], batch size: 17, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:56:18,674 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0297, 2.6071, 2.1046, 2.0120, 1.4818, 1.5279, 2.1739, 1.4485], + device='cuda:2'), covar=tensor([0.1683, 0.1477, 0.1483, 0.1766, 0.2532, 0.2133, 0.1075, 0.2082], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0213, 0.0169, 0.0205, 0.0201, 0.0184, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 08:56:19,787 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-27 08:56:34,938 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:56:36,808 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5028, 1.7792, 1.9305, 2.0528, 1.8784, 2.0416, 2.0237, 2.0027], + device='cuda:2'), covar=tensor([0.4158, 0.6050, 0.4740, 0.4841, 0.5835, 0.7090, 0.5774, 0.5251], + device='cuda:2'), in_proj_covar=tensor([0.0332, 0.0372, 0.0316, 0.0333, 0.0344, 0.0394, 0.0353, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 08:56:55,338 INFO [optim.py:369] (2/7) 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 08:56:56,107 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 4450, loss[loss=0.1747, simple_loss=0.2408, pruned_loss=0.05429, over 4870.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2478, pruned_loss=0.05385, over 952421.66 frames. ], batch size: 31, lr: 3.43e-03, grad_scale: 32.0 +2023-04-27 08:57:33,651 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:58:04,791 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 4500, loss[loss=0.1582, simple_loss=0.2302, pruned_loss=0.04308, over 4126.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2507, pruned_loss=0.0552, over 952831.93 frames. ], batch size: 17, lr: 3.43e-03, grad_scale: 64.0 +2023-04-27 08:58:52,084 INFO [optim.py:369] (2/7) 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 08:59:04,442 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:59:14,179 INFO [finetune.py:976] (2/7) Epoch 16, batch 4550, loss[loss=0.1929, simple_loss=0.2578, pruned_loss=0.06399, over 4718.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2523, pruned_loss=0.05595, over 953447.88 frames. ], batch size: 54, lr: 3.43e-03, grad_scale: 64.0 +2023-04-27 08:59:37,651 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90486.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 08:59:55,978 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90496.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:00:15,160 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7961, 3.6757, 2.7175, 4.4163, 3.8324, 3.8056, 1.5993, 3.7234], + device='cuda:2'), covar=tensor([0.1779, 0.1201, 0.3013, 0.1658, 0.2989, 0.1865, 0.5819, 0.2455], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0212, 0.0251, 0.0303, 0.0297, 0.0248, 0.0271, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 09:00:15,727 INFO [finetune.py:976] (2/7) Epoch 16, batch 4600, loss[loss=0.1745, simple_loss=0.2492, pruned_loss=0.04993, over 4773.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2519, pruned_loss=0.05543, over 955546.70 frames. ], batch size: 28, lr: 3.43e-03, grad_scale: 64.0 +2023-04-27 09:00:18,318 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:00:18,918 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90521.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 09:00:22,695 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 09:00:37,737 INFO [optim.py:369] (2/7) 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 09:00:37,962 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 09:00:43,138 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90557.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:00:46,454 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6454, 1.5764, 0.9280, 1.3321, 1.6850, 1.5493, 1.4013, 1.4502], + device='cuda:2'), covar=tensor([0.0495, 0.0352, 0.0331, 0.0532, 0.0268, 0.0487, 0.0472, 0.0563], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 09:00:49,342 INFO [finetune.py:976] (2/7) Epoch 16, batch 4650, loss[loss=0.1745, simple_loss=0.2367, pruned_loss=0.05611, over 4929.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2491, pruned_loss=0.05478, over 956186.30 frames. ], batch size: 38, lr: 3.42e-03, grad_scale: 64.0 +2023-04-27 09:00:51,268 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90569.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:01:03,239 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90588.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:01:23,260 INFO [finetune.py:976] (2/7) Epoch 16, batch 4700, loss[loss=0.1594, simple_loss=0.2177, pruned_loss=0.05056, over 4818.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2471, pruned_loss=0.05417, over 956332.30 frames. ], batch size: 39, lr: 3.42e-03, grad_scale: 64.0 +2023-04-27 09:01:36,023 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5327, 3.3500, 0.9549, 1.6671, 1.9479, 2.2846, 1.9113, 1.0288], + device='cuda:2'), covar=tensor([0.1402, 0.0889, 0.1937, 0.1330, 0.1048, 0.1024, 0.1528, 0.1781], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0244, 0.0137, 0.0121, 0.0131, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:01:45,413 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 16, batch 4750, loss[loss=0.1962, simple_loss=0.2785, pruned_loss=0.05694, over 4821.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2456, pruned_loss=0.05338, over 957731.58 frames. ], batch size: 39, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:02:51,158 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:03:07,869 INFO [finetune.py:976] (2/7) Epoch 16, batch 4800, loss[loss=0.1551, simple_loss=0.2263, pruned_loss=0.04198, over 4744.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2479, pruned_loss=0.05416, over 957183.59 frames. ], batch size: 26, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:03:07,997 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3310, 2.0485, 2.2939, 2.7694, 2.6417, 2.3073, 1.9142, 2.4862], + device='cuda:2'), covar=tensor([0.0788, 0.1005, 0.0569, 0.0513, 0.0488, 0.0764, 0.0721, 0.0486], + device='cuda:2'), in_proj_covar=tensor([0.0186, 0.0198, 0.0179, 0.0170, 0.0175, 0.0179, 0.0150, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:03:19,958 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5193, 1.5852, 0.6657, 1.2543, 1.6756, 1.3965, 1.2860, 1.3904], + device='cuda:2'), covar=tensor([0.0529, 0.0380, 0.0380, 0.0556, 0.0288, 0.0528, 0.0507, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0029, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0029], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 09:03:36,199 INFO [optim.py:369] (2/7) 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 09:03:47,210 INFO [finetune.py:976] (2/7) Epoch 16, batch 4850, loss[loss=0.2043, simple_loss=0.2652, pruned_loss=0.07167, over 4850.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2516, pruned_loss=0.05545, over 957420.52 frames. ], batch size: 31, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:03:49,732 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6826, 1.0219, 1.6784, 2.1164, 1.7601, 1.6241, 1.7117, 1.6416], + device='cuda:2'), covar=tensor([0.4713, 0.6392, 0.6023, 0.5838, 0.5531, 0.7387, 0.7449, 0.8083], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0405, 0.0493, 0.0507, 0.0446, 0.0472, 0.0477, 0.0482], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:04:00,363 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90786.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:04:01,590 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5122, 3.2783, 2.7603, 3.0517, 2.4272, 2.7562, 2.8777, 2.1783], + device='cuda:2'), covar=tensor([0.2132, 0.1219, 0.0816, 0.1218, 0.3175, 0.1382, 0.1860, 0.3169], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0312, 0.0224, 0.0283, 0.0316, 0.0264, 0.0253, 0.0269], + device='cuda:2'), out_proj_covar=tensor([1.1619e-04, 1.2395e-04, 8.9320e-05, 1.1255e-04, 1.2878e-04, 1.0514e-04, + 1.0229e-04, 1.0709e-04], device='cuda:2') +2023-04-27 09:04:36,231 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 4900, loss[loss=0.2357, simple_loss=0.2877, pruned_loss=0.09183, over 4885.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2533, pruned_loss=0.05613, over 957360.46 frames. ], batch size: 32, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:05:00,533 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9498, 1.2156, 4.6161, 4.3054, 4.0135, 4.2811, 4.0301, 4.0914], + device='cuda:2'), covar=tensor([0.6412, 0.6002, 0.0999, 0.1815, 0.1205, 0.1261, 0.2465, 0.1392], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0304, 0.0403, 0.0404, 0.0347, 0.0409, 0.0309, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:05:01,135 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90834.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:05:21,461 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90849.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:05:21,696 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 09:05:22,604 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:05:32,384 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8463, 1.4188, 1.9468, 2.3646, 1.9286, 1.7851, 1.9171, 1.8871], + device='cuda:2'), covar=tensor([0.4824, 0.7090, 0.6690, 0.5989, 0.5953, 0.8032, 0.8063, 0.8976], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0405, 0.0493, 0.0507, 0.0447, 0.0471, 0.0477, 0.0482], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:05:38,747 INFO [finetune.py:976] (2/7) Epoch 16, batch 4950, loss[loss=0.2698, simple_loss=0.3268, pruned_loss=0.1064, over 4824.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2542, pruned_loss=0.05651, over 954928.48 frames. ], batch size: 39, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:05:54,237 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:06:03,062 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 09:06:08,377 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90910.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:06:12,378 INFO [finetune.py:976] (2/7) Epoch 16, batch 5000, loss[loss=0.1445, simple_loss=0.227, pruned_loss=0.03098, over 4842.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2514, pruned_loss=0.05549, over 951832.72 frames. ], batch size: 47, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:06:27,002 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90936.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:06:36,100 INFO [optim.py:369] (2/7) 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 09:06:46,205 INFO [finetune.py:976] (2/7) Epoch 16, batch 5050, loss[loss=0.1577, simple_loss=0.2215, pruned_loss=0.04692, over 4746.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2487, pruned_loss=0.05472, over 953311.78 frames. ], batch size: 26, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:07:13,892 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91007.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:07:19,801 INFO [finetune.py:976] (2/7) Epoch 16, batch 5100, loss[loss=0.1836, simple_loss=0.2403, pruned_loss=0.06348, over 4901.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2451, pruned_loss=0.05313, over 955102.94 frames. ], batch size: 36, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:07:22,201 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4447, 1.7226, 1.5844, 2.2661, 2.4505, 2.0299, 1.9488, 1.7381], + device='cuda:2'), covar=tensor([0.2146, 0.1905, 0.1751, 0.1578, 0.1210, 0.1971, 0.2312, 0.2158], + device='cuda:2'), in_proj_covar=tensor([0.0302, 0.0307, 0.0347, 0.0285, 0.0323, 0.0308, 0.0298, 0.0360], + device='cuda:2'), out_proj_covar=tensor([6.2427e-05, 6.4083e-05, 7.3996e-05, 5.8008e-05, 6.7258e-05, 6.4901e-05, + 6.3081e-05, 7.6835e-05], device='cuda:2') +2023-04-27 09:07:43,899 INFO [optim.py:369] (2/7) 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 09:07:46,425 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91055.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:07:51,973 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3541, 2.0366, 2.4059, 2.7810, 2.7269, 2.1966, 1.8805, 2.3844], + device='cuda:2'), covar=tensor([0.0900, 0.1041, 0.0606, 0.0572, 0.0589, 0.0946, 0.0853, 0.0632], + device='cuda:2'), in_proj_covar=tensor([0.0190, 0.0202, 0.0182, 0.0173, 0.0178, 0.0182, 0.0153, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:07:53,571 INFO [finetune.py:976] (2/7) Epoch 16, batch 5150, loss[loss=0.2016, simple_loss=0.2759, pruned_loss=0.06359, over 4762.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2457, pruned_loss=0.0534, over 955615.48 frames. ], batch size: 54, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:07:53,790 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-04-27 09:08:33,728 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 09:08:48,474 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 5200, loss[loss=0.204, simple_loss=0.2708, pruned_loss=0.06856, over 4916.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2483, pruned_loss=0.05438, over 955937.13 frames. ], batch size: 38, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:09:17,736 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7694, 2.2167, 0.8613, 1.1645, 1.3857, 1.1222, 2.4904, 1.2980], + device='cuda:2'), covar=tensor([0.0749, 0.0624, 0.0745, 0.1496, 0.0572, 0.1164, 0.0370, 0.0803], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:09:38,601 INFO [optim.py:369] (2/7) 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 09:09:39,315 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91152.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:09:49,131 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 09:09:51,507 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91163.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:09:53,296 INFO [finetune.py:976] (2/7) Epoch 16, batch 5250, loss[loss=0.2027, simple_loss=0.2713, pruned_loss=0.06705, over 4751.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2498, pruned_loss=0.0545, over 956116.97 frames. ], batch size: 27, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:10:21,398 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.6193, 4.6056, 3.1561, 5.3466, 4.6770, 4.5914, 2.2537, 4.5966], + device='cuda:2'), covar=tensor([0.1571, 0.0861, 0.2682, 0.0790, 0.2690, 0.1508, 0.5233, 0.2006], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0210, 0.0248, 0.0301, 0.0293, 0.0245, 0.0269, 0.0268], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 09:10:44,541 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91200.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:10:47,643 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 16, batch 5300, loss[loss=0.1597, simple_loss=0.2354, pruned_loss=0.04202, over 4776.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2496, pruned_loss=0.05443, over 955582.32 frames. ], batch size: 28, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:11:21,497 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-04-27 09:11:25,281 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 16, batch 5350, loss[loss=0.1787, simple_loss=0.2479, pruned_loss=0.0547, over 4805.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2498, pruned_loss=0.05402, over 957682.49 frames. ], batch size: 41, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:11:47,134 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91285.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:12:08,305 INFO [finetune.py:976] (2/7) Epoch 16, batch 5400, loss[loss=0.1812, simple_loss=0.2548, pruned_loss=0.05379, over 4816.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2473, pruned_loss=0.05316, over 954841.70 frames. ], batch size: 39, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:12:12,803 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9133, 2.4866, 1.9964, 1.7971, 1.3967, 1.4146, 1.9855, 1.3410], + device='cuda:2'), covar=tensor([0.1504, 0.1388, 0.1338, 0.1795, 0.2250, 0.1920, 0.0934, 0.2052], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0214, 0.0170, 0.0205, 0.0202, 0.0186, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 09:12:17,127 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 09:12:18,839 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 09:12:28,675 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 09:12:42,142 INFO [finetune.py:976] (2/7) Epoch 16, batch 5450, loss[loss=0.1505, simple_loss=0.219, pruned_loss=0.04096, over 4777.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2453, pruned_loss=0.0527, over 956592.09 frames. ], batch size: 26, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:13:15,971 INFO [finetune.py:976] (2/7) Epoch 16, batch 5500, loss[loss=0.2058, simple_loss=0.2771, pruned_loss=0.0673, over 4864.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2423, pruned_loss=0.0515, over 956007.86 frames. ], batch size: 44, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:13:17,374 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8667, 1.1305, 1.5146, 1.6427, 1.6175, 1.7400, 1.5653, 1.5745], + device='cuda:2'), covar=tensor([0.3556, 0.4871, 0.4435, 0.4367, 0.5281, 0.6901, 0.4660, 0.4611], + device='cuda:2'), in_proj_covar=tensor([0.0332, 0.0372, 0.0318, 0.0333, 0.0345, 0.0395, 0.0353, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 09:13:19,805 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0755, 2.6792, 2.0087, 2.1490, 1.4740, 1.5490, 2.0953, 1.3404], + device='cuda:2'), covar=tensor([0.1554, 0.1191, 0.1412, 0.1556, 0.2323, 0.2013, 0.0984, 0.2031], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0212, 0.0168, 0.0204, 0.0201, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 09:13:38,281 INFO [optim.py:369] (2/7) 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 09:13:49,958 INFO [finetune.py:976] (2/7) Epoch 16, batch 5550, loss[loss=0.1811, simple_loss=0.2591, pruned_loss=0.05154, over 4816.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2438, pruned_loss=0.05219, over 955899.71 frames. ], batch size: 39, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:13:56,294 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.18 vs. limit=5.0 +2023-04-27 09:14:14,721 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91505.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:14:32,824 INFO [finetune.py:976] (2/7) Epoch 16, batch 5600, loss[loss=0.2067, simple_loss=0.2788, pruned_loss=0.06726, over 4792.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2481, pruned_loss=0.05343, over 954602.15 frames. ], batch size: 51, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:14:34,718 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1761, 1.6190, 2.0426, 2.4000, 2.0267, 1.5871, 1.1856, 1.7299], + device='cuda:2'), covar=tensor([0.3486, 0.3465, 0.1866, 0.2371, 0.2783, 0.2891, 0.4410, 0.2389], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0247, 0.0227, 0.0315, 0.0217, 0.0231, 0.0228, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 09:15:04,817 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91542.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:15:04,855 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0970, 1.5909, 1.9416, 2.1631, 1.8727, 1.5043, 1.0789, 1.6285], + device='cuda:2'), covar=tensor([0.3261, 0.3130, 0.1688, 0.2285, 0.2605, 0.2699, 0.4549, 0.2037], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0247, 0.0227, 0.0315, 0.0217, 0.0231, 0.0228, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 09:15:04,870 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6118, 0.9577, 1.5840, 2.0780, 1.6903, 1.5118, 1.5393, 1.5673], + device='cuda:2'), covar=tensor([0.4594, 0.6759, 0.6580, 0.5806, 0.5840, 0.7854, 0.7521, 0.8564], + device='cuda:2'), in_proj_covar=tensor([0.0418, 0.0407, 0.0493, 0.0504, 0.0447, 0.0470, 0.0476, 0.0483], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:15:15,398 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91553.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:15:28,932 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-04-27 09:15:29,903 INFO [finetune.py:976] (2/7) Epoch 16, batch 5650, loss[loss=0.2102, simple_loss=0.2877, pruned_loss=0.06632, over 4815.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2518, pruned_loss=0.05504, over 955398.43 frames. ], batch size: 38, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:16:20,702 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91603.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:16:34,226 INFO [finetune.py:976] (2/7) Epoch 16, batch 5700, loss[loss=0.1741, simple_loss=0.2329, pruned_loss=0.05764, over 4372.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2489, pruned_loss=0.05465, over 936111.35 frames. ], batch size: 19, lr: 3.42e-03, grad_scale: 32.0 +2023-04-27 09:17:06,040 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91641.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:17:19,306 INFO [finetune.py:976] (2/7) Epoch 17, batch 0, loss[loss=0.1847, simple_loss=0.2569, pruned_loss=0.05632, over 4901.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2569, pruned_loss=0.05632, over 4901.00 frames. ], batch size: 36, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:17:19,306 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 09:17:21,454 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6872, 1.0215, 1.6488, 2.2134, 1.8433, 1.6716, 1.6703, 1.6377], + device='cuda:2'), covar=tensor([0.4560, 0.6743, 0.5893, 0.5830, 0.5739, 0.7688, 0.7825, 0.7699], + device='cuda:2'), in_proj_covar=tensor([0.0420, 0.0408, 0.0496, 0.0506, 0.0449, 0.0471, 0.0479, 0.0485], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:17:21,516 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3331, 1.3102, 3.8925, 3.5527, 3.4610, 3.6958, 3.7890, 3.3944], + device='cuda:2'), covar=tensor([0.7353, 0.5606, 0.1194, 0.2158, 0.1316, 0.1601, 0.0748, 0.1615], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0303, 0.0400, 0.0400, 0.0345, 0.0408, 0.0307, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:17:22,026 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4325, 1.2962, 1.6745, 1.6350, 1.3155, 1.2070, 1.3365, 0.9633], + device='cuda:2'), covar=tensor([0.0494, 0.0574, 0.0400, 0.0458, 0.0721, 0.1036, 0.0471, 0.0545], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0075, 0.0097, 0.0075, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:17:40,834 INFO [finetune.py:1010] (2/7) Epoch 17, validation: loss=0.1535, simple_loss=0.2247, pruned_loss=0.04111, over 2265189.00 frames. +2023-04-27 09:17:40,835 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 09:17:45,697 INFO [optim.py:369] (2/7) 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 09:18:20,574 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 09:18:24,711 INFO [finetune.py:976] (2/7) Epoch 17, batch 50, loss[loss=0.1558, simple_loss=0.2278, pruned_loss=0.04191, over 4842.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2521, pruned_loss=0.05625, over 217100.79 frames. ], batch size: 44, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:18:32,664 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7338, 1.7437, 1.7579, 1.3610, 1.8072, 1.4194, 2.3475, 1.4886], + device='cuda:2'), covar=tensor([0.3718, 0.2021, 0.5256, 0.2922, 0.1803, 0.2608, 0.1585, 0.5135], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0348, 0.0432, 0.0359, 0.0385, 0.0383, 0.0374, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:18:39,440 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-27 09:18:45,331 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1545, 3.0385, 0.8310, 1.5388, 1.6633, 2.0667, 1.7523, 0.9863], + device='cuda:2'), covar=tensor([0.2085, 0.1681, 0.2626, 0.1838, 0.1390, 0.1446, 0.1994, 0.2257], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0243, 0.0137, 0.0121, 0.0132, 0.0153, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:18:57,084 INFO [finetune.py:976] (2/7) Epoch 17, batch 100, loss[loss=0.1765, simple_loss=0.2384, pruned_loss=0.05727, over 4738.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2458, pruned_loss=0.05325, over 382140.78 frames. ], batch size: 54, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:19:02,435 INFO [optim.py:369] (2/7) 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 09:19:03,985 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 09:19:30,053 INFO [finetune.py:976] (2/7) Epoch 17, batch 150, loss[loss=0.1614, simple_loss=0.2401, pruned_loss=0.04139, over 4865.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2403, pruned_loss=0.05059, over 511143.07 frames. ], batch size: 34, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:20:03,535 INFO [finetune.py:976] (2/7) Epoch 17, batch 200, loss[loss=0.167, simple_loss=0.2362, pruned_loss=0.04893, over 4915.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2403, pruned_loss=0.05156, over 610442.35 frames. ], batch size: 35, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:20:04,227 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1101, 1.4642, 1.2837, 1.6362, 1.5604, 1.9378, 1.3093, 3.5668], + device='cuda:2'), covar=tensor([0.0617, 0.0794, 0.0860, 0.1256, 0.0664, 0.0539, 0.0769, 0.0155], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 09:20:08,957 INFO [optim.py:369] (2/7) 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 09:20:35,884 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91891.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:20:36,993 INFO [finetune.py:976] (2/7) Epoch 17, batch 250, loss[loss=0.2052, simple_loss=0.2777, pruned_loss=0.06634, over 4754.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2449, pruned_loss=0.05304, over 687437.40 frames. ], batch size: 59, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:20:40,158 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91898.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:20:41,407 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5852, 1.4705, 1.9568, 1.8656, 1.4259, 1.2862, 1.5504, 0.9834], + device='cuda:2'), covar=tensor([0.0480, 0.0655, 0.0339, 0.0640, 0.0679, 0.1172, 0.0625, 0.0642], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0069, 0.0068, 0.0067, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:21:03,331 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4643, 1.7321, 1.8632, 1.9423, 1.8581, 1.8696, 1.9558, 1.9273], + device='cuda:2'), covar=tensor([0.3867, 0.5288, 0.4698, 0.4451, 0.5387, 0.7430, 0.5196, 0.4982], + device='cuda:2'), in_proj_covar=tensor([0.0332, 0.0372, 0.0318, 0.0332, 0.0345, 0.0396, 0.0353, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 09:21:08,453 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91941.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:21:08,493 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1867, 2.2144, 1.8725, 1.9242, 2.3307, 1.7998, 2.8802, 1.6882], + device='cuda:2'), covar=tensor([0.3661, 0.1863, 0.4464, 0.3203, 0.1798, 0.2742, 0.1166, 0.4235], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0347, 0.0431, 0.0357, 0.0384, 0.0382, 0.0373, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:21:10,055 INFO [finetune.py:976] (2/7) Epoch 17, batch 300, loss[loss=0.1791, simple_loss=0.25, pruned_loss=0.05407, over 4862.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.25, pruned_loss=0.05471, over 748924.68 frames. ], batch size: 31, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:21:15,845 INFO [optim.py:369] (2/7) 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 09:21:16,626 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91952.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:21:18,455 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91955.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:21:51,281 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91981.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:21:56,126 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:22:04,085 INFO [finetune.py:976] (2/7) Epoch 17, batch 350, loss[loss=0.2291, simple_loss=0.2986, pruned_loss=0.07984, over 4891.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2514, pruned_loss=0.0547, over 793493.09 frames. ], batch size: 43, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:22:36,921 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92016.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:23:09,172 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92042.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:23:09,648 INFO [finetune.py:976] (2/7) Epoch 17, batch 400, loss[loss=0.1721, simple_loss=0.2395, pruned_loss=0.05237, over 4779.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2516, pruned_loss=0.05432, over 830183.39 frames. ], batch size: 51, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:23:21,384 INFO [optim.py:369] (2/7) 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 09:24:03,172 INFO [finetune.py:976] (2/7) Epoch 17, batch 450, loss[loss=0.1454, simple_loss=0.2156, pruned_loss=0.03763, over 4789.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2497, pruned_loss=0.05377, over 856731.37 frames. ], batch size: 29, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:24:26,093 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 09:24:32,731 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5902, 0.6726, 1.4168, 1.9062, 1.6233, 1.4832, 1.4570, 1.5525], + device='cuda:2'), covar=tensor([0.5787, 0.7548, 0.8097, 0.9006, 0.7238, 0.9160, 0.9340, 0.9575], + device='cuda:2'), in_proj_covar=tensor([0.0421, 0.0409, 0.0497, 0.0508, 0.0450, 0.0473, 0.0480, 0.0485], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:24:36,847 INFO [finetune.py:976] (2/7) Epoch 17, batch 500, loss[loss=0.1358, simple_loss=0.1994, pruned_loss=0.03606, over 4700.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2474, pruned_loss=0.05296, over 878614.65 frames. ], batch size: 23, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:24:42,153 INFO [optim.py:369] (2/7) 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 09:25:10,286 INFO [finetune.py:976] (2/7) Epoch 17, batch 550, loss[loss=0.2348, simple_loss=0.2879, pruned_loss=0.09087, over 4811.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2448, pruned_loss=0.05252, over 896595.46 frames. ], batch size: 51, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:25:13,459 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92198.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 09:25:26,161 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0164, 2.2692, 1.0125, 1.2559, 1.6735, 1.2631, 2.8734, 1.4776], + device='cuda:2'), covar=tensor([0.0655, 0.0727, 0.0787, 0.1332, 0.0513, 0.0994, 0.0264, 0.0738], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0049, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:25:31,469 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-04-27 09:25:44,127 INFO [finetune.py:976] (2/7) Epoch 17, batch 600, loss[loss=0.1504, simple_loss=0.2167, pruned_loss=0.04209, over 4836.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2451, pruned_loss=0.05297, over 908907.34 frames. ], batch size: 25, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:25:46,052 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92246.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:25:46,672 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92247.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:25:49,030 INFO [optim.py:369] (2/7) 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 09:26:17,387 INFO [finetune.py:976] (2/7) Epoch 17, batch 650, loss[loss=0.165, simple_loss=0.2393, pruned_loss=0.04531, over 4862.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2478, pruned_loss=0.05346, over 919115.93 frames. ], batch size: 31, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:26:22,370 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0812, 2.5234, 0.7592, 1.4144, 1.4142, 1.7974, 1.5107, 0.8141], + device='cuda:2'), covar=tensor([0.1517, 0.1059, 0.1883, 0.1408, 0.1188, 0.1032, 0.1666, 0.1778], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0245, 0.0139, 0.0122, 0.0133, 0.0154, 0.0118, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:26:29,964 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92311.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:26:56,489 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92337.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:27:00,059 INFO [finetune.py:976] (2/7) Epoch 17, batch 700, loss[loss=0.1881, simple_loss=0.26, pruned_loss=0.05808, over 4921.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2479, pruned_loss=0.05316, over 926945.90 frames. ], batch size: 42, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:27:10,671 INFO [optim.py:369] (2/7) 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 09:27:44,599 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1538, 1.8862, 2.0993, 2.4529, 2.4819, 2.0038, 1.6536, 2.3095], + device='cuda:2'), covar=tensor([0.0954, 0.1141, 0.0708, 0.0671, 0.0676, 0.0970, 0.0878, 0.0591], + device='cuda:2'), in_proj_covar=tensor([0.0191, 0.0202, 0.0183, 0.0174, 0.0177, 0.0183, 0.0153, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:27:49,356 INFO [finetune.py:976] (2/7) Epoch 17, batch 750, loss[loss=0.1651, simple_loss=0.2401, pruned_loss=0.04512, over 4822.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2488, pruned_loss=0.05302, over 933427.70 frames. ], batch size: 25, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:28:06,987 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0591, 1.2932, 5.1968, 4.8828, 4.5451, 5.0545, 4.6177, 4.5582], + device='cuda:2'), covar=tensor([0.6544, 0.6679, 0.1035, 0.1747, 0.0944, 0.1728, 0.1265, 0.1588], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0306, 0.0401, 0.0405, 0.0347, 0.0410, 0.0308, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:28:44,566 INFO [finetune.py:976] (2/7) Epoch 17, batch 800, loss[loss=0.1452, simple_loss=0.218, pruned_loss=0.03619, over 4724.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.249, pruned_loss=0.05308, over 941043.17 frames. ], batch size: 27, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:28:54,787 INFO [optim.py:369] (2/7) 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 09:29:31,911 INFO [finetune.py:976] (2/7) Epoch 17, batch 850, loss[loss=0.1847, simple_loss=0.2532, pruned_loss=0.05812, over 4904.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2478, pruned_loss=0.05298, over 945036.04 frames. ], batch size: 35, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:29:52,947 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1583, 1.4780, 4.9887, 4.7046, 4.3350, 4.7197, 4.3878, 4.3396], + device='cuda:2'), covar=tensor([0.6583, 0.6021, 0.0983, 0.1757, 0.1017, 0.1200, 0.1500, 0.1668], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0303, 0.0398, 0.0402, 0.0344, 0.0406, 0.0306, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:30:05,439 INFO [finetune.py:976] (2/7) Epoch 17, batch 900, loss[loss=0.1419, simple_loss=0.216, pruned_loss=0.03393, over 4792.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2454, pruned_loss=0.05228, over 946854.81 frames. ], batch size: 29, lr: 3.41e-03, grad_scale: 32.0 +2023-04-27 09:30:07,971 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 09:30:11,075 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7198, 1.5479, 1.9638, 2.0528, 1.8290, 1.7350, 1.7858, 1.8535], + device='cuda:2'), covar=tensor([0.6108, 0.8217, 0.8635, 0.9974, 0.8055, 1.0968, 1.0795, 1.1640], + device='cuda:2'), in_proj_covar=tensor([0.0420, 0.0407, 0.0496, 0.0506, 0.0450, 0.0472, 0.0479, 0.0483], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:30:19,043 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5578, 1.4707, 1.8763, 1.9171, 1.4293, 1.2561, 1.5569, 0.8900], + device='cuda:2'), covar=tensor([0.0555, 0.0636, 0.0371, 0.0580, 0.0871, 0.1179, 0.0647, 0.0685], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0069, 0.0068, 0.0075, 0.0096, 0.0075, 0.0068], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:30:36,116 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9687, 2.4363, 0.9157, 1.1048, 1.6256, 1.1115, 3.1758, 1.3205], + device='cuda:2'), covar=tensor([0.0837, 0.0848, 0.0958, 0.1898, 0.0730, 0.1430, 0.0455, 0.1074], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:30:38,494 INFO [finetune.py:976] (2/7) Epoch 17, batch 950, loss[loss=0.1717, simple_loss=0.2443, pruned_loss=0.04952, over 4816.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2435, pruned_loss=0.05203, over 948525.59 frames. ], batch size: 40, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:30:39,830 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92595.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:30:49,718 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92611.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:30:59,819 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:31:07,665 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92637.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:31:12,292 INFO [finetune.py:976] (2/7) Epoch 17, batch 1000, loss[loss=0.2221, simple_loss=0.2812, pruned_loss=0.08146, over 4906.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2463, pruned_loss=0.05337, over 948957.87 frames. ], batch size: 35, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:31:17,225 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:31:40,696 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92685.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:31:41,318 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92686.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:31:45,965 INFO [finetune.py:976] (2/7) Epoch 17, batch 1050, loss[loss=0.1517, simple_loss=0.2399, pruned_loss=0.0318, over 4903.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2492, pruned_loss=0.05373, over 951667.74 frames. ], batch size: 36, lr: 3.40e-03, grad_scale: 64.0 +2023-04-27 09:31:56,955 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8590, 2.3610, 1.8220, 1.8379, 1.3615, 1.3917, 1.9269, 1.3144], + device='cuda:2'), covar=tensor([0.1723, 0.1361, 0.1523, 0.1636, 0.2371, 0.2102, 0.1040, 0.2134], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0214, 0.0169, 0.0205, 0.0202, 0.0185, 0.0157, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 09:32:03,083 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:32:30,324 INFO [finetune.py:976] (2/7) Epoch 17, batch 1100, loss[loss=0.2121, simple_loss=0.288, pruned_loss=0.06812, over 4925.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2499, pruned_loss=0.0533, over 952880.45 frames. ], batch size: 38, lr: 3.40e-03, grad_scale: 64.0 +2023-04-27 09:32:36,189 INFO [optim.py:369] (2/7) 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 09:33:11,523 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92781.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:33:32,088 INFO [finetune.py:976] (2/7) Epoch 17, batch 1150, loss[loss=0.1845, simple_loss=0.2558, pruned_loss=0.05657, over 4746.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2511, pruned_loss=0.05354, over 953802.46 frames. ], batch size: 54, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:33:32,341 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-27 09:34:19,592 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92831.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:34:39,529 INFO [finetune.py:976] (2/7) Epoch 17, batch 1200, loss[loss=0.1578, simple_loss=0.2294, pruned_loss=0.04313, over 4858.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2494, pruned_loss=0.05314, over 951729.10 frames. ], batch size: 49, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:34:50,261 INFO [optim.py:369] (2/7) 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 09:35:44,789 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 1250, loss[loss=0.1743, simple_loss=0.2322, pruned_loss=0.05814, over 4856.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2468, pruned_loss=0.05295, over 952111.39 frames. ], batch size: 31, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:35:48,975 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92899.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:36:06,414 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2890, 1.6581, 1.4865, 1.9051, 1.7480, 1.9090, 1.4945, 3.7040], + device='cuda:2'), covar=tensor([0.0558, 0.0741, 0.0747, 0.1106, 0.0605, 0.0478, 0.0706, 0.0139], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 09:36:25,917 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 1300, loss[loss=0.1576, simple_loss=0.2298, pruned_loss=0.04265, over 4941.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2428, pruned_loss=0.0513, over 954503.54 frames. ], batch size: 33, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:36:57,404 INFO [optim.py:369] (2/7) 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 09:37:08,811 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92960.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:10,226 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 09:37:23,989 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92975.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:32,384 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0544, 2.4836, 1.0625, 1.3652, 1.9308, 1.2671, 3.3444, 1.6289], + device='cuda:2'), covar=tensor([0.0680, 0.0688, 0.0766, 0.1231, 0.0500, 0.0986, 0.0231, 0.0668], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0049, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:37:33,568 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:33,617 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7040, 2.0708, 1.7302, 1.9169, 1.4773, 1.7542, 1.7044, 1.3708], + device='cuda:2'), covar=tensor([0.1550, 0.0976, 0.0764, 0.1066, 0.2876, 0.0891, 0.1558, 0.2078], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0305, 0.0220, 0.0277, 0.0312, 0.0258, 0.0249, 0.0266], + device='cuda:2'), out_proj_covar=tensor([1.1421e-04, 1.2137e-04, 8.7365e-05, 1.1002e-04, 1.2696e-04, 1.0281e-04, + 1.0069e-04, 1.0563e-04], device='cuda:2') +2023-04-27 09:37:40,823 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92992.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:37:41,318 INFO [finetune.py:976] (2/7) Epoch 17, batch 1350, loss[loss=0.1981, simple_loss=0.2858, pruned_loss=0.05516, over 4825.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2448, pruned_loss=0.05311, over 953657.25 frames. ], batch size: 40, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:37:54,393 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93010.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:38:16,128 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:38:16,141 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:38:16,170 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3150, 1.8249, 2.2294, 2.7597, 2.2050, 1.7642, 1.5834, 1.9892], + device='cuda:2'), covar=tensor([0.3861, 0.3496, 0.1843, 0.2247, 0.2897, 0.3028, 0.4298, 0.2399], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0245, 0.0223, 0.0312, 0.0215, 0.0228, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 09:38:26,348 INFO [finetune.py:976] (2/7) Epoch 17, batch 1400, loss[loss=0.1641, simple_loss=0.2505, pruned_loss=0.03879, over 4870.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2471, pruned_loss=0.05311, over 955455.85 frames. ], batch size: 34, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:38:44,539 INFO [optim.py:369] (2/7) 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 09:39:08,806 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93071.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:39:11,825 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93076.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:39:28,198 INFO [finetune.py:976] (2/7) Epoch 17, batch 1450, loss[loss=0.1804, simple_loss=0.2565, pruned_loss=0.05214, over 4853.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2495, pruned_loss=0.05333, over 954612.21 frames. ], batch size: 31, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:39:35,634 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8377, 1.4053, 1.3809, 1.5991, 1.9040, 1.5912, 1.2682, 1.3324], + device='cuda:2'), covar=tensor([0.1598, 0.1456, 0.2140, 0.1321, 0.1020, 0.1600, 0.2034, 0.2298], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0310, 0.0351, 0.0288, 0.0326, 0.0310, 0.0300, 0.0365], + device='cuda:2'), out_proj_covar=tensor([6.2977e-05, 6.4685e-05, 7.4828e-05, 5.8596e-05, 6.7801e-05, 6.5308e-05, + 6.3314e-05, 7.7918e-05], device='cuda:2') +2023-04-27 09:39:36,183 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7554, 1.4962, 4.6759, 4.3663, 4.0599, 4.5031, 4.2681, 4.0795], + device='cuda:2'), covar=tensor([0.6973, 0.6511, 0.0929, 0.1624, 0.1191, 0.1924, 0.1095, 0.1456], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0305, 0.0402, 0.0402, 0.0347, 0.0407, 0.0309, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:39:37,280 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93097.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:39:40,710 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9687, 2.3080, 0.9824, 1.2849, 1.8319, 1.2079, 3.0178, 1.5551], + device='cuda:2'), covar=tensor([0.0639, 0.0548, 0.0749, 0.1307, 0.0474, 0.1021, 0.0254, 0.0653], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0049, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:39:51,616 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 09:40:01,747 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:40:20,400 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.6089, 4.4829, 2.9343, 5.2710, 4.5636, 4.4943, 2.1356, 4.7010], + device='cuda:2'), covar=tensor([0.1423, 0.1022, 0.3537, 0.0887, 0.2738, 0.1661, 0.5265, 0.1818], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0213, 0.0250, 0.0304, 0.0297, 0.0249, 0.0273, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 09:40:20,473 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2189, 2.7313, 2.1701, 2.1973, 1.5988, 1.5536, 2.2058, 1.4688], + device='cuda:2'), covar=tensor([0.1659, 0.1571, 0.1260, 0.1657, 0.2297, 0.1933, 0.0958, 0.1962], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0213, 0.0169, 0.0205, 0.0201, 0.0185, 0.0157, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 09:40:24,592 INFO [finetune.py:976] (2/7) Epoch 17, batch 1500, loss[loss=0.1859, simple_loss=0.2584, pruned_loss=0.05675, over 4916.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2519, pruned_loss=0.05477, over 953217.14 frames. ], batch size: 37, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:40:31,494 INFO [optim.py:369] (2/7) 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 09:40:54,015 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:41:06,569 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93187.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:41:15,566 INFO [finetune.py:976] (2/7) Epoch 17, batch 1550, loss[loss=0.2062, simple_loss=0.2578, pruned_loss=0.07725, over 4782.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2517, pruned_loss=0.05512, over 954646.86 frames. ], batch size: 26, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:41:38,289 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6726, 1.3896, 1.3043, 1.4152, 1.8566, 1.5163, 1.2396, 1.2192], + device='cuda:2'), covar=tensor([0.1374, 0.1253, 0.1627, 0.1162, 0.0727, 0.1494, 0.1995, 0.1970], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0311, 0.0351, 0.0289, 0.0327, 0.0311, 0.0301, 0.0366], + device='cuda:2'), out_proj_covar=tensor([6.3031e-05, 6.4916e-05, 7.4822e-05, 5.8855e-05, 6.8030e-05, 6.5533e-05, + 6.3558e-05, 7.8091e-05], device='cuda:2') +2023-04-27 09:41:49,255 INFO [finetune.py:976] (2/7) Epoch 17, batch 1600, loss[loss=0.178, simple_loss=0.2456, pruned_loss=0.05519, over 4819.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2504, pruned_loss=0.0552, over 952651.72 frames. ], batch size: 33, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:41:54,717 INFO [optim.py:369] (2/7) 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 09:41:57,602 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93255.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:42:06,629 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 09:42:15,841 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93281.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:42:19,488 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93287.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:42:23,065 INFO [finetune.py:976] (2/7) Epoch 17, batch 1650, loss[loss=0.1505, simple_loss=0.2185, pruned_loss=0.04123, over 4801.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2488, pruned_loss=0.05476, over 954929.92 frames. ], batch size: 25, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:42:29,355 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1501, 1.6315, 1.4163, 2.0687, 2.1842, 1.8500, 1.7822, 1.5137], + device='cuda:2'), covar=tensor([0.2150, 0.1755, 0.1863, 0.1472, 0.1123, 0.2085, 0.2276, 0.2302], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0312, 0.0352, 0.0290, 0.0328, 0.0311, 0.0301, 0.0366], + device='cuda:2'), out_proj_covar=tensor([6.3348e-05, 6.5089e-05, 7.5041e-05, 5.9047e-05, 6.8241e-05, 6.5644e-05, + 6.3595e-05, 7.8277e-05], device='cuda:2') +2023-04-27 09:42:32,299 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6975, 1.0580, 1.6696, 2.1702, 1.7702, 1.6165, 1.6420, 1.6931], + device='cuda:2'), covar=tensor([0.4328, 0.6047, 0.5858, 0.5380, 0.5493, 0.7589, 0.7326, 0.7700], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0406, 0.0495, 0.0504, 0.0448, 0.0472, 0.0478, 0.0483], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:43:03,618 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:04,869 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93331.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:17,621 INFO [finetune.py:976] (2/7) Epoch 17, batch 1700, loss[loss=0.1428, simple_loss=0.2135, pruned_loss=0.03603, over 4757.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2445, pruned_loss=0.05266, over 955855.43 frames. ], batch size: 27, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:43:22,705 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-04-27 09:43:23,104 INFO [optim.py:369] (2/7) 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 09:43:33,690 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93366.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:41,241 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93376.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:43:51,040 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93392.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 09:43:51,578 INFO [finetune.py:976] (2/7) Epoch 17, batch 1750, loss[loss=0.2029, simple_loss=0.272, pruned_loss=0.06688, over 4757.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.248, pruned_loss=0.05409, over 956306.50 frames. ], batch size: 54, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:44:03,403 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-04-27 09:44:13,251 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93424.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:44:36,169 INFO [finetune.py:976] (2/7) Epoch 17, batch 1800, loss[loss=0.1667, simple_loss=0.24, pruned_loss=0.04673, over 4763.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2504, pruned_loss=0.05429, over 954496.49 frames. ], batch size: 26, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:44:47,781 INFO [optim.py:369] (2/7) 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 09:45:05,917 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:45:22,724 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93487.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:45:26,249 INFO [finetune.py:976] (2/7) Epoch 17, batch 1850, loss[loss=0.2128, simple_loss=0.2848, pruned_loss=0.07036, over 4827.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2516, pruned_loss=0.05466, over 956338.39 frames. ], batch size: 47, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:45:44,481 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5398, 0.7486, 1.5250, 1.9154, 1.5873, 1.5026, 1.5213, 1.5337], + device='cuda:2'), covar=tensor([0.4104, 0.5926, 0.5482, 0.5424, 0.5293, 0.6452, 0.6776, 0.7524], + device='cuda:2'), in_proj_covar=tensor([0.0420, 0.0406, 0.0495, 0.0505, 0.0448, 0.0473, 0.0478, 0.0483], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:46:27,191 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93535.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:46:29,062 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4642, 1.5737, 1.4595, 1.7454, 1.6839, 2.0321, 1.3499, 3.6882], + device='cuda:2'), covar=tensor([0.0575, 0.0826, 0.0846, 0.1225, 0.0662, 0.0491, 0.0824, 0.0140], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 09:46:37,237 INFO [finetune.py:976] (2/7) Epoch 17, batch 1900, loss[loss=0.181, simple_loss=0.2506, pruned_loss=0.05564, over 4820.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2539, pruned_loss=0.0554, over 958998.09 frames. ], batch size: 39, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:46:37,986 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4477, 2.4168, 1.8447, 2.1342, 2.3326, 1.9063, 3.1831, 1.6501], + device='cuda:2'), covar=tensor([0.3614, 0.2124, 0.5093, 0.3240, 0.2036, 0.2724, 0.1696, 0.4881], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0346, 0.0432, 0.0357, 0.0383, 0.0383, 0.0371, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:46:42,803 INFO [optim.py:369] (2/7) 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 09:46:44,101 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8079, 1.7334, 1.6032, 1.2952, 1.7790, 1.4734, 2.2249, 1.4120], + device='cuda:2'), covar=tensor([0.3265, 0.1706, 0.4578, 0.2702, 0.1616, 0.2216, 0.1373, 0.4645], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0346, 0.0432, 0.0357, 0.0383, 0.0383, 0.0371, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:46:44,733 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:03,046 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5229, 1.2768, 4.2257, 3.9483, 3.6639, 3.9999, 3.9379, 3.7545], + device='cuda:2'), covar=tensor([0.7023, 0.6156, 0.1066, 0.1734, 0.1283, 0.1916, 0.1741, 0.1418], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0305, 0.0399, 0.0404, 0.0347, 0.0407, 0.0309, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:47:06,562 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93587.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:07,822 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7862, 2.0007, 1.2154, 1.4212, 2.1963, 1.7000, 1.5465, 1.5691], + device='cuda:2'), covar=tensor([0.0455, 0.0299, 0.0283, 0.0492, 0.0247, 0.0483, 0.0448, 0.0503], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 09:47:10,125 INFO [finetune.py:976] (2/7) Epoch 17, batch 1950, loss[loss=0.1875, simple_loss=0.2467, pruned_loss=0.06421, over 4861.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2509, pruned_loss=0.05437, over 957582.43 frames. ], batch size: 31, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:47:10,823 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8612, 1.2467, 3.2545, 3.0176, 2.9012, 3.1966, 3.1656, 2.8473], + device='cuda:2'), covar=tensor([0.7065, 0.5411, 0.1411, 0.2020, 0.1423, 0.2018, 0.1902, 0.1639], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0304, 0.0398, 0.0403, 0.0346, 0.0406, 0.0308, 0.0362], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 09:47:16,344 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93603.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:35,268 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93631.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:37,608 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93635.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:47:48,923 INFO [finetune.py:976] (2/7) Epoch 17, batch 2000, loss[loss=0.1672, simple_loss=0.2337, pruned_loss=0.05035, over 4792.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2465, pruned_loss=0.05267, over 954984.16 frames. ], batch size: 51, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:47:54,465 INFO [optim.py:369] (2/7) 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 09:48:03,152 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93666.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:11,984 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:21,378 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 2050, loss[loss=0.1518, simple_loss=0.2279, pruned_loss=0.03786, over 4829.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2431, pruned_loss=0.05166, over 953206.64 frames. ], batch size: 40, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:48:23,766 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-27 09:48:35,562 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93714.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:48:53,239 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 2100, loss[loss=0.1794, simple_loss=0.2495, pruned_loss=0.05469, over 4691.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2437, pruned_loss=0.05238, over 954821.56 frames. ], batch size: 23, lr: 3.40e-03, grad_scale: 32.0 +2023-04-27 09:49:01,809 INFO [optim.py:369] (2/7) 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 09:49:11,172 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7277, 1.3788, 1.3949, 1.3985, 1.8302, 1.4855, 1.2201, 1.3034], + device='cuda:2'), covar=tensor([0.1551, 0.1295, 0.1873, 0.1261, 0.0801, 0.1560, 0.1923, 0.2193], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0315, 0.0352, 0.0292, 0.0330, 0.0312, 0.0303, 0.0368], + device='cuda:2'), out_proj_covar=tensor([6.3856e-05, 6.5575e-05, 7.5120e-05, 5.9408e-05, 6.8680e-05, 6.5730e-05, + 6.3855e-05, 7.8554e-05], device='cuda:2') +2023-04-27 09:49:12,960 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93770.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:49:13,577 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93771.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:49:17,400 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 09:49:30,244 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6240, 1.6899, 0.7518, 1.3508, 1.7627, 1.4928, 1.3829, 1.4694], + device='cuda:2'), covar=tensor([0.0539, 0.0394, 0.0404, 0.0594, 0.0308, 0.0549, 0.0570, 0.0610], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 09:49:30,854 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5737, 1.6601, 0.7700, 1.2427, 1.9316, 1.4584, 1.3234, 1.3842], + device='cuda:2'), covar=tensor([0.0528, 0.0407, 0.0361, 0.0595, 0.0269, 0.0537, 0.0535, 0.0625], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 09:49:33,699 INFO [finetune.py:976] (2/7) Epoch 17, batch 2150, loss[loss=0.2325, simple_loss=0.3057, pruned_loss=0.07964, over 4847.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2472, pruned_loss=0.05371, over 956524.00 frames. ], batch size: 47, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:50:13,740 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:50:26,751 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93831.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:50:47,864 INFO [finetune.py:976] (2/7) Epoch 17, batch 2200, loss[loss=0.1655, simple_loss=0.2454, pruned_loss=0.04276, over 4809.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2483, pruned_loss=0.05369, over 956261.37 frames. ], batch size: 51, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:50:59,348 INFO [optim.py:369] (2/7) 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 09:51:41,201 INFO [finetune.py:976] (2/7) Epoch 17, batch 2250, loss[loss=0.1623, simple_loss=0.2312, pruned_loss=0.04665, over 4707.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2487, pruned_loss=0.05359, over 955350.58 frames. ], batch size: 23, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:51:57,961 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-27 09:52:14,469 INFO [finetune.py:976] (2/7) Epoch 17, batch 2300, loss[loss=0.1877, simple_loss=0.2609, pruned_loss=0.0573, over 4922.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2485, pruned_loss=0.05336, over 953779.22 frames. ], batch size: 42, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:52:20,959 INFO [optim.py:369] (2/7) 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 09:52:40,176 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 2350, loss[loss=0.1959, simple_loss=0.2549, pruned_loss=0.06846, over 4901.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2472, pruned_loss=0.05325, over 953251.59 frames. ], batch size: 32, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:53:40,606 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2064, 2.5620, 1.0456, 1.3339, 2.0328, 1.2393, 3.6249, 1.7793], + device='cuda:2'), covar=tensor([0.0657, 0.0787, 0.0891, 0.1355, 0.0570, 0.1054, 0.0343, 0.0707], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0075, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:53:48,602 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94042.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:53:49,094 INFO [finetune.py:976] (2/7) Epoch 17, batch 2400, loss[loss=0.1883, simple_loss=0.2648, pruned_loss=0.05595, over 4915.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2451, pruned_loss=0.05264, over 954587.86 frames. ], batch size: 37, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:53:56,108 INFO [optim.py:369] (2/7) 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 09:54:23,068 INFO [finetune.py:976] (2/7) Epoch 17, batch 2450, loss[loss=0.2046, simple_loss=0.2775, pruned_loss=0.06585, over 4846.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2437, pruned_loss=0.05251, over 954485.40 frames. ], batch size: 47, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:54:46,626 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94126.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:54:57,072 INFO [finetune.py:976] (2/7) Epoch 17, batch 2500, loss[loss=0.1348, simple_loss=0.2139, pruned_loss=0.02779, over 4779.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2452, pruned_loss=0.0533, over 955698.49 frames. ], batch size: 28, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:55:03,648 INFO [optim.py:369] (2/7) 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 09:55:06,696 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5828, 1.4247, 1.8890, 1.9058, 1.4300, 1.2702, 1.5929, 0.9526], + device='cuda:2'), covar=tensor([0.0521, 0.0749, 0.0395, 0.0561, 0.0705, 0.1137, 0.0595, 0.0761], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0068, 0.0068, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:55:12,003 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 09:55:31,027 INFO [finetune.py:976] (2/7) Epoch 17, batch 2550, loss[loss=0.2105, simple_loss=0.2858, pruned_loss=0.06762, over 4823.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2485, pruned_loss=0.05423, over 956387.76 frames. ], batch size: 40, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:55:42,856 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2209, 2.6916, 0.9992, 1.4022, 1.9847, 1.3713, 3.6810, 1.8767], + device='cuda:2'), covar=tensor([0.0630, 0.0589, 0.0767, 0.1228, 0.0518, 0.0889, 0.0179, 0.0584], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0076, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:55:52,788 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94223.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 09:56:08,755 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3187, 1.8147, 2.2293, 2.6733, 2.1841, 1.7938, 1.4269, 1.9157], + device='cuda:2'), covar=tensor([0.3576, 0.3262, 0.1707, 0.2480, 0.2681, 0.2698, 0.4497, 0.2386], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0248, 0.0228, 0.0318, 0.0219, 0.0232, 0.0231, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 09:56:09,849 INFO [finetune.py:976] (2/7) Epoch 17, batch 2600, loss[loss=0.1687, simple_loss=0.235, pruned_loss=0.0512, over 4813.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2498, pruned_loss=0.05468, over 956838.67 frames. ], batch size: 38, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:56:19,533 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2450, 2.6241, 0.9586, 1.2990, 1.9687, 1.4377, 3.5744, 1.8019], + device='cuda:2'), covar=tensor([0.0638, 0.0624, 0.0836, 0.1427, 0.0584, 0.0971, 0.0262, 0.0679], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 09:56:21,227 INFO [optim.py:369] (2/7) 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 09:56:31,033 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 09:56:54,703 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9817, 2.5177, 2.1064, 2.4704, 1.6837, 2.2608, 2.1348, 1.5522], + device='cuda:2'), covar=tensor([0.1998, 0.1263, 0.0822, 0.1036, 0.3410, 0.1104, 0.1810, 0.2775], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0306, 0.0219, 0.0278, 0.0311, 0.0257, 0.0249, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1382e-04, 1.2187e-04, 8.7148e-05, 1.1028e-04, 1.2651e-04, 1.0229e-04, + 1.0082e-04, 1.0507e-04], device='cuda:2') +2023-04-27 09:57:00,087 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4324, 1.7768, 1.6815, 2.2693, 2.4951, 2.0332, 2.0045, 1.7499], + device='cuda:2'), covar=tensor([0.1536, 0.1538, 0.1641, 0.1417, 0.1096, 0.1773, 0.1930, 0.2248], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0312, 0.0350, 0.0289, 0.0327, 0.0309, 0.0300, 0.0364], + device='cuda:2'), out_proj_covar=tensor([6.3408e-05, 6.5016e-05, 7.4687e-05, 5.8785e-05, 6.8096e-05, 6.5163e-05, + 6.3232e-05, 7.7722e-05], device='cuda:2') +2023-04-27 09:57:00,089 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94284.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 09:57:04,407 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 09:57:05,478 INFO [finetune.py:976] (2/7) Epoch 17, batch 2650, loss[loss=0.2021, simple_loss=0.2682, pruned_loss=0.068, over 4838.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2501, pruned_loss=0.05415, over 958283.60 frames. ], batch size: 49, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:57:34,891 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:57:38,464 INFO [finetune.py:976] (2/7) Epoch 17, batch 2700, loss[loss=0.1751, simple_loss=0.2497, pruned_loss=0.05025, over 4750.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2491, pruned_loss=0.05355, over 955744.31 frames. ], batch size: 28, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:57:41,634 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94348.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:57:43,966 INFO [optim.py:369] (2/7) 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 09:57:45,455 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 09:58:17,886 INFO [finetune.py:976] (2/7) Epoch 17, batch 2750, loss[loss=0.1664, simple_loss=0.2279, pruned_loss=0.05242, over 4834.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2476, pruned_loss=0.05389, over 957370.14 frames. ], batch size: 30, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:58:38,889 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94409.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 09:59:01,614 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94426.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 09:59:11,571 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4892, 3.6252, 1.1973, 1.8763, 2.0420, 2.5207, 2.0706, 0.9836], + device='cuda:2'), covar=tensor([0.1420, 0.0815, 0.1793, 0.1289, 0.1117, 0.0966, 0.1514, 0.2036], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0244, 0.0138, 0.0121, 0.0132, 0.0154, 0.0118, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 09:59:24,236 INFO [finetune.py:976] (2/7) Epoch 17, batch 2800, loss[loss=0.1738, simple_loss=0.2431, pruned_loss=0.05222, over 4821.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2447, pruned_loss=0.05313, over 956227.61 frames. ], batch size: 38, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 09:59:35,021 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) Epoch 17, batch 2850, loss[loss=0.1596, simple_loss=0.2488, pruned_loss=0.03522, over 4833.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2441, pruned_loss=0.05284, over 955841.67 frames. ], batch size: 39, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:00:40,995 INFO [finetune.py:976] (2/7) Epoch 17, batch 2900, loss[loss=0.186, simple_loss=0.234, pruned_loss=0.06905, over 4167.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2476, pruned_loss=0.05416, over 954715.29 frames. ], batch size: 18, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:00:46,389 INFO [optim.py:369] (2/7) 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 10:01:03,844 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94579.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 10:01:14,247 INFO [finetune.py:976] (2/7) Epoch 17, batch 2950, loss[loss=0.1795, simple_loss=0.2635, pruned_loss=0.04775, over 4761.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2508, pruned_loss=0.05508, over 955839.48 frames. ], batch size: 28, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:02:00,070 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94637.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:02:04,041 INFO [finetune.py:976] (2/7) Epoch 17, batch 3000, loss[loss=0.1784, simple_loss=0.2544, pruned_loss=0.05116, over 4836.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2521, pruned_loss=0.05587, over 955727.59 frames. ], batch size: 49, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:02:04,041 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 10:02:27,439 INFO [finetune.py:1010] (2/7) Epoch 17, validation: loss=0.1526, simple_loss=0.2233, pruned_loss=0.04089, over 2265189.00 frames. +2023-04-27 10:02:27,440 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 10:02:39,394 INFO [optim.py:369] (2/7) 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 10:03:10,439 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:03:10,508 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0481, 1.0412, 1.2686, 1.2095, 1.0140, 0.9588, 1.0180, 0.5934], + device='cuda:2'), covar=tensor([0.0594, 0.0598, 0.0485, 0.0555, 0.0736, 0.1195, 0.0459, 0.0702], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0070, 0.0069, 0.0068, 0.0076, 0.0096, 0.0075, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:03:15,761 INFO [finetune.py:976] (2/7) Epoch 17, batch 3050, loss[loss=0.1573, simple_loss=0.2441, pruned_loss=0.03524, over 4861.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2521, pruned_loss=0.05549, over 954617.80 frames. ], batch size: 31, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:03:24,032 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94704.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:03:47,900 INFO [finetune.py:976] (2/7) Epoch 17, batch 3100, loss[loss=0.1754, simple_loss=0.2518, pruned_loss=0.04948, over 4894.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2498, pruned_loss=0.05433, over 955263.66 frames. ], batch size: 32, lr: 3.39e-03, grad_scale: 32.0 +2023-04-27 10:03:53,731 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-04-27 10:03:55,523 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 10:03:55,871 INFO [optim.py:369] (2/7) 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 10:04:07,407 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 10:04:13,368 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1586, 4.5568, 0.8344, 2.4708, 2.9474, 2.9352, 2.6025, 1.0138], + device='cuda:2'), covar=tensor([0.1256, 0.0851, 0.2228, 0.1165, 0.0815, 0.1151, 0.1558, 0.2035], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0244, 0.0138, 0.0121, 0.0132, 0.0154, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:04:21,243 INFO [finetune.py:976] (2/7) Epoch 17, batch 3150, loss[loss=0.1375, simple_loss=0.2164, pruned_loss=0.02929, over 4932.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2462, pruned_loss=0.0529, over 951734.43 frames. ], batch size: 33, lr: 3.39e-03, grad_scale: 64.0 +2023-04-27 10:04:26,872 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2144, 2.9803, 2.3136, 2.3734, 1.5990, 1.5136, 2.4484, 1.6389], + device='cuda:2'), covar=tensor([0.1761, 0.1430, 0.1442, 0.1652, 0.2347, 0.1982, 0.0990, 0.2082], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0214, 0.0170, 0.0207, 0.0202, 0.0186, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 10:05:23,158 INFO [finetune.py:976] (2/7) Epoch 17, batch 3200, loss[loss=0.1707, simple_loss=0.237, pruned_loss=0.05222, over 4900.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2437, pruned_loss=0.05197, over 950240.43 frames. ], batch size: 35, lr: 3.39e-03, grad_scale: 64.0 +2023-04-27 10:05:34,599 INFO [optim.py:369] (2/7) 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 10:05:48,287 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.22 vs. limit=5.0 +2023-04-27 10:05:53,673 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94879.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:06:02,200 INFO [finetune.py:976] (2/7) Epoch 17, batch 3250, loss[loss=0.2317, simple_loss=0.2923, pruned_loss=0.08555, over 4855.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2456, pruned_loss=0.05323, over 951701.85 frames. ], batch size: 49, lr: 3.39e-03, grad_scale: 64.0 +2023-04-27 10:06:22,391 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 10:06:26,489 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94927.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:06:36,228 INFO [finetune.py:976] (2/7) Epoch 17, batch 3300, loss[loss=0.2182, simple_loss=0.2778, pruned_loss=0.07936, over 4897.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2497, pruned_loss=0.05458, over 954272.78 frames. ], batch size: 35, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:06:45,993 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-04-27 10:06:47,632 INFO [optim.py:369] (2/7) 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 10:07:30,061 INFO [finetune.py:976] (2/7) Epoch 17, batch 3350, loss[loss=0.1658, simple_loss=0.2431, pruned_loss=0.04424, over 4782.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.251, pruned_loss=0.05493, over 953219.47 frames. ], batch size: 26, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:07:36,976 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95004.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:08:03,820 INFO [finetune.py:976] (2/7) Epoch 17, batch 3400, loss[loss=0.2033, simple_loss=0.2808, pruned_loss=0.06289, over 4840.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2516, pruned_loss=0.05496, over 952542.32 frames. ], batch size: 47, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:08:09,293 INFO [optim.py:369] (2/7) 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 10:08:09,364 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95052.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:08:21,285 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-04-27 10:08:23,992 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6641, 1.4887, 1.3095, 1.5891, 1.9576, 1.5615, 1.3605, 1.2839], + device='cuda:2'), covar=tensor([0.1751, 0.1563, 0.1724, 0.1330, 0.0830, 0.1709, 0.2336, 0.2187], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0316, 0.0356, 0.0292, 0.0330, 0.0313, 0.0302, 0.0369], + device='cuda:2'), out_proj_covar=tensor([6.4269e-05, 6.5963e-05, 7.5927e-05, 5.9235e-05, 6.8445e-05, 6.5940e-05, + 6.3781e-05, 7.8775e-05], device='cuda:2') +2023-04-27 10:08:24,603 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6001, 1.9260, 2.0220, 2.1180, 1.9937, 2.0584, 2.1567, 2.0898], + device='cuda:2'), covar=tensor([0.3737, 0.5004, 0.4549, 0.4811, 0.5419, 0.7483, 0.5291, 0.5052], + device='cuda:2'), in_proj_covar=tensor([0.0329, 0.0371, 0.0317, 0.0332, 0.0342, 0.0394, 0.0353, 0.0323], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:08:37,257 INFO [finetune.py:976] (2/7) Epoch 17, batch 3450, loss[loss=0.1918, simple_loss=0.2648, pruned_loss=0.05934, over 4800.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2519, pruned_loss=0.05542, over 954528.45 frames. ], batch size: 45, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:08:44,629 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95105.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:09:11,014 INFO [finetune.py:976] (2/7) Epoch 17, batch 3500, loss[loss=0.1367, simple_loss=0.221, pruned_loss=0.02621, over 4919.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2493, pruned_loss=0.05457, over 953794.30 frames. ], batch size: 37, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:09:16,410 INFO [optim.py:369] (2/7) 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 10:09:17,179 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8309, 2.0907, 1.7552, 1.3749, 1.3729, 1.3474, 1.7598, 1.2862], + device='cuda:2'), covar=tensor([0.1789, 0.1468, 0.1490, 0.1908, 0.2381, 0.1952, 0.1061, 0.2079], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0212, 0.0168, 0.0205, 0.0200, 0.0184, 0.0156, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 10:09:25,625 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95166.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:09:44,908 INFO [finetune.py:976] (2/7) Epoch 17, batch 3550, loss[loss=0.1461, simple_loss=0.212, pruned_loss=0.04008, over 4766.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2462, pruned_loss=0.05329, over 954295.62 frames. ], batch size: 28, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:10:42,813 INFO [finetune.py:976] (2/7) Epoch 17, batch 3600, loss[loss=0.1904, simple_loss=0.2661, pruned_loss=0.05737, over 4814.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.243, pruned_loss=0.05229, over 954493.58 frames. ], batch size: 41, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:10:54,067 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:11:10,457 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95268.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:11:21,421 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 10:11:29,924 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9613, 1.6125, 1.7879, 2.1163, 2.1231, 1.7984, 1.6715, 1.9905], + device='cuda:2'), covar=tensor([0.0707, 0.1038, 0.0595, 0.0496, 0.0550, 0.0726, 0.0708, 0.0522], + device='cuda:2'), in_proj_covar=tensor([0.0190, 0.0201, 0.0182, 0.0171, 0.0177, 0.0180, 0.0152, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:11:34,140 INFO [finetune.py:976] (2/7) Epoch 17, batch 3650, loss[loss=0.1603, simple_loss=0.2255, pruned_loss=0.04753, over 4222.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2435, pruned_loss=0.05233, over 951752.02 frames. ], batch size: 18, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:11:51,439 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:11:57,167 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6442, 1.2709, 4.2665, 3.9989, 3.7016, 3.9711, 3.9021, 3.7553], + device='cuda:2'), covar=tensor([0.7031, 0.6058, 0.0980, 0.1486, 0.1065, 0.1938, 0.2198, 0.1443], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0304, 0.0400, 0.0405, 0.0346, 0.0405, 0.0309, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:11:57,812 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:12:05,320 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 3700, loss[loss=0.1833, simple_loss=0.2609, pruned_loss=0.05286, over 4826.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2471, pruned_loss=0.0532, over 951261.12 frames. ], batch size: 33, lr: 3.38e-03, grad_scale: 64.0 +2023-04-27 10:12:13,171 INFO [optim.py:369] (2/7) 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 10:12:41,441 INFO [finetune.py:976] (2/7) Epoch 17, batch 3750, loss[loss=0.2128, simple_loss=0.2772, pruned_loss=0.07418, over 4879.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2493, pruned_loss=0.05411, over 950921.64 frames. ], batch size: 32, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:12:46,019 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95400.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:13:06,436 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2750, 1.8276, 2.1031, 2.4833, 2.4928, 2.0091, 1.7836, 2.3041], + device='cuda:2'), covar=tensor([0.0797, 0.1122, 0.0660, 0.0606, 0.0610, 0.0809, 0.0741, 0.0553], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0199, 0.0180, 0.0169, 0.0175, 0.0179, 0.0151, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:13:14,397 INFO [finetune.py:976] (2/7) Epoch 17, batch 3800, loss[loss=0.2102, simple_loss=0.2834, pruned_loss=0.06848, over 4800.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2508, pruned_loss=0.0548, over 951896.98 frames. ], batch size: 40, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:13:20,913 INFO [optim.py:369] (2/7) 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 10:13:22,942 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-27 10:13:25,883 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95461.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:13:33,854 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 10:13:48,140 INFO [finetune.py:976] (2/7) Epoch 17, batch 3850, loss[loss=0.1501, simple_loss=0.2313, pruned_loss=0.03443, over 4911.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2495, pruned_loss=0.05425, over 950845.06 frames. ], batch size: 46, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:13:53,087 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 10:14:20,427 INFO [finetune.py:976] (2/7) Epoch 17, batch 3900, loss[loss=0.1391, simple_loss=0.2091, pruned_loss=0.03451, over 4819.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2479, pruned_loss=0.05433, over 951999.23 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:14:27,544 INFO [optim.py:369] (2/7) 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 10:14:27,664 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2805, 1.5394, 1.3513, 1.7497, 1.6480, 1.8074, 1.4129, 3.3609], + device='cuda:2'), covar=tensor([0.0573, 0.0763, 0.0772, 0.1154, 0.0610, 0.0527, 0.0725, 0.0128], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 10:14:31,987 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8509, 1.4415, 2.0118, 2.4012, 1.9632, 1.8688, 1.9073, 1.8733], + device='cuda:2'), covar=tensor([0.4645, 0.6602, 0.6454, 0.5556, 0.5788, 0.7246, 0.8117, 0.9302], + device='cuda:2'), in_proj_covar=tensor([0.0419, 0.0406, 0.0495, 0.0502, 0.0447, 0.0473, 0.0479, 0.0484], + device='cuda:2'), out_proj_covar=tensor([1.0075e-04, 9.9850e-05, 1.1125e-04, 1.1972e-04, 1.0721e-04, 1.1354e-04, + 1.1376e-04, 1.1434e-04], device='cuda:2') +2023-04-27 10:14:34,997 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3911, 2.9680, 0.9045, 1.4325, 2.1915, 1.2800, 3.9798, 1.9876], + device='cuda:2'), covar=tensor([0.0655, 0.0961, 0.0927, 0.1236, 0.0537, 0.0981, 0.0340, 0.0600], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 10:14:52,351 INFO [finetune.py:976] (2/7) Epoch 17, batch 3950, loss[loss=0.2019, simple_loss=0.2492, pruned_loss=0.07727, over 4766.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2448, pruned_loss=0.05323, over 954706.76 frames. ], batch size: 26, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:14:58,354 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9704, 2.3329, 2.2091, 2.3459, 2.0988, 2.2363, 2.2867, 2.2241], + device='cuda:2'), covar=tensor([0.4584, 0.6551, 0.4968, 0.4958, 0.6040, 0.7413, 0.6517, 0.5597], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0371, 0.0317, 0.0332, 0.0344, 0.0394, 0.0354, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:15:08,536 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:15:09,187 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:15:10,973 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6461, 2.9783, 1.0602, 1.6373, 2.5036, 1.6288, 4.3168, 2.3010], + device='cuda:2'), covar=tensor([0.0634, 0.0713, 0.0840, 0.1296, 0.0504, 0.0964, 0.0213, 0.0603], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0075, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 10:15:13,451 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95624.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:15:31,265 INFO [finetune.py:976] (2/7) Epoch 17, batch 4000, loss[loss=0.1681, simple_loss=0.2425, pruned_loss=0.04684, over 4871.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2432, pruned_loss=0.05245, over 955915.19 frames. ], batch size: 31, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:15:43,874 INFO [optim.py:369] (2/7) 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 10:16:16,162 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95678.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:16:24,323 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2575, 1.5356, 1.6535, 1.8115, 1.7122, 1.7453, 1.8055, 1.7109], + device='cuda:2'), covar=tensor([0.4237, 0.5676, 0.4634, 0.4567, 0.5592, 0.7261, 0.4854, 0.4948], + device='cuda:2'), in_proj_covar=tensor([0.0333, 0.0374, 0.0319, 0.0335, 0.0347, 0.0397, 0.0356, 0.0327], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:16:36,027 INFO [finetune.py:976] (2/7) Epoch 17, batch 4050, loss[loss=0.1723, simple_loss=0.2556, pruned_loss=0.04455, over 4760.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2464, pruned_loss=0.05385, over 953187.94 frames. ], batch size: 28, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:16:37,790 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95695.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:17:31,214 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0857, 1.7713, 2.1020, 2.5508, 2.5015, 2.1621, 1.6976, 2.2903], + device='cuda:2'), covar=tensor([0.0860, 0.1074, 0.0689, 0.0475, 0.0547, 0.0765, 0.0770, 0.0486], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0200, 0.0181, 0.0170, 0.0176, 0.0178, 0.0151, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:17:36,636 INFO [finetune.py:976] (2/7) Epoch 17, batch 4100, loss[loss=0.1747, simple_loss=0.2527, pruned_loss=0.04837, over 4755.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2491, pruned_loss=0.05413, over 955826.30 frames. ], batch size: 28, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:17:43,691 INFO [optim.py:369] (2/7) 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 10:17:50,086 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95761.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:18:09,903 INFO [finetune.py:976] (2/7) Epoch 17, batch 4150, loss[loss=0.1806, simple_loss=0.2674, pruned_loss=0.04691, over 4922.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2503, pruned_loss=0.05461, over 953734.15 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:18:21,840 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95809.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:18:24,199 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95812.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:18:24,879 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-27 10:18:43,506 INFO [finetune.py:976] (2/7) Epoch 17, batch 4200, loss[loss=0.189, simple_loss=0.2712, pruned_loss=0.05338, over 4814.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2507, pruned_loss=0.0543, over 954007.45 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:18:46,042 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95847.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:18:49,672 INFO [optim.py:369] (2/7) 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 10:19:02,014 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3068, 1.9448, 2.4975, 2.7964, 2.3655, 2.2265, 2.3239, 2.2881], + device='cuda:2'), covar=tensor([0.4492, 0.6173, 0.6111, 0.5610, 0.5513, 0.7979, 0.7900, 0.8459], + device='cuda:2'), in_proj_covar=tensor([0.0422, 0.0408, 0.0499, 0.0506, 0.0451, 0.0476, 0.0482, 0.0487], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:19:04,194 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95873.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:19:16,329 INFO [finetune.py:976] (2/7) Epoch 17, batch 4250, loss[loss=0.1822, simple_loss=0.2669, pruned_loss=0.04877, over 4825.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2494, pruned_loss=0.05356, over 955037.17 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:19:16,460 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4911, 1.4114, 1.8654, 1.8237, 1.3820, 1.2584, 1.5954, 0.9543], + device='cuda:2'), covar=tensor([0.0587, 0.0717, 0.0373, 0.0592, 0.0751, 0.1171, 0.0553, 0.0646], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0070, 0.0068, 0.0068, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:19:26,049 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95908.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:19:28,955 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3481, 3.3011, 2.5601, 3.9102, 3.4025, 3.3211, 1.4074, 3.3874], + device='cuda:2'), covar=tensor([0.1993, 0.1566, 0.3152, 0.2334, 0.3134, 0.2319, 0.5909, 0.2541], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0213, 0.0249, 0.0304, 0.0297, 0.0249, 0.0271, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:19:31,943 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:19:35,084 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2293, 1.7039, 2.0717, 2.5849, 2.0845, 1.6741, 1.4416, 1.8971], + device='cuda:2'), covar=tensor([0.3151, 0.3109, 0.1630, 0.2170, 0.2525, 0.2572, 0.4312, 0.2071], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0248, 0.0229, 0.0318, 0.0219, 0.0233, 0.0230, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 10:19:36,855 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95924.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:19:48,686 INFO [finetune.py:976] (2/7) Epoch 17, batch 4300, loss[loss=0.1529, simple_loss=0.2376, pruned_loss=0.03409, over 4828.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2458, pruned_loss=0.05226, over 955483.05 frames. ], batch size: 30, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:19:54,839 INFO [optim.py:369] (2/7) 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 10:20:03,626 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:20:09,895 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95973.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:20:14,637 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 4350, loss[loss=0.1561, simple_loss=0.2421, pruned_loss=0.03508, over 4786.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2421, pruned_loss=0.05063, over 954818.50 frames. ], batch size: 29, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:20:23,767 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95995.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:21:07,905 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96041.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:21:08,980 INFO [finetune.py:976] (2/7) Epoch 17, batch 4400, loss[loss=0.1878, simple_loss=0.2627, pruned_loss=0.05642, over 4808.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2437, pruned_loss=0.05164, over 954287.70 frames. ], batch size: 41, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:21:09,043 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96043.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:21:15,036 INFO [optim.py:369] (2/7) 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 10:21:21,397 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7447, 1.3780, 1.3774, 1.4056, 1.8718, 1.5715, 1.2830, 1.3692], + device='cuda:2'), covar=tensor([0.1371, 0.1256, 0.1768, 0.1295, 0.0761, 0.1369, 0.1707, 0.1825], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0313, 0.0351, 0.0290, 0.0328, 0.0310, 0.0300, 0.0366], + device='cuda:2'), out_proj_covar=tensor([6.4099e-05, 6.5286e-05, 7.4872e-05, 5.8842e-05, 6.8198e-05, 6.5345e-05, + 6.3237e-05, 7.7978e-05], device='cuda:2') +2023-04-27 10:21:34,765 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:21:48,839 INFO [finetune.py:976] (2/7) Epoch 17, batch 4450, loss[loss=0.1793, simple_loss=0.2545, pruned_loss=0.05205, over 4903.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2456, pruned_loss=0.05175, over 954036.56 frames. ], batch size: 36, lr: 3.38e-03, grad_scale: 32.0 +2023-04-27 10:21:59,259 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5466, 2.0306, 2.5361, 3.1390, 2.9204, 2.4591, 2.0138, 2.6418], + device='cuda:2'), covar=tensor([0.0794, 0.1159, 0.0697, 0.0601, 0.0538, 0.0751, 0.0773, 0.0626], + device='cuda:2'), in_proj_covar=tensor([0.0189, 0.0201, 0.0181, 0.0171, 0.0176, 0.0180, 0.0151, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:22:36,180 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 10:22:53,791 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 4500, loss[loss=0.1616, simple_loss=0.2272, pruned_loss=0.04798, over 4883.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2482, pruned_loss=0.05268, over 954755.98 frames. ], batch size: 35, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:23:01,012 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4614, 3.3812, 0.9536, 1.8881, 1.8645, 2.4421, 2.0224, 0.9930], + device='cuda:2'), covar=tensor([0.1423, 0.0915, 0.1914, 0.1221, 0.1073, 0.0979, 0.1460, 0.2032], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0242, 0.0136, 0.0120, 0.0132, 0.0152, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:23:02,754 INFO [optim.py:369] (2/7) 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 10:23:14,572 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 10:23:16,181 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:23:17,978 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96168.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 10:23:58,142 INFO [finetune.py:976] (2/7) Epoch 17, batch 4550, loss[loss=0.1961, simple_loss=0.2621, pruned_loss=0.065, over 4913.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2514, pruned_loss=0.05459, over 954899.03 frames. ], batch size: 36, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:24:10,600 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96203.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:24:21,979 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1315, 0.8497, 0.9127, 0.8047, 1.2542, 1.0641, 0.9301, 0.9554], + device='cuda:2'), covar=tensor([0.1583, 0.1327, 0.1944, 0.1540, 0.0908, 0.1230, 0.1441, 0.1967], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0314, 0.0352, 0.0288, 0.0328, 0.0310, 0.0300, 0.0366], + device='cuda:2'), out_proj_covar=tensor([6.3909e-05, 6.5386e-05, 7.4990e-05, 5.8464e-05, 6.8112e-05, 6.5283e-05, + 6.3286e-05, 7.8009e-05], device='cuda:2') +2023-04-27 10:24:42,292 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96226.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:24:54,976 INFO [finetune.py:976] (2/7) Epoch 17, batch 4600, loss[loss=0.1657, simple_loss=0.2368, pruned_loss=0.04728, over 4898.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.249, pruned_loss=0.05299, over 953541.11 frames. ], batch size: 35, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:25:01,052 INFO [optim.py:369] (2/7) 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 10:25:01,845 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.07 vs. limit=5.0 +2023-04-27 10:25:12,848 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 10:25:13,335 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:25:25,582 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7344, 1.2841, 1.8465, 2.2967, 1.9134, 1.7322, 1.8202, 1.7588], + device='cuda:2'), covar=tensor([0.4275, 0.6369, 0.5701, 0.5010, 0.5212, 0.7288, 0.7073, 0.7549], + device='cuda:2'), in_proj_covar=tensor([0.0421, 0.0408, 0.0497, 0.0504, 0.0449, 0.0475, 0.0482, 0.0486], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:25:27,832 INFO [finetune.py:976] (2/7) Epoch 17, batch 4650, loss[loss=0.2087, simple_loss=0.2566, pruned_loss=0.08038, over 4815.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2475, pruned_loss=0.05313, over 955028.81 frames. ], batch size: 39, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:25:32,033 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4393, 3.2767, 2.6477, 2.8697, 2.3262, 2.8554, 2.7941, 2.0496], + device='cuda:2'), covar=tensor([0.2512, 0.1483, 0.0864, 0.1300, 0.3209, 0.1243, 0.2078, 0.3283], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0305, 0.0218, 0.0278, 0.0311, 0.0258, 0.0248, 0.0265], + device='cuda:2'), out_proj_covar=tensor([1.1358e-04, 1.2135e-04, 8.6581e-05, 1.1023e-04, 1.2646e-04, 1.0241e-04, + 1.0015e-04, 1.0540e-04], device='cuda:2') +2023-04-27 10:25:35,115 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 10:25:45,390 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:25:47,235 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96324.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:25:56,371 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:26:01,178 INFO [finetune.py:976] (2/7) Epoch 17, batch 4700, loss[loss=0.1541, simple_loss=0.2211, pruned_loss=0.04351, over 4908.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2437, pruned_loss=0.05192, over 955612.66 frames. ], batch size: 36, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:26:03,064 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4118, 1.7157, 1.5835, 1.9613, 1.8413, 2.1848, 1.4827, 3.8126], + device='cuda:2'), covar=tensor([0.0522, 0.0766, 0.0765, 0.1119, 0.0587, 0.0423, 0.0732, 0.0149], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 10:26:07,579 INFO [optim.py:369] (2/7) 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 10:26:35,976 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4725, 1.4312, 1.7817, 1.8140, 1.3886, 1.1725, 1.5296, 1.0218], + device='cuda:2'), covar=tensor([0.0805, 0.0700, 0.0464, 0.0649, 0.0890, 0.1316, 0.0620, 0.0703], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0075, 0.0095, 0.0074, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:26:38,852 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:26:48,559 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:26:55,845 INFO [finetune.py:976] (2/7) Epoch 17, batch 4750, loss[loss=0.1747, simple_loss=0.2495, pruned_loss=0.05, over 4798.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.242, pruned_loss=0.05163, over 956129.78 frames. ], batch size: 29, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:27:07,410 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 10:27:42,418 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0260, 1.5598, 1.8815, 2.2843, 1.8608, 1.5175, 1.1368, 1.7052], + device='cuda:2'), covar=tensor([0.3003, 0.2914, 0.1621, 0.1861, 0.2466, 0.2467, 0.4095, 0.1939], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0247, 0.0226, 0.0316, 0.0217, 0.0230, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 10:27:44,796 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 4800, loss[loss=0.147, simple_loss=0.2151, pruned_loss=0.03946, over 4763.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2458, pruned_loss=0.05328, over 955529.79 frames. ], batch size: 26, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:27:54,028 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0591, 1.5613, 1.9640, 2.4008, 1.9390, 1.5316, 1.2796, 1.7169], + device='cuda:2'), covar=tensor([0.3085, 0.3099, 0.1633, 0.2105, 0.2612, 0.2571, 0.4446, 0.2068], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0248, 0.0228, 0.0317, 0.0218, 0.0232, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 10:27:56,400 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96451.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:27:57,985 INFO [optim.py:369] (2/7) 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 10:28:07,159 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96468.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:28:34,487 INFO [finetune.py:976] (2/7) Epoch 17, batch 4850, loss[loss=0.2274, simple_loss=0.285, pruned_loss=0.0849, over 4847.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2485, pruned_loss=0.05366, over 954683.20 frames. ], batch size: 47, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:28:35,123 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7606, 1.0504, 1.7217, 2.2369, 1.8270, 1.6927, 1.6881, 1.7074], + device='cuda:2'), covar=tensor([0.4804, 0.6844, 0.6310, 0.5745, 0.6303, 0.8107, 0.8089, 0.7703], + device='cuda:2'), in_proj_covar=tensor([0.0421, 0.0408, 0.0497, 0.0504, 0.0450, 0.0476, 0.0481, 0.0486], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:28:39,375 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.37 vs. limit=5.0 +2023-04-27 10:28:41,624 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96503.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:28:49,918 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96521.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:29:07,106 INFO [finetune.py:976] (2/7) Epoch 17, batch 4900, loss[loss=0.2012, simple_loss=0.2679, pruned_loss=0.06721, over 4862.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2494, pruned_loss=0.05384, over 954986.78 frames. ], batch size: 44, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:29:13,477 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96551.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:29:15,080 INFO [optim.py:369] (2/7) 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 10:29:21,150 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3883, 1.3913, 1.7969, 1.7253, 1.2939, 1.1089, 1.4894, 0.9148], + device='cuda:2'), covar=tensor([0.0728, 0.0778, 0.0453, 0.0724, 0.0912, 0.1246, 0.0690, 0.0717], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0074, 0.0095, 0.0074, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:29:39,869 INFO [finetune.py:976] (2/7) Epoch 17, batch 4950, loss[loss=0.1964, simple_loss=0.263, pruned_loss=0.06494, over 4804.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2509, pruned_loss=0.05406, over 955776.62 frames. ], batch size: 40, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:30:09,369 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96636.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:30:13,519 INFO [finetune.py:976] (2/7) Epoch 17, batch 5000, loss[loss=0.1773, simple_loss=0.2491, pruned_loss=0.05272, over 4739.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2492, pruned_loss=0.05296, over 955995.00 frames. ], batch size: 59, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:30:21,474 INFO [optim.py:369] (2/7) 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 10:30:38,902 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:30:39,083 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 10:30:41,301 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96684.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:30:46,661 INFO [finetune.py:976] (2/7) Epoch 17, batch 5050, loss[loss=0.1894, simple_loss=0.2548, pruned_loss=0.06201, over 4908.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2478, pruned_loss=0.05306, over 956825.23 frames. ], batch size: 36, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:30:59,708 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-27 10:31:15,679 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:31:19,859 INFO [finetune.py:976] (2/7) Epoch 17, batch 5100, loss[loss=0.1788, simple_loss=0.243, pruned_loss=0.05727, over 4208.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2455, pruned_loss=0.05308, over 956864.13 frames. ], batch size: 65, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:31:21,741 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96746.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:31:23,065 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 10:31:25,901 INFO [optim.py:369] (2/7) 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 10:31:54,160 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:32:15,345 INFO [finetune.py:976] (2/7) Epoch 17, batch 5150, loss[loss=0.1535, simple_loss=0.2229, pruned_loss=0.04207, over 4738.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2454, pruned_loss=0.05312, over 954664.49 frames. ], batch size: 23, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:32:22,307 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7350, 2.2459, 1.9504, 1.7523, 1.2903, 1.3770, 2.0394, 1.2525], + device='cuda:2'), covar=tensor([0.1789, 0.1667, 0.1482, 0.1933, 0.2504, 0.2088, 0.0931, 0.2241], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0212, 0.0167, 0.0204, 0.0200, 0.0183, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 10:32:46,410 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:32:48,221 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8953, 2.9173, 2.2143, 3.3165, 2.9600, 2.9084, 1.2433, 2.8160], + device='cuda:2'), covar=tensor([0.2278, 0.1657, 0.3498, 0.2911, 0.3588, 0.2226, 0.5981, 0.3060], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0214, 0.0251, 0.0308, 0.0299, 0.0250, 0.0274, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:33:03,921 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 5200, loss[loss=0.1593, simple_loss=0.2387, pruned_loss=0.03992, over 4750.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2482, pruned_loss=0.05391, over 952625.83 frames. ], batch size: 54, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:33:24,747 INFO [optim.py:369] (2/7) 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 10:33:47,098 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96869.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:33:53,354 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 10:34:02,258 INFO [finetune.py:976] (2/7) Epoch 17, batch 5250, loss[loss=0.19, simple_loss=0.27, pruned_loss=0.05494, over 4934.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2502, pruned_loss=0.05404, over 953191.82 frames. ], batch size: 42, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:34:16,526 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 10:34:17,668 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4354, 1.7820, 1.7840, 1.8919, 1.7526, 1.8223, 1.9168, 1.8345], + device='cuda:2'), covar=tensor([0.4273, 0.5696, 0.4977, 0.4577, 0.5892, 0.7750, 0.5565, 0.5320], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0371, 0.0318, 0.0333, 0.0343, 0.0393, 0.0354, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:34:29,509 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8264, 2.1962, 0.8593, 1.1834, 1.4974, 1.1286, 2.4655, 1.3780], + device='cuda:2'), covar=tensor([0.0670, 0.0573, 0.0652, 0.1185, 0.0471, 0.0981, 0.0262, 0.0676], + device='cuda:2'), in_proj_covar=tensor([0.0052, 0.0066, 0.0049, 0.0047, 0.0050, 0.0053, 0.0075, 0.0052], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 10:34:35,558 INFO [finetune.py:976] (2/7) Epoch 17, batch 5300, loss[loss=0.1639, simple_loss=0.2113, pruned_loss=0.05825, over 4140.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2514, pruned_loss=0.05405, over 953314.39 frames. ], batch size: 18, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:34:38,177 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 10:34:41,675 INFO [optim.py:369] (2/7) 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 10:35:00,589 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 5350, loss[loss=0.1953, simple_loss=0.2567, pruned_loss=0.06702, over 4903.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2513, pruned_loss=0.05401, over 954032.29 frames. ], batch size: 37, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:35:13,317 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97000.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:35:28,352 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0647, 1.3960, 5.1667, 4.8610, 4.5101, 4.9002, 4.6367, 4.5107], + device='cuda:2'), covar=tensor([0.6404, 0.6058, 0.1068, 0.1951, 0.1133, 0.1241, 0.1173, 0.1500], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0304, 0.0400, 0.0404, 0.0348, 0.0403, 0.0309, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:35:33,058 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97028.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:35:42,641 INFO [finetune.py:976] (2/7) Epoch 17, batch 5400, loss[loss=0.1757, simple_loss=0.2465, pruned_loss=0.05247, over 4819.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2489, pruned_loss=0.05333, over 953669.10 frames. ], batch size: 39, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:35:44,589 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97046.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:35:45,964 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 10:35:48,747 INFO [optim.py:369] (2/7) 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 10:35:53,796 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97061.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:36:15,419 INFO [finetune.py:976] (2/7) Epoch 17, batch 5450, loss[loss=0.13, simple_loss=0.2019, pruned_loss=0.02902, over 4731.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2456, pruned_loss=0.05271, over 952324.12 frames. ], batch size: 59, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:36:16,110 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97094.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:36:40,622 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 17, batch 5500, loss[loss=0.1673, simple_loss=0.2299, pruned_loss=0.05232, over 4848.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2429, pruned_loss=0.05184, over 953169.48 frames. ], batch size: 47, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:36:54,343 INFO [optim.py:369] (2/7) 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 10:36:54,471 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6709, 2.7215, 2.0680, 2.4079, 2.5877, 2.2909, 3.4565, 1.8639], + device='cuda:2'), covar=tensor([0.3102, 0.1782, 0.3916, 0.2677, 0.1706, 0.2440, 0.1440, 0.4089], + device='cuda:2'), in_proj_covar=tensor([0.0334, 0.0343, 0.0422, 0.0352, 0.0378, 0.0377, 0.0367, 0.0415], + device='cuda:2'), out_proj_covar=tensor([9.9753e-05, 1.0319e-04, 1.2845e-04, 1.0637e-04, 1.1296e-04, 1.1292e-04, + 1.0814e-04, 1.2597e-04], device='cuda:2') +2023-04-27 10:37:37,879 INFO [finetune.py:976] (2/7) Epoch 17, batch 5550, loss[loss=0.2217, simple_loss=0.2958, pruned_loss=0.07379, over 4815.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.243, pruned_loss=0.05177, over 953274.79 frames. ], batch size: 40, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:37:38,662 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2974, 2.7677, 2.2419, 2.2656, 1.7986, 1.7806, 2.3268, 1.7012], + device='cuda:2'), covar=tensor([0.1538, 0.1321, 0.1295, 0.1428, 0.2187, 0.1770, 0.0927, 0.1922], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0213, 0.0168, 0.0205, 0.0201, 0.0184, 0.0156, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 10:38:15,511 INFO [finetune.py:976] (2/7) Epoch 17, batch 5600, loss[loss=0.1958, simple_loss=0.2773, pruned_loss=0.05712, over 4903.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2477, pruned_loss=0.05299, over 953792.63 frames. ], batch size: 43, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:38:26,564 INFO [optim.py:369] (2/7) 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 10:38:37,064 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97260.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:38:47,529 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1524, 1.4479, 1.3358, 1.6745, 1.6274, 1.8143, 1.3812, 3.3135], + device='cuda:2'), covar=tensor([0.0593, 0.0834, 0.0816, 0.1172, 0.0612, 0.0543, 0.0809, 0.0149], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 10:39:14,493 INFO [finetune.py:976] (2/7) Epoch 17, batch 5650, loss[loss=0.1896, simple_loss=0.2697, pruned_loss=0.05475, over 4824.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2513, pruned_loss=0.05419, over 953027.39 frames. ], batch size: 33, lr: 3.37e-03, grad_scale: 32.0 +2023-04-27 10:39:47,534 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:40:07,563 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6517, 3.2036, 1.0868, 1.8954, 2.0547, 2.3200, 2.0342, 1.2488], + device='cuda:2'), covar=tensor([0.1289, 0.1023, 0.1845, 0.1110, 0.0864, 0.1001, 0.1407, 0.1586], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0243, 0.0138, 0.0121, 0.0133, 0.0154, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:40:16,392 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97342.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:40:16,914 INFO [finetune.py:976] (2/7) Epoch 17, batch 5700, loss[loss=0.1466, simple_loss=0.2046, pruned_loss=0.04431, over 3824.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2479, pruned_loss=0.0537, over 935090.87 frames. ], batch size: 16, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:40:19,361 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 10:40:24,677 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:40:30,699 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.02 vs. limit=5.0 +2023-04-27 10:40:46,576 INFO [finetune.py:976] (2/7) Epoch 18, batch 0, loss[loss=0.1215, simple_loss=0.2007, pruned_loss=0.02109, over 4746.00 frames. ], tot_loss[loss=0.1215, simple_loss=0.2007, pruned_loss=0.02109, over 4746.00 frames. ], batch size: 27, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:40:46,577 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 10:41:03,166 INFO [finetune.py:1010] (2/7) Epoch 18, validation: loss=0.1537, simple_loss=0.225, pruned_loss=0.04121, over 2265189.00 frames. +2023-04-27 10:41:03,166 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 10:41:05,468 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4157, 3.0286, 1.0152, 1.8099, 1.8213, 2.2646, 1.8764, 1.0310], + device='cuda:2'), covar=tensor([0.1287, 0.0873, 0.1732, 0.1094, 0.0965, 0.0856, 0.1303, 0.1747], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0244, 0.0138, 0.0121, 0.0133, 0.0154, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:41:05,495 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97374.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:41:28,957 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97403.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:41:32,012 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97408.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:41:34,596 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 10:41:40,832 INFO [finetune.py:976] (2/7) Epoch 18, batch 50, loss[loss=0.1947, simple_loss=0.2669, pruned_loss=0.06123, over 4823.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2518, pruned_loss=0.05478, over 217438.77 frames. ], batch size: 39, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:41:49,625 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:41:51,453 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97435.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:41:52,228 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-04-27 10:42:00,566 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0603, 2.7206, 2.2697, 2.5270, 2.0353, 2.2955, 2.2272, 1.6396], + device='cuda:2'), covar=tensor([0.2093, 0.1467, 0.0856, 0.1443, 0.3037, 0.1311, 0.1884, 0.3012], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0306, 0.0220, 0.0280, 0.0312, 0.0260, 0.0251, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1454e-04, 1.2182e-04, 8.7461e-05, 1.1112e-04, 1.2686e-04, 1.0321e-04, + 1.0139e-04, 1.0613e-04], device='cuda:2') +2023-04-27 10:42:02,221 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 100, loss[loss=0.1328, simple_loss=0.2084, pruned_loss=0.02863, over 4795.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2432, pruned_loss=0.05207, over 378599.49 frames. ], batch size: 51, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:42:22,013 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97480.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:42:47,722 INFO [finetune.py:976] (2/7) Epoch 18, batch 150, loss[loss=0.17, simple_loss=0.2197, pruned_loss=0.06012, over 4347.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2393, pruned_loss=0.05101, over 506877.77 frames. ], batch size: 65, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:42:48,445 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7165, 1.6482, 1.6205, 1.3766, 1.7954, 1.5028, 2.2211, 1.4593], + device='cuda:2'), covar=tensor([0.3502, 0.1814, 0.5083, 0.2702, 0.1644, 0.2247, 0.1552, 0.4556], + device='cuda:2'), in_proj_covar=tensor([0.0332, 0.0340, 0.0420, 0.0349, 0.0376, 0.0373, 0.0366, 0.0412], + device='cuda:2'), out_proj_covar=tensor([9.9171e-05, 1.0229e-04, 1.2790e-04, 1.0555e-04, 1.1242e-04, 1.1192e-04, + 1.0779e-04, 1.2514e-04], device='cuda:2') +2023-04-27 10:43:08,512 INFO [optim.py:369] (2/7) 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 10:43:09,274 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8797, 1.4346, 1.4542, 1.6777, 2.0749, 1.6824, 1.3895, 1.3616], + device='cuda:2'), covar=tensor([0.1406, 0.1502, 0.1922, 0.1197, 0.0793, 0.1518, 0.2270, 0.2144], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0311, 0.0350, 0.0286, 0.0327, 0.0308, 0.0298, 0.0366], + device='cuda:2'), out_proj_covar=tensor([6.3389e-05, 6.4776e-05, 7.4442e-05, 5.8153e-05, 6.7986e-05, 6.4763e-05, + 6.2758e-05, 7.8089e-05], device='cuda:2') +2023-04-27 10:43:09,851 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:43:14,689 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9514, 1.4309, 1.5301, 1.7251, 2.1677, 1.6853, 1.4518, 1.4114], + device='cuda:2'), covar=tensor([0.1529, 0.1599, 0.1904, 0.1222, 0.0690, 0.1590, 0.1968, 0.2242], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0310, 0.0349, 0.0286, 0.0327, 0.0307, 0.0297, 0.0365], + device='cuda:2'), out_proj_covar=tensor([6.3280e-05, 6.4646e-05, 7.4276e-05, 5.8043e-05, 6.7898e-05, 6.4648e-05, + 6.2625e-05, 7.7962e-05], device='cuda:2') +2023-04-27 10:43:19,853 INFO [finetune.py:976] (2/7) Epoch 18, batch 200, loss[loss=0.1276, simple_loss=0.196, pruned_loss=0.0296, over 4888.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2403, pruned_loss=0.05174, over 607340.92 frames. ], batch size: 32, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:43:55,476 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:43:55,521 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:43:58,889 INFO [finetune.py:976] (2/7) Epoch 18, batch 250, loss[loss=0.218, simple_loss=0.2831, pruned_loss=0.07643, over 4859.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2448, pruned_loss=0.05321, over 686160.22 frames. ], batch size: 31, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:44:42,422 INFO [optim.py:369] (2/7) 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 10:44:49,684 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 300, loss[loss=0.2065, simple_loss=0.2727, pruned_loss=0.07014, over 4742.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2474, pruned_loss=0.05283, over 745113.20 frames. ], batch size: 27, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:45:09,979 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97698.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 10:45:47,048 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:46:05,896 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0613, 1.8406, 2.0778, 2.4481, 2.4204, 2.0037, 1.6866, 2.1900], + device='cuda:2'), covar=tensor([0.0936, 0.1105, 0.0684, 0.0651, 0.0621, 0.0904, 0.0861, 0.0634], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0198, 0.0178, 0.0168, 0.0173, 0.0176, 0.0149, 0.0176], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:46:07,663 INFO [finetune.py:976] (2/7) Epoch 18, batch 350, loss[loss=0.1953, simple_loss=0.26, pruned_loss=0.06529, over 4889.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2496, pruned_loss=0.05361, over 792952.91 frames. ], batch size: 35, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:46:18,776 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97730.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:46:28,131 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97735.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:46:51,633 INFO [optim.py:369] (2/7) 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 10:47:08,346 INFO [finetune.py:976] (2/7) Epoch 18, batch 400, loss[loss=0.1721, simple_loss=0.2367, pruned_loss=0.05374, over 4797.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2501, pruned_loss=0.05302, over 830035.89 frames. ], batch size: 25, lr: 3.36e-03, grad_scale: 64.0 +2023-04-27 10:47:17,499 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8081, 1.4854, 1.9858, 2.3648, 1.8774, 1.8361, 1.8896, 1.8163], + device='cuda:2'), covar=tensor([0.5292, 0.7369, 0.7543, 0.6449, 0.7063, 0.9041, 0.9632, 0.9237], + device='cuda:2'), in_proj_covar=tensor([0.0421, 0.0407, 0.0498, 0.0505, 0.0450, 0.0477, 0.0484, 0.0488], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:47:42,004 INFO [finetune.py:976] (2/7) Epoch 18, batch 450, loss[loss=0.1772, simple_loss=0.247, pruned_loss=0.0537, over 4753.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2471, pruned_loss=0.05216, over 857468.03 frames. ], batch size: 27, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:47:44,581 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0889, 1.8384, 2.0984, 2.4210, 2.4290, 2.0767, 1.6358, 2.2678], + device='cuda:2'), covar=tensor([0.0852, 0.1094, 0.0646, 0.0613, 0.0617, 0.0816, 0.0833, 0.0563], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0198, 0.0178, 0.0169, 0.0174, 0.0178, 0.0150, 0.0176], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:48:05,035 INFO [optim.py:369] (2/7) 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 10:48:15,396 INFO [finetune.py:976] (2/7) Epoch 18, batch 500, loss[loss=0.1533, simple_loss=0.228, pruned_loss=0.03936, over 4806.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.246, pruned_loss=0.05223, over 879729.23 frames. ], batch size: 51, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:48:16,099 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3978, 1.2870, 4.1509, 3.8287, 3.6815, 3.9121, 3.9545, 3.5926], + device='cuda:2'), covar=tensor([0.7368, 0.6177, 0.1344, 0.2352, 0.1218, 0.1754, 0.1259, 0.1756], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0304, 0.0401, 0.0405, 0.0348, 0.0404, 0.0310, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:48:26,732 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97888.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:48:28,101 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-27 10:48:33,164 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-04-27 10:48:42,840 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:48:48,923 INFO [finetune.py:976] (2/7) Epoch 18, batch 550, loss[loss=0.1888, simple_loss=0.2533, pruned_loss=0.06221, over 4230.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2434, pruned_loss=0.05194, over 896640.89 frames. ], batch size: 66, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:48:55,893 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 10:48:59,950 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8919, 1.7295, 1.9158, 2.2006, 2.2770, 1.8709, 1.5136, 2.0948], + device='cuda:2'), covar=tensor([0.0867, 0.1121, 0.0724, 0.0683, 0.0673, 0.0894, 0.0868, 0.0593], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0200, 0.0180, 0.0170, 0.0176, 0.0179, 0.0151, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:49:04,062 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6470, 1.5813, 4.3770, 4.1025, 3.8403, 4.2230, 4.1153, 3.8295], + device='cuda:2'), covar=tensor([0.6864, 0.5595, 0.1104, 0.1774, 0.1197, 0.2354, 0.1294, 0.1588], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0305, 0.0402, 0.0406, 0.0348, 0.0405, 0.0311, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:49:08,024 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8808, 2.1535, 1.8102, 1.5730, 1.3576, 1.3937, 1.9126, 1.3145], + device='cuda:2'), covar=tensor([0.1756, 0.1582, 0.1650, 0.1984, 0.2648, 0.2132, 0.1143, 0.2242], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0213, 0.0168, 0.0205, 0.0199, 0.0184, 0.0156, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 10:49:09,152 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97949.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:12,035 INFO [optim.py:369] (2/7) 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 10:49:18,169 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97964.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:22,478 INFO [finetune.py:976] (2/7) Epoch 18, batch 600, loss[loss=0.2327, simple_loss=0.3134, pruned_loss=0.07601, over 4157.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2443, pruned_loss=0.05212, over 908407.87 frames. ], batch size: 65, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:49:23,144 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3324, 3.0971, 0.9467, 1.8077, 1.6206, 2.3840, 1.8390, 1.1293], + device='cuda:2'), covar=tensor([0.1419, 0.0892, 0.1876, 0.1203, 0.1147, 0.0912, 0.1370, 0.1874], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0242, 0.0137, 0.0120, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:49:35,803 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6187, 1.6638, 1.8002, 1.3571, 1.8410, 1.5717, 2.3096, 1.6029], + device='cuda:2'), covar=tensor([0.3889, 0.1852, 0.4636, 0.2729, 0.1460, 0.2153, 0.1507, 0.4303], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0346, 0.0427, 0.0356, 0.0381, 0.0379, 0.0371, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:49:41,022 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:46,924 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98003.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:49:57,856 INFO [finetune.py:976] (2/7) Epoch 18, batch 650, loss[loss=0.1873, simple_loss=0.2412, pruned_loss=0.0667, over 4783.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2474, pruned_loss=0.05347, over 917314.55 frames. ], batch size: 26, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:50:03,542 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98030.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 10:50:03,574 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98030.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:50:14,374 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98046.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:50:18,391 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 18, batch 700, loss[loss=0.182, simple_loss=0.2609, pruned_loss=0.05153, over 4827.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2498, pruned_loss=0.05346, over 926134.85 frames. ], batch size: 25, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:50:58,804 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:51:33,407 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98106.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:51:45,513 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8075, 2.7815, 2.1663, 3.2610, 2.7919, 2.7987, 1.2389, 2.7137], + device='cuda:2'), covar=tensor([0.2032, 0.1688, 0.3611, 0.3121, 0.2883, 0.2210, 0.5512, 0.3083], + device='cuda:2'), in_proj_covar=tensor([0.0239, 0.0210, 0.0245, 0.0301, 0.0294, 0.0246, 0.0268, 0.0267], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:51:53,719 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8231, 1.4681, 1.3818, 1.5964, 2.0310, 1.6789, 1.3853, 1.2883], + device='cuda:2'), covar=tensor([0.1570, 0.1471, 0.2003, 0.1461, 0.0896, 0.1394, 0.1929, 0.2260], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0312, 0.0351, 0.0287, 0.0329, 0.0310, 0.0300, 0.0366], + device='cuda:2'), out_proj_covar=tensor([6.3701e-05, 6.5053e-05, 7.4740e-05, 5.8196e-05, 6.8415e-05, 6.5210e-05, + 6.3188e-05, 7.8106e-05], device='cuda:2') +2023-04-27 10:51:54,809 INFO [finetune.py:976] (2/7) Epoch 18, batch 750, loss[loss=0.1656, simple_loss=0.2451, pruned_loss=0.04306, over 4229.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2502, pruned_loss=0.05372, over 933071.40 frames. ], batch size: 65, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:52:31,990 INFO [optim.py:369] (2/7) 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 10:52:42,186 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98167.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:52:44,509 INFO [finetune.py:976] (2/7) Epoch 18, batch 800, loss[loss=0.1752, simple_loss=0.233, pruned_loss=0.05868, over 4933.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2482, pruned_loss=0.05276, over 937965.46 frames. ], batch size: 38, lr: 3.36e-03, grad_scale: 32.0 +2023-04-27 10:53:10,399 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 850, loss[loss=0.1625, simple_loss=0.2442, pruned_loss=0.04034, over 4869.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2458, pruned_loss=0.05185, over 941238.70 frames. ], batch size: 31, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:53:32,031 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98244.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:53:38,452 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98259.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:53:46,219 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7911, 2.2094, 1.6885, 1.5590, 1.3100, 1.3353, 1.8081, 1.2458], + device='cuda:2'), covar=tensor([0.1508, 0.1300, 0.1403, 0.1731, 0.2113, 0.1790, 0.0876, 0.1917], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0214, 0.0169, 0.0206, 0.0201, 0.0185, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 10:53:51,234 INFO [finetune.py:976] (2/7) Epoch 18, batch 900, loss[loss=0.1497, simple_loss=0.2204, pruned_loss=0.03953, over 4810.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2434, pruned_loss=0.05115, over 945042.40 frames. ], batch size: 51, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:54:13,178 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0644, 2.4158, 2.1600, 2.3801, 1.8355, 2.1541, 2.2397, 1.7274], + device='cuda:2'), covar=tensor([0.1639, 0.0852, 0.0652, 0.0929, 0.2605, 0.0941, 0.1491, 0.2056], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0309, 0.0222, 0.0283, 0.0317, 0.0263, 0.0253, 0.0269], + device='cuda:2'), out_proj_covar=tensor([1.1626e-04, 1.2292e-04, 8.8366e-05, 1.1236e-04, 1.2878e-04, 1.0439e-04, + 1.0233e-04, 1.0694e-04], device='cuda:2') +2023-04-27 10:54:24,149 INFO [finetune.py:976] (2/7) Epoch 18, batch 950, loss[loss=0.181, simple_loss=0.2472, pruned_loss=0.05739, over 4755.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2416, pruned_loss=0.05079, over 948047.20 frames. ], batch size: 26, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:54:30,321 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98330.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 10:54:36,452 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4352, 1.3530, 1.7766, 1.7729, 1.3558, 1.2121, 1.5096, 0.9062], + device='cuda:2'), covar=tensor([0.0607, 0.0655, 0.0406, 0.0695, 0.0840, 0.1035, 0.0602, 0.0712], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0069, 0.0067, 0.0067, 0.0075, 0.0095, 0.0073, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:54:44,944 INFO [optim.py:369] (2/7) 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 10:54:57,897 INFO [finetune.py:976] (2/7) Epoch 18, batch 1000, loss[loss=0.1648, simple_loss=0.2483, pruned_loss=0.04064, over 4776.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2435, pruned_loss=0.0508, over 950678.27 frames. ], batch size: 29, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:55:02,792 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98378.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 10:55:22,436 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98410.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:55:29,960 INFO [finetune.py:976] (2/7) Epoch 18, batch 1050, loss[loss=0.1611, simple_loss=0.2451, pruned_loss=0.03856, over 4856.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2465, pruned_loss=0.05122, over 953417.24 frames. ], batch size: 44, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:55:52,030 INFO [optim.py:369] (2/7) 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 10:55:56,965 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:56:08,982 INFO [finetune.py:976] (2/7) Epoch 18, batch 1100, loss[loss=0.2188, simple_loss=0.283, pruned_loss=0.07725, over 4809.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.25, pruned_loss=0.05306, over 954314.50 frames. ], batch size: 39, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:56:09,114 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:56:43,236 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9974, 0.9672, 1.1541, 1.1198, 1.0113, 0.9006, 0.9650, 0.5238], + device='cuda:2'), covar=tensor([0.0581, 0.0488, 0.0452, 0.0488, 0.0733, 0.1117, 0.0463, 0.0700], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0075, 0.0095, 0.0074, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 10:56:54,545 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7658, 1.1830, 1.7765, 2.2810, 1.8674, 1.7221, 1.7540, 1.7379], + device='cuda:2'), covar=tensor([0.4768, 0.6436, 0.6501, 0.5766, 0.5924, 0.8011, 0.8551, 0.8707], + device='cuda:2'), in_proj_covar=tensor([0.0417, 0.0404, 0.0493, 0.0499, 0.0447, 0.0474, 0.0480, 0.0484], + device='cuda:2'), out_proj_covar=tensor([1.0042e-04, 9.9326e-05, 1.1085e-04, 1.1927e-04, 1.0716e-04, 1.1388e-04, + 1.1396e-04, 1.1425e-04], device='cuda:2') +2023-04-27 10:57:13,273 INFO [finetune.py:976] (2/7) Epoch 18, batch 1150, loss[loss=0.1445, simple_loss=0.2178, pruned_loss=0.03553, over 4769.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2498, pruned_loss=0.05307, over 952010.75 frames. ], batch size: 27, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:57:45,475 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98543.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:57:46,074 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98544.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:57:57,934 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 1200, loss[loss=0.1386, simple_loss=0.2084, pruned_loss=0.03438, over 4761.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.248, pruned_loss=0.05281, over 952164.39 frames. ], batch size: 26, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:58:15,587 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2038, 1.4148, 1.7743, 1.8980, 1.8043, 1.8690, 1.8473, 1.8417], + device='cuda:2'), covar=tensor([0.4430, 0.5169, 0.4152, 0.4427, 0.5198, 0.6902, 0.4924, 0.4371], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0372, 0.0319, 0.0333, 0.0344, 0.0395, 0.0355, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 10:58:28,911 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98592.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:58:36,400 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98604.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 10:58:46,684 INFO [finetune.py:976] (2/7) Epoch 18, batch 1250, loss[loss=0.1995, simple_loss=0.2514, pruned_loss=0.0738, over 4909.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2461, pruned_loss=0.05238, over 952206.94 frames. ], batch size: 36, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:59:10,194 INFO [optim.py:369] (2/7) 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 10:59:20,476 INFO [finetune.py:976] (2/7) Epoch 18, batch 1300, loss[loss=0.1321, simple_loss=0.2074, pruned_loss=0.02842, over 4754.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2431, pruned_loss=0.05175, over 952793.93 frames. ], batch size: 28, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 10:59:38,653 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7238, 1.7619, 1.8297, 1.3221, 1.8584, 1.6179, 2.3215, 1.5970], + device='cuda:2'), covar=tensor([0.3537, 0.1753, 0.4349, 0.2668, 0.1502, 0.2197, 0.1457, 0.4241], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0342, 0.0422, 0.0351, 0.0377, 0.0376, 0.0367, 0.0414], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 10:59:53,697 INFO [finetune.py:976] (2/7) Epoch 18, batch 1350, loss[loss=0.1564, simple_loss=0.2166, pruned_loss=0.04805, over 4216.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2436, pruned_loss=0.0523, over 952448.34 frames. ], batch size: 18, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:00:17,035 INFO [optim.py:369] (2/7) 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 11:00:22,053 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98762.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:00:24,433 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:00:24,459 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3660, 1.8015, 1.6693, 2.0367, 2.0118, 2.1180, 1.6955, 3.6478], + device='cuda:2'), covar=tensor([0.0520, 0.0656, 0.0657, 0.0954, 0.0518, 0.0637, 0.0661, 0.0162], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0044, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 11:00:27,464 INFO [finetune.py:976] (2/7) Epoch 18, batch 1400, loss[loss=0.1723, simple_loss=0.2303, pruned_loss=0.05711, over 4745.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2465, pruned_loss=0.0527, over 953720.18 frames. ], batch size: 23, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:00:28,878 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 11:00:47,555 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9779, 2.3903, 2.0383, 2.3833, 1.6788, 2.1321, 1.9808, 1.5626], + device='cuda:2'), covar=tensor([0.1968, 0.1349, 0.0898, 0.1177, 0.3563, 0.1155, 0.2163, 0.2756], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0307, 0.0221, 0.0281, 0.0315, 0.0261, 0.0251, 0.0268], + device='cuda:2'), out_proj_covar=tensor([1.1537e-04, 1.2227e-04, 8.7837e-05, 1.1143e-04, 1.2796e-04, 1.0380e-04, + 1.0172e-04, 1.0636e-04], device='cuda:2') +2023-04-27 11:00:48,301 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5387, 1.4828, 1.8470, 1.9075, 1.4541, 1.2972, 1.6462, 0.9555], + device='cuda:2'), covar=tensor([0.0636, 0.0684, 0.0392, 0.0609, 0.0735, 0.1202, 0.0558, 0.0811], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0069, 0.0068, 0.0068, 0.0075, 0.0096, 0.0074, 0.0068], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:00:54,324 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98810.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:00:56,164 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9061, 1.8614, 2.2253, 2.3517, 1.7917, 1.5792, 1.9654, 1.1347], + device='cuda:2'), covar=tensor([0.0595, 0.0712, 0.0389, 0.0656, 0.0695, 0.1035, 0.0640, 0.0817], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0068, 0.0075, 0.0096, 0.0074, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:01:00,975 INFO [finetune.py:976] (2/7) Epoch 18, batch 1450, loss[loss=0.2254, simple_loss=0.2939, pruned_loss=0.07848, over 4203.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2482, pruned_loss=0.05285, over 953646.54 frames. ], batch size: 65, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:01:07,759 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 11:01:24,436 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 1500, loss[loss=0.1676, simple_loss=0.2485, pruned_loss=0.04331, over 4889.00 frames. ], tot_loss[loss=0.178, simple_loss=0.249, pruned_loss=0.05348, over 953151.47 frames. ], batch size: 35, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:01:52,798 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98890.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:02:02,779 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98895.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:02:11,048 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98899.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:02:35,628 INFO [finetune.py:976] (2/7) Epoch 18, batch 1550, loss[loss=0.1738, simple_loss=0.2601, pruned_loss=0.04377, over 4897.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2487, pruned_loss=0.0529, over 952446.25 frames. ], batch size: 36, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:03:17,463 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 11:03:25,939 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 1600, loss[loss=0.1572, simple_loss=0.2364, pruned_loss=0.03898, over 4869.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2461, pruned_loss=0.0518, over 953364.48 frames. ], batch size: 31, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:03:41,831 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98971.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:03:59,205 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 11:04:14,630 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9377, 1.1793, 1.6660, 1.7978, 1.7375, 1.7884, 1.6724, 1.6476], + device='cuda:2'), covar=tensor([0.4289, 0.5372, 0.4374, 0.4510, 0.5378, 0.7362, 0.5173, 0.4673], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0372, 0.0319, 0.0332, 0.0342, 0.0393, 0.0354, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 11:04:15,772 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3608, 3.2510, 0.8058, 1.6941, 1.6076, 2.2972, 1.8551, 1.1144], + device='cuda:2'), covar=tensor([0.1768, 0.1361, 0.2545, 0.1648, 0.1341, 0.1317, 0.1543, 0.2039], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0243, 0.0137, 0.0121, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:04:20,540 INFO [finetune.py:976] (2/7) Epoch 18, batch 1650, loss[loss=0.1577, simple_loss=0.239, pruned_loss=0.03819, over 4827.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2434, pruned_loss=0.05107, over 951664.76 frames. ], batch size: 40, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:04:27,354 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:04:40,860 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99051.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:04:42,973 INFO [optim.py:369] (2/7) 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 11:04:50,875 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99066.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:04:53,814 INFO [finetune.py:976] (2/7) Epoch 18, batch 1700, loss[loss=0.1645, simple_loss=0.2401, pruned_loss=0.04443, over 4767.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2422, pruned_loss=0.05138, over 952195.10 frames. ], batch size: 26, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:05:21,695 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99112.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:05:22,859 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99114.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:05:27,557 INFO [finetune.py:976] (2/7) Epoch 18, batch 1750, loss[loss=0.1742, simple_loss=0.24, pruned_loss=0.05423, over 4797.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.244, pruned_loss=0.05245, over 952413.90 frames. ], batch size: 25, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:05:50,007 INFO [optim.py:369] (2/7) 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 11:05:59,198 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-27 11:06:01,215 INFO [finetune.py:976] (2/7) Epoch 18, batch 1800, loss[loss=0.1886, simple_loss=0.2603, pruned_loss=0.05849, over 4858.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2481, pruned_loss=0.05345, over 952263.75 frames. ], batch size: 44, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:06:07,532 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 11:06:18,772 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99199.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:22,390 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1329, 1.7412, 2.0030, 2.2406, 1.9633, 1.6175, 1.2128, 1.7405], + device='cuda:2'), covar=tensor([0.3467, 0.3182, 0.1804, 0.2053, 0.2547, 0.2728, 0.4009, 0.1853], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0246, 0.0227, 0.0315, 0.0218, 0.0231, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 11:06:34,181 INFO [finetune.py:976] (2/7) Epoch 18, batch 1850, loss[loss=0.1973, simple_loss=0.2667, pruned_loss=0.06394, over 4224.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.251, pruned_loss=0.05477, over 953503.99 frames. ], batch size: 66, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:06:39,641 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:49,923 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99247.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:53,449 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99251.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:06:55,190 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5700, 2.7537, 2.3328, 2.4732, 2.8140, 2.2797, 3.7074, 1.9964], + device='cuda:2'), covar=tensor([0.3906, 0.1982, 0.3905, 0.3018, 0.1963, 0.2691, 0.1295, 0.4141], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0348, 0.0429, 0.0355, 0.0383, 0.0381, 0.0372, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:06:55,632 INFO [optim.py:369] (2/7) 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 11:07:06,160 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99268.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:07:07,876 INFO [finetune.py:976] (2/7) Epoch 18, batch 1900, loss[loss=0.1883, simple_loss=0.2537, pruned_loss=0.0614, over 4724.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2517, pruned_loss=0.05425, over 954040.48 frames. ], batch size: 59, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:07:20,136 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 1950, loss[loss=0.1526, simple_loss=0.228, pruned_loss=0.03856, over 4826.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2493, pruned_loss=0.05284, over 955141.69 frames. ], batch size: 33, lr: 3.35e-03, grad_scale: 32.0 +2023-04-27 11:08:05,849 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99324.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:08:06,407 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0176, 1.4112, 4.6464, 4.3133, 4.0611, 4.2745, 4.1244, 4.1455], + device='cuda:2'), covar=tensor([0.6486, 0.5539, 0.1062, 0.1745, 0.1079, 0.1547, 0.2286, 0.1369], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0306, 0.0402, 0.0405, 0.0349, 0.0403, 0.0310, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:08:07,589 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99327.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:08:09,392 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:08:30,193 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 2000, loss[loss=0.148, simple_loss=0.2123, pruned_loss=0.04181, over 4832.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2461, pruned_loss=0.05223, over 953832.42 frames. ], batch size: 49, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:09:12,995 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:09:44,014 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99407.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:09:45,196 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7428, 1.0086, 1.6614, 2.2525, 1.8114, 1.6272, 1.6331, 1.6477], + device='cuda:2'), covar=tensor([0.4539, 0.6389, 0.6183, 0.5565, 0.5987, 0.7978, 0.7371, 0.7942], + device='cuda:2'), in_proj_covar=tensor([0.0423, 0.0407, 0.0498, 0.0502, 0.0451, 0.0477, 0.0482, 0.0488], + device='cuda:2'), out_proj_covar=tensor([1.0169e-04, 9.9921e-05, 1.1199e-04, 1.1978e-04, 1.0813e-04, 1.1463e-04, + 1.1452e-04, 1.1512e-04], device='cuda:2') +2023-04-27 11:09:54,354 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 2050, loss[loss=0.1684, simple_loss=0.2342, pruned_loss=0.05132, over 4864.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2434, pruned_loss=0.05157, over 954559.00 frames. ], batch size: 44, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:10:00,962 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7309, 4.7423, 1.2849, 2.7664, 2.9558, 3.4593, 3.0628, 1.4506], + device='cuda:2'), covar=tensor([0.1022, 0.0904, 0.1999, 0.1118, 0.0825, 0.0889, 0.1182, 0.1726], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0244, 0.0138, 0.0121, 0.0133, 0.0154, 0.0119, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:10:13,806 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 11:10:15,959 INFO [optim.py:369] (2/7) 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 11:10:26,027 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99468.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:10:28,723 INFO [finetune.py:976] (2/7) Epoch 18, batch 2100, loss[loss=0.1549, simple_loss=0.235, pruned_loss=0.0374, over 4777.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2425, pruned_loss=0.05101, over 956043.39 frames. ], batch size: 28, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:10:35,448 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99481.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:10:43,931 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2682, 2.9385, 2.5681, 2.4033, 1.6685, 1.7514, 2.6692, 1.7354], + device='cuda:2'), covar=tensor([0.1573, 0.1381, 0.1220, 0.1457, 0.2257, 0.1766, 0.0844, 0.1897], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0213, 0.0168, 0.0205, 0.0201, 0.0185, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 11:10:57,623 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4799, 1.3295, 0.5524, 1.1835, 1.4153, 1.3682, 1.2840, 1.2702], + device='cuda:2'), covar=tensor([0.0501, 0.0377, 0.0390, 0.0550, 0.0287, 0.0488, 0.0495, 0.0569], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0044, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 11:11:02,725 INFO [finetune.py:976] (2/7) Epoch 18, batch 2150, loss[loss=0.1879, simple_loss=0.2753, pruned_loss=0.05022, over 4839.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2474, pruned_loss=0.05287, over 957448.02 frames. ], batch size: 47, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:11:08,207 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99529.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:11:11,750 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4802, 1.4321, 1.4201, 1.0633, 1.4487, 1.2574, 1.7429, 1.3375], + device='cuda:2'), covar=tensor([0.3514, 0.1790, 0.5266, 0.2641, 0.1574, 0.2096, 0.1802, 0.4817], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0348, 0.0428, 0.0355, 0.0383, 0.0381, 0.0371, 0.0421], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:11:18,932 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99546.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:21,995 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 11:11:32,855 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99568.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:35,034 INFO [finetune.py:976] (2/7) Epoch 18, batch 2200, loss[loss=0.1832, simple_loss=0.2642, pruned_loss=0.05107, over 4801.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.249, pruned_loss=0.0532, over 957715.87 frames. ], batch size: 40, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:11:45,031 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99585.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:47,524 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99589.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:51,022 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99599.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:11:54,095 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7578, 2.8637, 2.3314, 2.5323, 2.9117, 2.4931, 3.7455, 2.1425], + device='cuda:2'), covar=tensor([0.3580, 0.1881, 0.3546, 0.3002, 0.1595, 0.2374, 0.1445, 0.3641], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0345, 0.0424, 0.0352, 0.0380, 0.0378, 0.0369, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:12:08,620 INFO [finetune.py:976] (2/7) Epoch 18, batch 2250, loss[loss=0.1511, simple_loss=0.2256, pruned_loss=0.03833, over 4772.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2489, pruned_loss=0.05273, over 956384.90 frames. ], batch size: 29, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:12:10,997 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99624.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:13,831 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99627.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:15,110 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:31,119 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 2300, loss[loss=0.1922, simple_loss=0.2561, pruned_loss=0.06415, over 4797.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2491, pruned_loss=0.0524, over 955733.42 frames. ], batch size: 25, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:12:45,388 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99675.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:49,372 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:12:49,398 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:13:12,575 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:13:21,174 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-04-27 11:13:26,439 INFO [finetune.py:976] (2/7) Epoch 18, batch 2350, loss[loss=0.1695, simple_loss=0.2477, pruned_loss=0.04565, over 4829.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2478, pruned_loss=0.05242, over 955553.61 frames. ], batch size: 41, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:13:33,096 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 11:13:54,522 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2490, 2.5385, 0.8159, 1.4562, 1.5856, 1.8716, 1.7142, 0.8467], + device='cuda:2'), covar=tensor([0.1570, 0.1275, 0.1983, 0.1500, 0.1205, 0.1070, 0.1602, 0.1756], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0243, 0.0137, 0.0121, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:13:56,972 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99741.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:14:08,177 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 11:14:16,629 INFO [optim.py:369] (2/7) 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 11:14:17,851 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99755.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:14:33,765 INFO [finetune.py:976] (2/7) Epoch 18, batch 2400, loss[loss=0.1628, simple_loss=0.2327, pruned_loss=0.04651, over 4914.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2451, pruned_loss=0.05199, over 955675.17 frames. ], batch size: 37, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:14:36,882 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 2450, loss[loss=0.1808, simple_loss=0.2458, pruned_loss=0.05788, over 4819.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2425, pruned_loss=0.05116, over 956863.34 frames. ], batch size: 39, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:15:37,742 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99824.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:15:46,845 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8957, 2.8375, 2.1721, 3.3379, 2.8833, 2.9255, 1.1455, 2.8523], + device='cuda:2'), covar=tensor([0.1982, 0.1573, 0.3033, 0.2613, 0.3351, 0.2089, 0.5825, 0.2713], + device='cuda:2'), in_proj_covar=tensor([0.0248, 0.0217, 0.0253, 0.0308, 0.0301, 0.0252, 0.0276, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 11:16:08,395 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 2500, loss[loss=0.1758, simple_loss=0.2483, pruned_loss=0.05163, over 4851.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2454, pruned_loss=0.05283, over 955940.41 frames. ], batch size: 47, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:16:28,671 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99885.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:16:36,190 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7261, 1.0972, 4.0095, 3.4857, 3.4756, 3.7132, 3.6136, 3.3888], + device='cuda:2'), covar=tensor([0.8649, 0.8838, 0.1779, 0.3507, 0.2631, 0.3408, 0.4278, 0.3054], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0310, 0.0407, 0.0410, 0.0352, 0.0407, 0.0314, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:16:52,953 INFO [finetune.py:976] (2/7) Epoch 18, batch 2550, loss[loss=0.2089, simple_loss=0.2729, pruned_loss=0.07251, over 4913.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.25, pruned_loss=0.05489, over 955587.50 frames. ], batch size: 37, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:16:54,879 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:16:54,896 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:10,138 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:16,096 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 2600, loss[loss=0.1687, simple_loss=0.2517, pruned_loss=0.0428, over 4909.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.25, pruned_loss=0.05439, over 952824.11 frames. ], batch size: 37, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:17:27,556 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:32,461 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:17:52,668 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100007.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:18:01,563 INFO [finetune.py:976] (2/7) Epoch 18, batch 2650, loss[loss=0.1488, simple_loss=0.2351, pruned_loss=0.03126, over 4812.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2511, pruned_loss=0.0542, over 952278.83 frames. ], batch size: 39, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:18:05,844 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100028.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:18:10,675 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:18:24,404 INFO [optim.py:369] (2/7) 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 11:18:33,071 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100068.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:18:35,203 INFO [finetune.py:976] (2/7) Epoch 18, batch 2700, loss[loss=0.1892, simple_loss=0.2442, pruned_loss=0.06712, over 4818.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2497, pruned_loss=0.05354, over 952429.72 frames. ], batch size: 38, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:18:38,370 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100076.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:19:37,243 INFO [finetune.py:976] (2/7) Epoch 18, batch 2750, loss[loss=0.1644, simple_loss=0.2364, pruned_loss=0.04624, over 4856.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2462, pruned_loss=0.05245, over 954904.84 frames. ], batch size: 31, lr: 3.34e-03, grad_scale: 64.0 +2023-04-27 11:19:39,079 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:19:39,123 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100124.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:20:11,927 INFO [optim.py:369] (2/7) 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 11:20:28,162 INFO [finetune.py:976] (2/7) Epoch 18, batch 2800, loss[loss=0.1476, simple_loss=0.2194, pruned_loss=0.03784, over 4897.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2436, pruned_loss=0.05199, over 955138.66 frames. ], batch size: 32, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:20:34,239 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100172.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:21:07,288 INFO [finetune.py:976] (2/7) Epoch 18, batch 2850, loss[loss=0.1731, simple_loss=0.239, pruned_loss=0.05358, over 4766.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2418, pruned_loss=0.05132, over 956681.68 frames. ], batch size: 27, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:21:09,225 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100224.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:21:29,444 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100238.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:21:39,941 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 11:22:14,797 INFO [finetune.py:976] (2/7) Epoch 18, batch 2900, loss[loss=0.1525, simple_loss=0.2156, pruned_loss=0.04473, over 4761.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2436, pruned_loss=0.05132, over 954606.84 frames. ], batch size: 26, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:22:15,466 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100272.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:22:21,005 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6710, 1.4755, 1.8423, 1.9801, 1.5046, 1.3901, 1.5769, 0.9630], + device='cuda:2'), covar=tensor([0.0533, 0.0766, 0.0455, 0.0656, 0.0844, 0.1232, 0.0661, 0.0753], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0069, 0.0068, 0.0067, 0.0074, 0.0096, 0.0073, 0.0067], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:22:28,745 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:22:48,676 INFO [finetune.py:976] (2/7) Epoch 18, batch 2950, loss[loss=0.2054, simple_loss=0.2682, pruned_loss=0.07126, over 4740.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2472, pruned_loss=0.05208, over 955234.56 frames. ], batch size: 59, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:22:55,465 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6932, 1.2150, 1.3424, 2.0789, 2.0206, 1.6302, 1.2969, 1.8002], + device='cuda:2'), covar=tensor([0.0937, 0.1854, 0.1259, 0.0608, 0.0634, 0.0989, 0.0949, 0.0701], + device='cuda:2'), in_proj_covar=tensor([0.0190, 0.0203, 0.0184, 0.0173, 0.0180, 0.0183, 0.0153, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:22:57,898 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:23:09,488 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-27 11:23:09,851 INFO [optim.py:369] (2/7) 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 11:23:15,825 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100363.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:23:22,041 INFO [finetune.py:976] (2/7) Epoch 18, batch 3000, loss[loss=0.1887, simple_loss=0.2667, pruned_loss=0.05536, over 4845.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2485, pruned_loss=0.05266, over 955841.34 frames. ], batch size: 44, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:23:22,041 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 11:23:27,736 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7226, 1.5510, 0.7217, 1.3907, 1.5767, 1.5813, 1.4521, 1.5469], + device='cuda:2'), covar=tensor([0.0518, 0.0386, 0.0389, 0.0541, 0.0280, 0.0521, 0.0501, 0.0561], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 11:23:32,627 INFO [finetune.py:1010] (2/7) Epoch 18, validation: loss=0.1524, simple_loss=0.2231, pruned_loss=0.04086, over 2265189.00 frames. +2023-04-27 11:23:32,628 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 11:23:41,191 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100384.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:23:52,517 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100402.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:24:04,563 INFO [finetune.py:976] (2/7) Epoch 18, batch 3050, loss[loss=0.1586, simple_loss=0.2378, pruned_loss=0.03972, over 4812.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.249, pruned_loss=0.05229, over 958011.73 frames. ], batch size: 33, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:24:43,544 INFO [optim.py:369] (2/7) 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 11:24:55,016 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100463.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:24:59,806 INFO [finetune.py:976] (2/7) Epoch 18, batch 3100, loss[loss=0.1697, simple_loss=0.2383, pruned_loss=0.05048, over 4836.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2471, pruned_loss=0.05161, over 957571.68 frames. ], batch size: 47, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:26:02,763 INFO [finetune.py:976] (2/7) Epoch 18, batch 3150, loss[loss=0.1559, simple_loss=0.2342, pruned_loss=0.03879, over 4931.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2437, pruned_loss=0.05069, over 957992.21 frames. ], batch size: 33, lr: 3.34e-03, grad_scale: 32.0 +2023-04-27 11:26:10,647 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 11:26:37,705 INFO [optim.py:369] (2/7) 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 11:26:57,343 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-04-27 11:26:59,499 INFO [finetune.py:976] (2/7) Epoch 18, batch 3200, loss[loss=0.1815, simple_loss=0.2494, pruned_loss=0.05674, over 4807.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2407, pruned_loss=0.0499, over 958527.27 frames. ], batch size: 29, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:27:33,060 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100594.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:28:06,113 INFO [finetune.py:976] (2/7) Epoch 18, batch 3250, loss[loss=0.236, simple_loss=0.2929, pruned_loss=0.08956, over 4933.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2421, pruned_loss=0.05073, over 956235.81 frames. ], batch size: 33, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:28:30,300 INFO [optim.py:369] (2/7) 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 11:28:31,013 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2778, 4.1820, 2.9670, 4.9601, 4.2455, 4.2660, 1.7699, 4.2697], + device='cuda:2'), covar=tensor([0.1562, 0.1155, 0.3773, 0.1002, 0.2815, 0.1511, 0.5921, 0.2174], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0214, 0.0249, 0.0303, 0.0296, 0.0247, 0.0270, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 11:28:35,367 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100663.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:28:37,816 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4295, 1.9028, 2.2893, 2.9690, 2.2840, 1.7900, 1.8896, 2.1538], + device='cuda:2'), covar=tensor([0.2960, 0.2918, 0.1437, 0.2219, 0.2496, 0.2411, 0.3475, 0.1980], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0244, 0.0224, 0.0311, 0.0217, 0.0229, 0.0226, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 11:28:40,637 INFO [finetune.py:976] (2/7) Epoch 18, batch 3300, loss[loss=0.1419, simple_loss=0.2133, pruned_loss=0.03531, over 4771.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2461, pruned_loss=0.05193, over 953830.22 frames. ], batch size: 26, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:29:07,778 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100711.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:29:13,782 INFO [finetune.py:976] (2/7) Epoch 18, batch 3350, loss[loss=0.1941, simple_loss=0.2706, pruned_loss=0.05882, over 4803.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2479, pruned_loss=0.05248, over 952734.43 frames. ], batch size: 45, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:29:19,925 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100730.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:29:37,230 INFO [optim.py:369] (2/7) 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 11:29:39,119 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100758.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:29:47,005 INFO [finetune.py:976] (2/7) Epoch 18, batch 3400, loss[loss=0.2074, simple_loss=0.2841, pruned_loss=0.06537, over 4888.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2499, pruned_loss=0.05396, over 950514.24 frames. ], batch size: 43, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:29:58,685 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0390, 0.9928, 1.2339, 1.1513, 1.0018, 0.9406, 0.9151, 0.4606], + device='cuda:2'), covar=tensor([0.0515, 0.0610, 0.0463, 0.0526, 0.0687, 0.1167, 0.0432, 0.0726], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0069, 0.0067, 0.0067, 0.0074, 0.0095, 0.0073, 0.0066], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:30:00,850 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100791.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 11:30:20,268 INFO [finetune.py:976] (2/7) Epoch 18, batch 3450, loss[loss=0.1736, simple_loss=0.2506, pruned_loss=0.04827, over 4865.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2491, pruned_loss=0.05318, over 951242.12 frames. ], batch size: 34, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:30:20,392 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3624, 2.4486, 2.0434, 2.0891, 2.4276, 2.0301, 3.0606, 1.5445], + device='cuda:2'), covar=tensor([0.2925, 0.1399, 0.3677, 0.2431, 0.1408, 0.2130, 0.0924, 0.4226], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0346, 0.0429, 0.0354, 0.0382, 0.0382, 0.0369, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:30:51,932 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-04-27 11:30:54,161 INFO [optim.py:369] (2/7) 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 11:31:09,717 INFO [finetune.py:976] (2/7) Epoch 18, batch 3500, loss[loss=0.1414, simple_loss=0.2183, pruned_loss=0.03226, over 4681.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2479, pruned_loss=0.05322, over 952861.55 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:31:30,613 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100894.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:31:49,439 INFO [finetune.py:976] (2/7) Epoch 18, batch 3550, loss[loss=0.1421, simple_loss=0.2193, pruned_loss=0.03241, over 4828.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2448, pruned_loss=0.05196, over 953088.13 frames. ], batch size: 30, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:31:51,448 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 11:31:56,221 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7539, 1.2956, 4.7995, 4.5263, 4.1985, 4.5628, 4.2875, 4.2208], + device='cuda:2'), covar=tensor([0.6966, 0.6009, 0.1131, 0.1781, 0.1008, 0.1401, 0.1617, 0.1515], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0305, 0.0401, 0.0406, 0.0348, 0.0401, 0.0310, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:32:02,728 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100942.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:32:23,487 INFO [optim.py:369] (2/7) 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 11:32:39,051 INFO [finetune.py:976] (2/7) Epoch 18, batch 3600, loss[loss=0.1404, simple_loss=0.2116, pruned_loss=0.03454, over 4690.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2417, pruned_loss=0.05122, over 954171.51 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:32:59,330 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2569, 1.7163, 2.1230, 2.4909, 2.0538, 1.6351, 1.3318, 1.7479], + device='cuda:2'), covar=tensor([0.3589, 0.3396, 0.1881, 0.2299, 0.2692, 0.3018, 0.4311, 0.2163], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0245, 0.0224, 0.0312, 0.0217, 0.0230, 0.0227, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 11:33:40,094 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8775, 2.6183, 2.0692, 2.0862, 1.3896, 1.3287, 2.1179, 1.3388], + device='cuda:2'), covar=tensor([0.1692, 0.1523, 0.1408, 0.1681, 0.2290, 0.2024, 0.0939, 0.2125], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0213, 0.0169, 0.0206, 0.0201, 0.0185, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 11:33:50,170 INFO [finetune.py:976] (2/7) Epoch 18, batch 3650, loss[loss=0.2133, simple_loss=0.2775, pruned_loss=0.07457, over 4894.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2449, pruned_loss=0.05255, over 952545.29 frames. ], batch size: 32, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:33:50,923 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1197, 1.3795, 1.3085, 1.7165, 1.5095, 1.5997, 1.3821, 2.4851], + device='cuda:2'), covar=tensor([0.0599, 0.0832, 0.0816, 0.1129, 0.0632, 0.0457, 0.0733, 0.0206], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0037, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 11:34:12,269 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1889, 2.4239, 1.2593, 1.5137, 1.9257, 1.4363, 3.1911, 1.9186], + device='cuda:2'), covar=tensor([0.0576, 0.0766, 0.0715, 0.1062, 0.0444, 0.0870, 0.0272, 0.0506], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 11:34:31,992 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101058.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:34:48,424 INFO [finetune.py:976] (2/7) Epoch 18, batch 3700, loss[loss=0.1654, simple_loss=0.2414, pruned_loss=0.04472, over 4831.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2466, pruned_loss=0.05283, over 947858.76 frames. ], batch size: 49, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:34:57,131 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-04-27 11:34:57,632 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101086.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:35:10,210 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3064, 1.3409, 4.1278, 3.8646, 3.6541, 3.9567, 3.8994, 3.6669], + device='cuda:2'), covar=tensor([0.7157, 0.5943, 0.1161, 0.1829, 0.1115, 0.2043, 0.1548, 0.1500], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0305, 0.0403, 0.0406, 0.0348, 0.0403, 0.0311, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:35:10,798 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101106.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:35:11,913 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101107.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:35:22,255 INFO [finetune.py:976] (2/7) Epoch 18, batch 3750, loss[loss=0.1687, simple_loss=0.2351, pruned_loss=0.05117, over 4883.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.249, pruned_loss=0.05359, over 950764.73 frames. ], batch size: 35, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:35:43,442 INFO [optim.py:369] (2/7) 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 11:35:54,468 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 3800, loss[loss=0.1692, simple_loss=0.2456, pruned_loss=0.04643, over 4749.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2494, pruned_loss=0.05391, over 950089.37 frames. ], batch size: 54, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:36:30,056 INFO [finetune.py:976] (2/7) Epoch 18, batch 3850, loss[loss=0.135, simple_loss=0.2146, pruned_loss=0.02768, over 4897.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2472, pruned_loss=0.0528, over 951269.96 frames. ], batch size: 43, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:36:41,673 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7563, 1.3010, 1.4686, 1.5625, 1.9748, 1.5868, 1.3545, 1.3316], + device='cuda:2'), covar=tensor([0.1563, 0.1491, 0.2229, 0.1631, 0.0863, 0.1624, 0.1896, 0.2135], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0315, 0.0354, 0.0290, 0.0331, 0.0310, 0.0303, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4214e-05, 6.5472e-05, 7.5122e-05, 5.8886e-05, 6.8790e-05, 6.5283e-05, + 6.3702e-05, 7.9442e-05], device='cuda:2') +2023-04-27 11:36:50,613 INFO [optim.py:369] (2/7) 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 11:37:02,693 INFO [finetune.py:976] (2/7) Epoch 18, batch 3900, loss[loss=0.1903, simple_loss=0.26, pruned_loss=0.06031, over 4939.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2449, pruned_loss=0.05222, over 952964.04 frames. ], batch size: 38, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:37:07,075 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 11:37:35,479 INFO [finetune.py:976] (2/7) Epoch 18, batch 3950, loss[loss=0.1918, simple_loss=0.2478, pruned_loss=0.06793, over 4903.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2421, pruned_loss=0.05122, over 953105.11 frames. ], batch size: 32, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:38:09,082 INFO [optim.py:369] (2/7) 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 11:38:19,927 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-04-27 11:38:30,025 INFO [finetune.py:976] (2/7) Epoch 18, batch 4000, loss[loss=0.1902, simple_loss=0.2648, pruned_loss=0.05786, over 4842.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2422, pruned_loss=0.05145, over 955566.95 frames. ], batch size: 47, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:38:49,921 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 11:38:52,340 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101386.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:39:01,083 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 11:39:35,326 INFO [finetune.py:976] (2/7) Epoch 18, batch 4050, loss[loss=0.1595, simple_loss=0.2424, pruned_loss=0.03827, over 4832.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2455, pruned_loss=0.05226, over 957466.68 frames. ], batch size: 30, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:39:55,120 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101434.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:40:00,672 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0958, 2.5349, 1.1059, 1.4984, 2.0137, 1.2532, 3.4666, 1.8073], + device='cuda:2'), covar=tensor([0.0650, 0.0756, 0.0793, 0.1214, 0.0519, 0.1020, 0.0211, 0.0627], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0048, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 11:40:07,832 INFO [optim.py:369] (2/7) 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 11:40:12,810 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101463.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:40:18,124 INFO [finetune.py:976] (2/7) Epoch 18, batch 4100, loss[loss=0.1622, simple_loss=0.2472, pruned_loss=0.03856, over 4812.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2465, pruned_loss=0.05264, over 953914.50 frames. ], batch size: 40, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:40:45,381 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5020, 1.6802, 1.2841, 1.1521, 1.1399, 1.1173, 1.2193, 1.0446], + device='cuda:2'), covar=tensor([0.1934, 0.1266, 0.1821, 0.1862, 0.2527, 0.2218, 0.1188, 0.2161], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0213, 0.0169, 0.0207, 0.0202, 0.0186, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 11:40:51,379 INFO [finetune.py:976] (2/7) Epoch 18, batch 4150, loss[loss=0.1615, simple_loss=0.2419, pruned_loss=0.04059, over 4831.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2491, pruned_loss=0.05389, over 952771.25 frames. ], batch size: 49, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:41:14,399 INFO [optim.py:369] (2/7) 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 11:41:20,111 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.53 vs. limit=5.0 +2023-04-27 11:41:24,218 INFO [finetune.py:976] (2/7) Epoch 18, batch 4200, loss[loss=0.1686, simple_loss=0.2442, pruned_loss=0.04648, over 4814.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2492, pruned_loss=0.05335, over 952796.18 frames. ], batch size: 33, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:41:45,337 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5855, 0.6806, 1.4367, 1.9856, 1.6836, 1.5043, 1.4973, 1.5232], + device='cuda:2'), covar=tensor([0.4314, 0.6563, 0.6230, 0.5851, 0.5688, 0.6802, 0.7214, 0.7875], + device='cuda:2'), in_proj_covar=tensor([0.0425, 0.0408, 0.0500, 0.0504, 0.0454, 0.0478, 0.0486, 0.0491], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:41:58,014 INFO [finetune.py:976] (2/7) Epoch 18, batch 4250, loss[loss=0.1698, simple_loss=0.2235, pruned_loss=0.05807, over 4194.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2473, pruned_loss=0.05278, over 952756.26 frames. ], batch size: 18, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:42:21,983 INFO [optim.py:369] (2/7) 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 11:42:31,623 INFO [finetune.py:976] (2/7) Epoch 18, batch 4300, loss[loss=0.1343, simple_loss=0.2053, pruned_loss=0.0317, over 4801.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2444, pruned_loss=0.05176, over 951833.11 frames. ], batch size: 25, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:42:40,096 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101684.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:43:04,507 INFO [finetune.py:976] (2/7) Epoch 18, batch 4350, loss[loss=0.1317, simple_loss=0.1963, pruned_loss=0.03353, over 4236.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2406, pruned_loss=0.05052, over 952196.21 frames. ], batch size: 18, lr: 3.33e-03, grad_scale: 32.0 +2023-04-27 11:43:08,414 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 11:43:13,815 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8977, 1.5991, 2.0338, 2.3722, 2.0000, 1.8927, 1.9446, 1.8755], + device='cuda:2'), covar=tensor([0.4496, 0.6380, 0.6458, 0.5875, 0.6083, 0.7626, 0.8075, 0.8248], + device='cuda:2'), in_proj_covar=tensor([0.0424, 0.0408, 0.0499, 0.0503, 0.0453, 0.0478, 0.0486, 0.0490], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:43:20,706 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101745.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:43:25,441 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0491, 1.5561, 1.8438, 2.0969, 1.8971, 1.4944, 1.1536, 1.6505], + device='cuda:2'), covar=tensor([0.3208, 0.3188, 0.1779, 0.2292, 0.2376, 0.2738, 0.4469, 0.2061], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0246, 0.0227, 0.0315, 0.0219, 0.0231, 0.0228, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 11:43:32,959 INFO [optim.py:369] (2/7) 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 11:43:43,340 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101763.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:43:48,146 INFO [finetune.py:976] (2/7) Epoch 18, batch 4400, loss[loss=0.2402, simple_loss=0.3043, pruned_loss=0.08809, over 4812.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2428, pruned_loss=0.05119, over 951982.70 frames. ], batch size: 45, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:44:26,417 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7368, 1.2611, 1.8843, 2.2631, 1.8498, 1.7307, 1.7569, 1.7363], + device='cuda:2'), covar=tensor([0.4701, 0.6560, 0.6358, 0.5772, 0.5808, 0.7676, 0.8339, 0.8866], + device='cuda:2'), in_proj_covar=tensor([0.0424, 0.0407, 0.0499, 0.0502, 0.0453, 0.0477, 0.0485, 0.0489], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:44:37,288 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 4450, loss[loss=0.191, simple_loss=0.2621, pruned_loss=0.05999, over 4903.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.247, pruned_loss=0.05286, over 952511.14 frames. ], batch size: 37, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:45:32,216 INFO [optim.py:369] (2/7) 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 11:45:42,461 INFO [finetune.py:976] (2/7) Epoch 18, batch 4500, loss[loss=0.1991, simple_loss=0.2712, pruned_loss=0.06347, over 4895.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2497, pruned_loss=0.05373, over 953927.79 frames. ], batch size: 35, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:45:58,500 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5172, 1.0953, 0.3275, 1.2163, 1.2226, 1.3993, 1.2955, 1.3301], + device='cuda:2'), covar=tensor([0.0534, 0.0411, 0.0427, 0.0587, 0.0305, 0.0522, 0.0527, 0.0598], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 11:46:03,973 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5897, 3.9936, 0.8003, 1.9938, 2.1736, 2.7062, 2.3065, 1.0348], + device='cuda:2'), covar=tensor([0.1440, 0.1027, 0.2148, 0.1320, 0.1102, 0.0991, 0.1501, 0.2059], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0138, 0.0122, 0.0134, 0.0154, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 11:46:15,938 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101920.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:46:16,436 INFO [finetune.py:976] (2/7) Epoch 18, batch 4550, loss[loss=0.1704, simple_loss=0.2436, pruned_loss=0.0486, over 4891.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2501, pruned_loss=0.05372, over 954689.52 frames. ], batch size: 32, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:46:38,530 INFO [optim.py:369] (2/7) 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 11:46:49,877 INFO [finetune.py:976] (2/7) Epoch 18, batch 4600, loss[loss=0.207, simple_loss=0.2751, pruned_loss=0.06948, over 4856.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2499, pruned_loss=0.05346, over 955692.49 frames. ], batch size: 44, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:46:56,094 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101981.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:46:57,270 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8747, 2.2949, 1.1462, 1.5420, 2.3474, 1.7892, 1.6422, 1.8684], + device='cuda:2'), covar=tensor([0.0492, 0.0330, 0.0302, 0.0553, 0.0228, 0.0501, 0.0496, 0.0526], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 11:46:59,696 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:47:24,660 INFO [finetune.py:976] (2/7) Epoch 18, batch 4650, loss[loss=0.1974, simple_loss=0.2611, pruned_loss=0.06684, over 4927.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2465, pruned_loss=0.05287, over 952762.05 frames. ], batch size: 33, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:47:36,310 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:47:41,251 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 18, batch 4700, loss[loss=0.169, simple_loss=0.2316, pruned_loss=0.05318, over 4820.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2437, pruned_loss=0.05204, over 955063.06 frames. ], batch size: 33, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:48:30,319 INFO [finetune.py:976] (2/7) Epoch 18, batch 4750, loss[loss=0.1813, simple_loss=0.2487, pruned_loss=0.05692, over 4904.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2422, pruned_loss=0.05131, over 955490.60 frames. ], batch size: 35, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 11:48:51,944 INFO [optim.py:369] (2/7) 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 11:49:09,304 INFO [finetune.py:976] (2/7) Epoch 18, batch 4800, loss[loss=0.1587, simple_loss=0.2295, pruned_loss=0.04388, over 4814.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2456, pruned_loss=0.05243, over 955173.10 frames. ], batch size: 25, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:50:15,442 INFO [finetune.py:976] (2/7) Epoch 18, batch 4850, loss[loss=0.152, simple_loss=0.2172, pruned_loss=0.04343, over 4246.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2493, pruned_loss=0.05334, over 954919.03 frames. ], batch size: 18, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:50:58,311 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 4900, loss[loss=0.1785, simple_loss=0.2585, pruned_loss=0.04924, over 4736.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2504, pruned_loss=0.05422, over 951319.12 frames. ], batch size: 54, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:51:23,239 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102276.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 11:52:25,540 INFO [finetune.py:976] (2/7) Epoch 18, batch 4950, loss[loss=0.2039, simple_loss=0.2789, pruned_loss=0.06447, over 4883.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2504, pruned_loss=0.05381, over 951583.54 frames. ], batch size: 32, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:52:50,470 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102340.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:52:57,588 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 11:53:13,941 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6894, 3.5682, 2.6427, 4.2214, 3.6792, 3.6382, 1.5664, 3.6337], + device='cuda:2'), covar=tensor([0.1838, 0.1312, 0.3004, 0.1717, 0.3013, 0.1868, 0.5729, 0.2383], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0215, 0.0251, 0.0305, 0.0300, 0.0250, 0.0273, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 11:53:14,474 INFO [finetune.py:976] (2/7) Epoch 18, batch 5000, loss[loss=0.1677, simple_loss=0.2387, pruned_loss=0.04837, over 4803.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2483, pruned_loss=0.05328, over 953566.34 frames. ], batch size: 25, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:53:28,310 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 5050, loss[loss=0.1822, simple_loss=0.2502, pruned_loss=0.05704, over 4899.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2463, pruned_loss=0.05343, over 954613.72 frames. ], batch size: 35, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:54:21,678 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 11:54:23,901 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 5100, loss[loss=0.1599, simple_loss=0.234, pruned_loss=0.04291, over 4767.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2428, pruned_loss=0.05159, over 956572.59 frames. ], batch size: 26, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:55:18,947 INFO [finetune.py:976] (2/7) Epoch 18, batch 5150, loss[loss=0.1688, simple_loss=0.2453, pruned_loss=0.04618, over 4872.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2429, pruned_loss=0.05156, over 957427.45 frames. ], batch size: 34, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:55:42,253 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8699, 1.4308, 1.4654, 1.5522, 2.0086, 1.6187, 1.3200, 1.3841], + device='cuda:2'), covar=tensor([0.1225, 0.1245, 0.1788, 0.1196, 0.0719, 0.1491, 0.1749, 0.1994], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0312, 0.0353, 0.0291, 0.0331, 0.0310, 0.0303, 0.0371], + device='cuda:2'), out_proj_covar=tensor([6.3885e-05, 6.4890e-05, 7.4814e-05, 5.9048e-05, 6.8666e-05, 6.5177e-05, + 6.3788e-05, 7.9124e-05], device='cuda:2') +2023-04-27 11:55:52,831 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 5200, loss[loss=0.1886, simple_loss=0.2613, pruned_loss=0.05798, over 4899.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2476, pruned_loss=0.05383, over 956579.31 frames. ], batch size: 36, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:56:11,163 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102576.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:56:25,410 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.22 vs. limit=5.0 +2023-04-27 11:56:27,499 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0186, 2.6349, 2.0629, 1.9852, 1.5224, 1.4441, 2.2432, 1.4391], + device='cuda:2'), covar=tensor([0.1706, 0.1576, 0.1393, 0.1840, 0.2428, 0.1986, 0.1003, 0.2082], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0214, 0.0169, 0.0206, 0.0201, 0.0186, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 11:56:42,080 INFO [finetune.py:976] (2/7) Epoch 18, batch 5250, loss[loss=0.1675, simple_loss=0.2387, pruned_loss=0.04814, over 4822.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2504, pruned_loss=0.05484, over 957679.23 frames. ], batch size: 30, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:56:42,801 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0634, 2.5640, 1.1843, 1.4499, 2.1635, 1.3649, 3.4292, 1.8625], + device='cuda:2'), covar=tensor([0.0646, 0.0617, 0.0842, 0.1200, 0.0487, 0.0929, 0.0199, 0.0584], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 11:56:43,956 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102624.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 11:56:45,169 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0581, 1.6634, 1.8859, 2.3019, 2.3853, 1.8797, 1.6715, 2.0065], + device='cuda:2'), covar=tensor([0.0778, 0.1115, 0.0683, 0.0539, 0.0516, 0.0820, 0.0755, 0.0580], + device='cuda:2'), in_proj_covar=tensor([0.0189, 0.0201, 0.0182, 0.0171, 0.0177, 0.0181, 0.0150, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 11:56:57,434 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 11:57:11,272 INFO [optim.py:369] (2/7) 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 11:57:26,882 INFO [finetune.py:976] (2/7) Epoch 18, batch 5300, loss[loss=0.194, simple_loss=0.2689, pruned_loss=0.05951, over 4806.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2512, pruned_loss=0.05421, over 957244.01 frames. ], batch size: 41, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:57:55,088 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 18, batch 5350, loss[loss=0.1905, simple_loss=0.2539, pruned_loss=0.06359, over 4721.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2506, pruned_loss=0.05308, over 957220.08 frames. ], batch size: 59, lr: 3.32e-03, grad_scale: 64.0 +2023-04-27 11:58:41,910 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.87 vs. limit=5.0 +2023-04-27 11:58:50,049 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8018, 2.1467, 2.0015, 2.1972, 1.9891, 2.0768, 2.0980, 2.0355], + device='cuda:2'), covar=tensor([0.4211, 0.6303, 0.5319, 0.4760, 0.5980, 0.7641, 0.6315, 0.5635], + device='cuda:2'), in_proj_covar=tensor([0.0331, 0.0372, 0.0319, 0.0331, 0.0343, 0.0392, 0.0355, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 11:59:00,895 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0406, 1.3684, 1.7767, 2.3368, 1.8953, 1.4290, 1.2237, 1.5999], + device='cuda:2'), covar=tensor([0.3114, 0.3794, 0.1831, 0.2301, 0.2735, 0.2657, 0.4582, 0.2328], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0247, 0.0227, 0.0314, 0.0220, 0.0232, 0.0229, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 11:59:22,079 INFO [optim.py:369] (2/7) 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 11:59:37,389 INFO [finetune.py:976] (2/7) Epoch 18, batch 5400, loss[loss=0.1941, simple_loss=0.2706, pruned_loss=0.0588, over 4932.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.248, pruned_loss=0.05255, over 958505.21 frames. ], batch size: 38, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:00:39,491 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5804, 1.5878, 4.3815, 4.1088, 3.8170, 4.1538, 4.0959, 3.8620], + device='cuda:2'), covar=tensor([0.6847, 0.5628, 0.0960, 0.1560, 0.1102, 0.1776, 0.1339, 0.1361], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0311, 0.0411, 0.0413, 0.0355, 0.0412, 0.0317, 0.0370], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:00:40,212 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.67 vs. limit=5.0 +2023-04-27 12:00:49,559 INFO [finetune.py:976] (2/7) Epoch 18, batch 5450, loss[loss=0.1445, simple_loss=0.2126, pruned_loss=0.03816, over 4362.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2437, pruned_loss=0.05078, over 958503.19 frames. ], batch size: 19, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:01:23,541 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102851.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:01:32,606 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 12:01:33,555 INFO [optim.py:369] (2/7) 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 12:01:42,110 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-04-27 12:01:44,220 INFO [finetune.py:976] (2/7) Epoch 18, batch 5500, loss[loss=0.1697, simple_loss=0.2414, pruned_loss=0.04897, over 4843.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2416, pruned_loss=0.05082, over 956039.95 frames. ], batch size: 47, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:01:59,964 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102887.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:02:09,934 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 12:02:17,498 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102912.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:02:19,354 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8505, 1.3144, 1.6816, 1.7064, 1.6777, 1.3634, 0.7552, 1.3293], + device='cuda:2'), covar=tensor([0.3433, 0.3503, 0.1834, 0.2538, 0.2730, 0.2789, 0.4585, 0.2238], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0248, 0.0228, 0.0315, 0.0220, 0.0232, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 12:02:23,854 INFO [finetune.py:976] (2/7) Epoch 18, batch 5550, loss[loss=0.1561, simple_loss=0.2271, pruned_loss=0.04255, over 4721.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.244, pruned_loss=0.05169, over 956519.04 frames. ], batch size: 23, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:02:28,413 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-27 12:02:40,490 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102948.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:02:45,151 INFO [optim.py:369] (2/7) 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 12:02:54,416 INFO [finetune.py:976] (2/7) Epoch 18, batch 5600, loss[loss=0.1905, simple_loss=0.2601, pruned_loss=0.0604, over 4867.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2479, pruned_loss=0.05286, over 953768.10 frames. ], batch size: 34, lr: 3.32e-03, grad_scale: 32.0 +2023-04-27 12:02:58,618 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 12:02:59,319 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-27 12:02:59,715 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2829, 2.7899, 2.4479, 2.6580, 1.8898, 2.3223, 2.2935, 1.7270], + device='cuda:2'), covar=tensor([0.1874, 0.1077, 0.0626, 0.0976, 0.3301, 0.1215, 0.2016, 0.2513], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0304, 0.0218, 0.0281, 0.0311, 0.0260, 0.0250, 0.0266], + device='cuda:2'), out_proj_covar=tensor([1.1536e-04, 1.2102e-04, 8.6465e-05, 1.1152e-04, 1.2647e-04, 1.0299e-04, + 1.0114e-04, 1.0552e-04], device='cuda:2') +2023-04-27 12:03:24,718 INFO [finetune.py:976] (2/7) Epoch 18, batch 5650, loss[loss=0.1788, simple_loss=0.2486, pruned_loss=0.05448, over 4887.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2499, pruned_loss=0.0529, over 953814.39 frames. ], batch size: 36, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:03:38,233 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 12:03:44,557 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4033, 2.8209, 1.2821, 1.8131, 2.2868, 1.4760, 3.8160, 2.1897], + device='cuda:2'), covar=tensor([0.0591, 0.0565, 0.0729, 0.1145, 0.0464, 0.0944, 0.0264, 0.0528], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 12:03:46,252 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 18, batch 5700, loss[loss=0.1656, simple_loss=0.2147, pruned_loss=0.05824, over 4182.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2454, pruned_loss=0.05218, over 935769.59 frames. ], batch size: 18, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:04:07,298 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5328, 2.5049, 2.2579, 2.1945, 2.6048, 2.2254, 3.1465, 1.9731], + device='cuda:2'), covar=tensor([0.3298, 0.1671, 0.3449, 0.3001, 0.1456, 0.2361, 0.1154, 0.3732], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0344, 0.0425, 0.0354, 0.0378, 0.0379, 0.0369, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:04:24,055 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103098.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:04:24,531 INFO [finetune.py:976] (2/7) Epoch 19, batch 0, loss[loss=0.1559, simple_loss=0.2277, pruned_loss=0.04203, over 4859.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2277, pruned_loss=0.04203, over 4859.00 frames. ], batch size: 31, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:04:24,531 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 12:04:35,097 INFO [finetune.py:1010] (2/7) Epoch 19, validation: loss=0.1545, simple_loss=0.2248, pruned_loss=0.04209, over 2265189.00 frames. +2023-04-27 12:04:35,097 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 12:04:56,251 INFO [zipformer.py:1188] (2/7) 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] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 12:05:07,368 INFO [finetune.py:976] (2/7) Epoch 19, batch 50, loss[loss=0.1608, simple_loss=0.2502, pruned_loss=0.03569, over 4244.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2511, pruned_loss=0.05398, over 215048.03 frames. ], batch size: 66, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:05:13,154 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-04-27 12:05:13,471 INFO [optim.py:369] (2/7) 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 12:05:15,443 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103159.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:05:57,690 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 100, loss[loss=0.1496, simple_loss=0.2173, pruned_loss=0.0409, over 4131.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2456, pruned_loss=0.05323, over 377376.23 frames. ], batch size: 18, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:06:13,374 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103207.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:07:01,892 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 150, loss[loss=0.1859, simple_loss=0.246, pruned_loss=0.06289, over 4896.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2402, pruned_loss=0.05167, over 504760.88 frames. ], batch size: 35, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:07:16,416 INFO [optim.py:369] (2/7) 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 12:07:35,077 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3595, 1.6585, 1.6183, 2.1660, 1.9122, 1.9927, 1.6272, 4.3470], + device='cuda:2'), covar=tensor([0.0569, 0.0740, 0.0781, 0.1125, 0.0621, 0.0549, 0.0714, 0.0120], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 12:07:45,051 INFO [finetune.py:976] (2/7) Epoch 19, batch 200, loss[loss=0.1277, simple_loss=0.201, pruned_loss=0.02725, over 4768.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2396, pruned_loss=0.05224, over 605963.98 frames. ], batch size: 28, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:07:48,061 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1557, 1.6341, 2.0193, 2.2222, 1.9978, 1.5572, 1.1243, 1.6478], + device='cuda:2'), covar=tensor([0.3315, 0.3185, 0.1648, 0.2006, 0.2745, 0.2805, 0.4339, 0.2016], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0246, 0.0226, 0.0313, 0.0219, 0.0231, 0.0227, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 12:08:14,337 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9822, 2.5348, 1.0777, 1.4284, 1.9401, 1.2168, 3.0965, 1.7305], + device='cuda:2'), covar=tensor([0.0637, 0.0595, 0.0751, 0.1111, 0.0413, 0.0906, 0.0231, 0.0529], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0052, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 12:08:33,911 INFO [finetune.py:976] (2/7) Epoch 19, batch 250, loss[loss=0.223, simple_loss=0.2981, pruned_loss=0.07391, over 4913.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2433, pruned_loss=0.05222, over 684953.30 frames. ], batch size: 43, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:08:44,150 INFO [optim.py:369] (2/7) 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 12:08:56,544 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103366.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 12:08:58,385 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3301, 1.5730, 1.5270, 1.8023, 1.6962, 1.8145, 1.4991, 3.0260], + device='cuda:2'), covar=tensor([0.0717, 0.0880, 0.0847, 0.1119, 0.0678, 0.0761, 0.0893, 0.0239], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0014, 0.0013, 0.0012, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 12:09:14,221 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.30 vs. limit=5.0 +2023-04-27 12:09:22,446 INFO [finetune.py:976] (2/7) Epoch 19, batch 300, loss[loss=0.1694, simple_loss=0.2428, pruned_loss=0.04803, over 4743.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2479, pruned_loss=0.05373, over 744719.96 frames. ], batch size: 54, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:09:42,603 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103427.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 12:09:55,325 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6135, 1.5560, 1.9324, 2.0176, 1.5370, 1.3567, 1.6921, 1.0999], + device='cuda:2'), covar=tensor([0.0555, 0.0676, 0.0410, 0.0583, 0.0736, 0.1203, 0.0593, 0.0670], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0075, 0.0094, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 12:09:55,811 INFO [finetune.py:976] (2/7) Epoch 19, batch 350, loss[loss=0.146, simple_loss=0.2213, pruned_loss=0.03533, over 4877.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2494, pruned_loss=0.05336, over 792364.88 frames. ], batch size: 32, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:09:58,374 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4272, 2.9679, 2.3277, 2.2761, 1.7521, 1.6983, 2.4989, 1.6787], + device='cuda:2'), covar=tensor([0.1518, 0.1342, 0.1345, 0.1631, 0.2175, 0.1810, 0.0882, 0.1865], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0212, 0.0168, 0.0204, 0.0200, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 12:09:58,942 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103454.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:10:00,580 INFO [optim.py:369] (2/7) 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 12:10:00,695 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8582, 2.2922, 0.8414, 1.2205, 1.4969, 1.1648, 2.4528, 1.3228], + device='cuda:2'), covar=tensor([0.0683, 0.0534, 0.0684, 0.1257, 0.0487, 0.1034, 0.0305, 0.0721], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 12:10:05,549 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 12:10:15,746 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9799, 2.4559, 0.9946, 1.4472, 1.9109, 1.2210, 3.2343, 1.6431], + device='cuda:2'), covar=tensor([0.0707, 0.0692, 0.0837, 0.1344, 0.0533, 0.1110, 0.0329, 0.0700], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 12:10:22,421 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 400, loss[loss=0.1803, simple_loss=0.2437, pruned_loss=0.05843, over 4883.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2495, pruned_loss=0.05315, over 828584.53 frames. ], batch size: 43, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:10:30,524 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 12:10:34,538 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103507.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:10:47,358 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.62 vs. limit=5.0 +2023-04-27 12:10:57,128 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8031, 2.1574, 1.7488, 1.5681, 1.3872, 1.3604, 1.7546, 1.2920], + device='cuda:2'), covar=tensor([0.1617, 0.1316, 0.1484, 0.1736, 0.2308, 0.1959, 0.1026, 0.2031], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0213, 0.0169, 0.0205, 0.0201, 0.0185, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 12:10:58,919 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103543.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:11:02,472 INFO [finetune.py:976] (2/7) Epoch 19, batch 450, loss[loss=0.2058, simple_loss=0.2661, pruned_loss=0.07272, over 4927.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2491, pruned_loss=0.05298, over 856955.62 frames. ], batch size: 38, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:11:06,184 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103555.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:11:06,717 INFO [optim.py:369] (2/7) 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 12:11:34,262 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3985, 3.2834, 2.5654, 3.9165, 3.4137, 3.3680, 1.5146, 3.3266], + device='cuda:2'), covar=tensor([0.1834, 0.1334, 0.2869, 0.1976, 0.2729, 0.1952, 0.5568, 0.2691], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0213, 0.0247, 0.0302, 0.0297, 0.0245, 0.0269, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:11:42,147 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103591.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:11:46,989 INFO [finetune.py:976] (2/7) Epoch 19, batch 500, loss[loss=0.147, simple_loss=0.221, pruned_loss=0.03649, over 4930.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2465, pruned_loss=0.05251, over 878192.45 frames. ], batch size: 38, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:12:16,136 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5482, 1.8873, 1.9907, 2.1054, 1.9365, 2.0282, 2.0992, 2.0499], + device='cuda:2'), covar=tensor([0.3816, 0.6241, 0.4905, 0.4939, 0.6172, 0.7558, 0.5701, 0.5304], + device='cuda:2'), in_proj_covar=tensor([0.0334, 0.0374, 0.0323, 0.0334, 0.0347, 0.0396, 0.0359, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:12:30,575 INFO [finetune.py:976] (2/7) Epoch 19, batch 550, loss[loss=0.1637, simple_loss=0.2283, pruned_loss=0.0496, over 4805.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2438, pruned_loss=0.05187, over 894384.40 frames. ], batch size: 25, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:12:30,668 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4762, 1.2559, 4.0250, 3.7602, 3.4852, 3.8197, 3.7047, 3.5977], + device='cuda:2'), covar=tensor([0.6842, 0.5727, 0.1056, 0.1692, 0.1188, 0.2036, 0.2467, 0.1409], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0307, 0.0406, 0.0408, 0.0350, 0.0408, 0.0314, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:12:34,853 INFO [optim.py:369] (2/7) 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 12:12:37,845 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.6610, 4.6084, 3.0922, 5.2713, 4.6991, 4.6102, 1.9927, 4.5936], + device='cuda:2'), covar=tensor([0.1625, 0.0930, 0.3189, 0.0878, 0.3875, 0.1406, 0.5784, 0.2060], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0213, 0.0248, 0.0303, 0.0297, 0.0245, 0.0269, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:13:01,821 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1567, 2.7313, 1.2572, 1.6048, 2.1440, 1.4438, 3.2176, 1.8405], + device='cuda:2'), covar=tensor([0.0584, 0.0710, 0.0713, 0.0962, 0.0407, 0.0840, 0.0209, 0.0510], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 12:13:02,464 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0201, 1.0275, 1.1605, 1.1574, 1.0019, 0.9197, 0.9501, 0.5169], + device='cuda:2'), covar=tensor([0.0497, 0.0551, 0.0423, 0.0471, 0.0688, 0.1111, 0.0509, 0.0670], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0067, 0.0067, 0.0066, 0.0074, 0.0094, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 12:13:04,202 INFO [finetune.py:976] (2/7) Epoch 19, batch 600, loss[loss=0.1653, simple_loss=0.2415, pruned_loss=0.04449, over 4803.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2457, pruned_loss=0.05257, over 908636.91 frames. ], batch size: 45, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:13:05,502 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1987, 1.4736, 1.3728, 1.6421, 1.5527, 1.7628, 1.3767, 3.1747], + device='cuda:2'), covar=tensor([0.0621, 0.0795, 0.0759, 0.1180, 0.0639, 0.0577, 0.0756, 0.0157], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 12:13:08,603 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 12:13:19,610 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103722.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 12:13:34,134 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6980, 1.2714, 1.8390, 2.2671, 1.8278, 1.7334, 1.7528, 1.7267], + device='cuda:2'), covar=tensor([0.4951, 0.6936, 0.6530, 0.5521, 0.6132, 0.7960, 0.8301, 0.8498], + device='cuda:2'), in_proj_covar=tensor([0.0424, 0.0408, 0.0499, 0.0501, 0.0452, 0.0479, 0.0484, 0.0490], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:13:43,198 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103740.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:13:53,493 INFO [finetune.py:976] (2/7) Epoch 19, batch 650, loss[loss=0.1591, simple_loss=0.2343, pruned_loss=0.04189, over 4905.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2485, pruned_loss=0.05302, over 919267.13 frames. ], batch size: 35, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:14:01,806 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103754.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:02,874 INFO [optim.py:369] (2/7) 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 12:14:14,195 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:30,683 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103788.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:37,252 INFO [finetune.py:976] (2/7) Epoch 19, batch 700, loss[loss=0.1797, simple_loss=0.2576, pruned_loss=0.05095, over 4896.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2502, pruned_loss=0.05352, over 927130.22 frames. ], batch size: 37, lr: 3.31e-03, grad_scale: 32.0 +2023-04-27 12:14:38,715 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103801.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:39,252 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103802.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:14:55,326 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103826.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:15:02,441 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103836.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:15:10,369 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0610, 2.4993, 0.8066, 1.4823, 1.4776, 1.8558, 1.5862, 0.7965], + device='cuda:2'), covar=tensor([0.1630, 0.1231, 0.1871, 0.1303, 0.1221, 0.1016, 0.1659, 0.1905], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0244, 0.0137, 0.0121, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 12:15:10,877 INFO [finetune.py:976] (2/7) Epoch 19, batch 750, loss[loss=0.205, simple_loss=0.2693, pruned_loss=0.0704, over 4743.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2505, pruned_loss=0.05311, over 932843.62 frames. ], batch size: 27, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:15:15,074 INFO [optim.py:369] (2/7) 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 12:15:21,227 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2652, 2.9440, 0.8847, 1.4737, 1.9563, 1.2351, 4.0038, 1.8767], + device='cuda:2'), covar=tensor([0.0732, 0.0889, 0.0977, 0.1282, 0.0577, 0.1075, 0.0215, 0.0625], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 12:15:28,411 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103876.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:15:37,232 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-27 12:15:38,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2345, 1.4407, 1.7821, 1.9121, 1.8477, 1.9510, 1.8514, 1.8255], + device='cuda:2'), covar=tensor([0.3697, 0.5333, 0.4315, 0.4743, 0.5574, 0.7280, 0.4767, 0.4646], + device='cuda:2'), in_proj_covar=tensor([0.0330, 0.0369, 0.0319, 0.0330, 0.0342, 0.0392, 0.0354, 0.0325], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:15:44,006 INFO [finetune.py:976] (2/7) Epoch 19, batch 800, loss[loss=0.1561, simple_loss=0.2426, pruned_loss=0.03484, over 4789.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2491, pruned_loss=0.0525, over 935850.74 frames. ], batch size: 29, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:16:09,178 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103937.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 12:16:17,416 INFO [finetune.py:976] (2/7) Epoch 19, batch 850, loss[loss=0.1378, simple_loss=0.2179, pruned_loss=0.02882, over 4757.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2471, pruned_loss=0.05209, over 940731.98 frames. ], batch size: 27, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:16:21,649 INFO [optim.py:369] (2/7) 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 12:16:59,793 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103990.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:17:05,172 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 900, loss[loss=0.1391, simple_loss=0.2107, pruned_loss=0.03375, over 4849.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2441, pruned_loss=0.05103, over 943748.70 frames. ], batch size: 47, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:17:43,379 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104022.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 12:18:18,255 INFO [finetune.py:976] (2/7) Epoch 19, batch 950, loss[loss=0.1617, simple_loss=0.2352, pruned_loss=0.04409, over 4913.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2418, pruned_loss=0.05017, over 949025.75 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:18:18,367 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=104049.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:18:19,611 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104051.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:18:27,839 INFO [optim.py:369] (2/7) 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 12:18:29,215 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104058.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:18:41,642 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104070.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 12:19:14,740 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104096.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:19:21,308 INFO [finetune.py:976] (2/7) Epoch 19, batch 1000, loss[loss=0.1793, simple_loss=0.2457, pruned_loss=0.05643, over 4863.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2438, pruned_loss=0.05119, over 948658.69 frames. ], batch size: 34, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:19:33,845 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:19:51,330 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104121.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:20:27,141 INFO [finetune.py:976] (2/7) Epoch 19, batch 1050, loss[loss=0.2318, simple_loss=0.2991, pruned_loss=0.08224, over 4754.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2469, pruned_loss=0.05175, over 951083.93 frames. ], batch size: 54, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:20:37,627 INFO [optim.py:369] (2/7) 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 12:21:01,825 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6829, 2.3143, 1.7817, 1.6463, 1.2730, 1.2675, 1.8493, 1.2244], + device='cuda:2'), covar=tensor([0.1675, 0.1247, 0.1396, 0.1704, 0.2368, 0.1984, 0.0975, 0.2008], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0214, 0.0169, 0.0206, 0.0202, 0.0186, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 12:21:05,398 INFO [finetune.py:976] (2/7) Epoch 19, batch 1100, loss[loss=0.173, simple_loss=0.2413, pruned_loss=0.05231, over 4902.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2476, pruned_loss=0.0518, over 949629.73 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:21:15,525 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 12:21:27,610 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104232.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 12:21:39,382 INFO [finetune.py:976] (2/7) Epoch 19, batch 1150, loss[loss=0.2055, simple_loss=0.2702, pruned_loss=0.07043, over 4753.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.247, pruned_loss=0.05167, over 950582.78 frames. ], batch size: 28, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:21:44,616 INFO [optim.py:369] (2/7) 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 12:21:46,022 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 12:22:03,396 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 12:22:09,404 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-27 12:22:12,702 INFO [finetune.py:976] (2/7) Epoch 19, batch 1200, loss[loss=0.1838, simple_loss=0.2531, pruned_loss=0.05726, over 4840.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2463, pruned_loss=0.05143, over 952404.25 frames. ], batch size: 47, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:22:44,235 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 1250, loss[loss=0.1209, simple_loss=0.1934, pruned_loss=0.02421, over 4763.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2442, pruned_loss=0.05093, over 950944.50 frames. ], batch size: 27, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:22:49,475 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104353.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:22:51,238 INFO [optim.py:369] (2/7) 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 12:22:51,377 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8291, 1.5109, 1.5019, 1.6657, 2.0814, 1.6524, 1.4273, 1.3436], + device='cuda:2'), covar=tensor([0.1287, 0.1399, 0.1534, 0.1240, 0.0674, 0.1543, 0.2095, 0.1893], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0311, 0.0348, 0.0289, 0.0328, 0.0307, 0.0300, 0.0367], + device='cuda:2'), out_proj_covar=tensor([6.3009e-05, 6.4581e-05, 7.3871e-05, 5.8714e-05, 6.8193e-05, 6.4513e-05, + 6.2959e-05, 7.8048e-05], device='cuda:2') +2023-04-27 12:23:06,333 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5872, 1.7006, 1.8437, 2.0406, 1.8309, 1.9614, 1.9847, 2.0065], + device='cuda:2'), covar=tensor([0.3611, 0.5751, 0.5267, 0.4555, 0.6020, 0.7505, 0.5607, 0.4884], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0373, 0.0322, 0.0334, 0.0345, 0.0395, 0.0359, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:23:45,732 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:23:46,493 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 12:23:47,995 INFO [finetune.py:976] (2/7) Epoch 19, batch 1300, loss[loss=0.2041, simple_loss=0.2749, pruned_loss=0.06662, over 4821.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2411, pruned_loss=0.04987, over 951330.87 frames. ], batch size: 40, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:23:56,156 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104405.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:24:18,862 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104421.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:24:42,124 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3109, 1.5591, 1.6791, 1.8425, 1.6896, 1.8174, 1.8005, 1.7733], + device='cuda:2'), covar=tensor([0.3875, 0.4767, 0.4401, 0.3962, 0.5303, 0.6720, 0.4738, 0.4250], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0372, 0.0322, 0.0333, 0.0345, 0.0394, 0.0358, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:24:42,672 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9831, 1.1884, 3.2140, 2.9773, 2.8701, 3.1232, 3.0700, 2.8918], + device='cuda:2'), covar=tensor([0.6833, 0.5324, 0.1306, 0.1952, 0.1304, 0.1945, 0.2738, 0.1402], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0307, 0.0405, 0.0407, 0.0351, 0.0408, 0.0314, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:24:44,516 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104444.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:24:53,501 INFO [finetune.py:976] (2/7) Epoch 19, batch 1350, loss[loss=0.1475, simple_loss=0.2231, pruned_loss=0.03597, over 4775.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2414, pruned_loss=0.05072, over 950596.97 frames. ], batch size: 26, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:25:03,939 INFO [optim.py:369] (2/7) 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 12:25:14,768 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 12:25:23,556 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104469.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:25:39,303 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9914, 2.4515, 2.0616, 2.3373, 1.6345, 2.0319, 2.0625, 1.5814], + device='cuda:2'), covar=tensor([0.2148, 0.1136, 0.0678, 0.1133, 0.3386, 0.1058, 0.1865, 0.2471], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0301, 0.0216, 0.0277, 0.0310, 0.0257, 0.0250, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1474e-04, 1.1940e-04, 8.5565e-05, 1.1002e-04, 1.2572e-04, 1.0178e-04, + 1.0072e-04, 1.0463e-04], device='cuda:2') +2023-04-27 12:25:58,402 INFO [finetune.py:976] (2/7) Epoch 19, batch 1400, loss[loss=0.1668, simple_loss=0.2437, pruned_loss=0.04499, over 4820.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2447, pruned_loss=0.05187, over 951272.40 frames. ], batch size: 39, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:26:25,271 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-27 12:26:30,570 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104532.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 12:26:33,712 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 12:26:38,482 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8139, 3.7557, 2.7935, 4.4302, 3.8466, 3.8226, 1.7777, 3.7476], + device='cuda:2'), covar=tensor([0.1767, 0.1190, 0.3059, 0.1626, 0.3562, 0.1627, 0.5410, 0.2444], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0215, 0.0249, 0.0308, 0.0301, 0.0249, 0.0272, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:26:41,374 INFO [finetune.py:976] (2/7) Epoch 19, batch 1450, loss[loss=0.1908, simple_loss=0.2589, pruned_loss=0.06137, over 4833.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2474, pruned_loss=0.05239, over 952861.94 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:26:43,967 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7509, 1.1047, 1.8603, 2.2151, 1.7919, 1.7039, 1.7567, 1.7238], + device='cuda:2'), covar=tensor([0.4518, 0.6730, 0.6125, 0.5527, 0.6317, 0.7995, 0.7473, 0.8117], + device='cuda:2'), in_proj_covar=tensor([0.0426, 0.0410, 0.0501, 0.0504, 0.0454, 0.0483, 0.0488, 0.0493], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:26:46,126 INFO [optim.py:369] (2/7) 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 12:26:54,562 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8470, 1.2166, 3.2828, 3.0435, 2.8839, 3.2218, 3.2131, 2.8681], + device='cuda:2'), covar=tensor([0.7709, 0.5593, 0.1417, 0.2242, 0.1485, 0.1804, 0.1488, 0.1630], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0307, 0.0404, 0.0406, 0.0351, 0.0407, 0.0314, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:27:03,311 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104580.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:27:14,736 INFO [finetune.py:976] (2/7) Epoch 19, batch 1500, loss[loss=0.2054, simple_loss=0.2725, pruned_loss=0.06913, over 4703.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2496, pruned_loss=0.05345, over 954792.00 frames. ], batch size: 59, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:27:15,012 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 12:27:46,867 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104646.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:27:48,604 INFO [finetune.py:976] (2/7) Epoch 19, batch 1550, loss[loss=0.1579, simple_loss=0.2235, pruned_loss=0.04614, over 4799.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2497, pruned_loss=0.05294, over 956298.81 frames. ], batch size: 25, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:27:51,164 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104653.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:27:53,064 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-04-27 12:27:53,360 INFO [optim.py:369] (2/7) 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 12:28:35,127 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104694.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:28:43,384 INFO [finetune.py:976] (2/7) Epoch 19, batch 1600, loss[loss=0.1815, simple_loss=0.2381, pruned_loss=0.06241, over 4872.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2461, pruned_loss=0.05155, over 956110.80 frames. ], batch size: 34, lr: 3.30e-03, grad_scale: 32.0 +2023-04-27 12:28:44,680 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104701.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:28:47,655 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104705.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:29:06,198 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7630, 1.6724, 4.4698, 3.9277, 3.9542, 4.1542, 4.0554, 3.7613], + device='cuda:2'), covar=tensor([0.9772, 0.7915, 0.1532, 0.3044, 0.2005, 0.3619, 0.2197, 0.2905], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0305, 0.0403, 0.0405, 0.0349, 0.0406, 0.0312, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:29:13,486 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.86 vs. limit=5.0 +2023-04-27 12:29:16,967 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1444, 2.5472, 0.7633, 1.4480, 1.6117, 1.8577, 1.6516, 0.8427], + device='cuda:2'), covar=tensor([0.1615, 0.1262, 0.2037, 0.1494, 0.1203, 0.1071, 0.1837, 0.1750], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0241, 0.0136, 0.0119, 0.0130, 0.0151, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 12:29:17,626 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5003, 1.6730, 1.8346, 1.9740, 1.8136, 1.8819, 1.9740, 1.9139], + device='cuda:2'), covar=tensor([0.4042, 0.5437, 0.4793, 0.4773, 0.5799, 0.7234, 0.5276, 0.5296], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0374, 0.0322, 0.0334, 0.0346, 0.0395, 0.0359, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:29:25,439 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3974, 1.7016, 1.6929, 2.2306, 2.4378, 1.9119, 1.8699, 1.7473], + device='cuda:2'), covar=tensor([0.1893, 0.2006, 0.2054, 0.1708, 0.1486, 0.2304, 0.2457, 0.2749], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0313, 0.0350, 0.0291, 0.0330, 0.0309, 0.0302, 0.0369], + device='cuda:2'), out_proj_covar=tensor([6.3326e-05, 6.5061e-05, 7.4257e-05, 5.8990e-05, 6.8452e-05, 6.5024e-05, + 6.3350e-05, 7.8614e-05], device='cuda:2') +2023-04-27 12:29:27,150 INFO [finetune.py:976] (2/7) Epoch 19, batch 1650, loss[loss=0.1686, simple_loss=0.2412, pruned_loss=0.04802, over 4900.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2439, pruned_loss=0.05075, over 957264.75 frames. ], batch size: 32, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:29:29,682 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104753.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:29:30,981 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3696, 1.8873, 2.2707, 2.7509, 2.2741, 1.8044, 1.6337, 2.0418], + device='cuda:2'), covar=tensor([0.3197, 0.3169, 0.1781, 0.2296, 0.2540, 0.2642, 0.4296, 0.2166], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0247, 0.0228, 0.0316, 0.0220, 0.0232, 0.0228, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 12:29:31,432 INFO [optim.py:369] (2/7) 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 12:29:35,817 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7114, 1.3304, 1.9213, 2.2318, 1.8607, 1.7482, 1.8215, 1.7792], + device='cuda:2'), covar=tensor([0.4626, 0.6552, 0.6254, 0.5438, 0.6057, 0.7928, 0.7918, 0.8624], + device='cuda:2'), in_proj_covar=tensor([0.0426, 0.0409, 0.0501, 0.0504, 0.0455, 0.0483, 0.0487, 0.0493], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:29:44,030 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7466, 3.6525, 2.6965, 4.3679, 3.6597, 3.8114, 1.5767, 3.7167], + device='cuda:2'), covar=tensor([0.1761, 0.1386, 0.3644, 0.1372, 0.4208, 0.1704, 0.5645, 0.2339], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0214, 0.0247, 0.0306, 0.0298, 0.0247, 0.0270, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:29:45,738 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8375, 1.2145, 1.3005, 1.6534, 1.9327, 1.5700, 1.3554, 1.2454], + device='cuda:2'), covar=tensor([0.1341, 0.1733, 0.2462, 0.1210, 0.0979, 0.1656, 0.2115, 0.2379], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0314, 0.0351, 0.0291, 0.0330, 0.0310, 0.0302, 0.0370], + device='cuda:2'), out_proj_covar=tensor([6.3373e-05, 6.5077e-05, 7.4366e-05, 5.8954e-05, 6.8473e-05, 6.5048e-05, + 6.3391e-05, 7.8693e-05], device='cuda:2') +2023-04-27 12:30:01,073 INFO [finetune.py:976] (2/7) Epoch 19, batch 1700, loss[loss=0.1483, simple_loss=0.2116, pruned_loss=0.04247, over 3716.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2421, pruned_loss=0.05018, over 955787.72 frames. ], batch size: 16, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:30:48,746 INFO [finetune.py:976] (2/7) Epoch 19, batch 1750, loss[loss=0.1527, simple_loss=0.2389, pruned_loss=0.03323, over 4851.00 frames. ], tot_loss[loss=0.172, simple_loss=0.243, pruned_loss=0.05052, over 954309.17 frames. ], batch size: 44, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:30:52,628 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 12:30:53,007 INFO [optim.py:369] (2/7) 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 12:30:53,151 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9370, 1.4627, 1.8042, 2.1362, 1.7354, 1.4270, 1.0420, 1.5475], + device='cuda:2'), covar=tensor([0.3377, 0.3518, 0.1856, 0.2158, 0.2727, 0.2720, 0.4453, 0.2206], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0246, 0.0226, 0.0315, 0.0219, 0.0231, 0.0228, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 12:31:37,544 INFO [finetune.py:976] (2/7) Epoch 19, batch 1800, loss[loss=0.1986, simple_loss=0.2736, pruned_loss=0.06185, over 4895.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2475, pruned_loss=0.05204, over 955803.32 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:32:10,885 INFO [finetune.py:976] (2/7) Epoch 19, batch 1850, loss[loss=0.1991, simple_loss=0.2597, pruned_loss=0.06931, over 4910.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2489, pruned_loss=0.05292, over 956508.16 frames. ], batch size: 46, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:32:15,623 INFO [optim.py:369] (2/7) 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 12:32:26,532 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2672, 1.2897, 1.4146, 1.5784, 1.5741, 1.2541, 0.9737, 1.4166], + device='cuda:2'), covar=tensor([0.0933, 0.1276, 0.0879, 0.0607, 0.0772, 0.0913, 0.0874, 0.0659], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0197, 0.0179, 0.0169, 0.0174, 0.0178, 0.0149, 0.0175], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:32:35,599 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 12:32:44,648 INFO [finetune.py:976] (2/7) Epoch 19, batch 1900, loss[loss=0.2094, simple_loss=0.2854, pruned_loss=0.06672, over 4843.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2501, pruned_loss=0.05316, over 957534.17 frames. ], batch size: 44, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:32:57,440 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.3044, 4.1784, 2.8804, 4.8909, 4.1806, 4.3605, 1.6746, 4.2271], + device='cuda:2'), covar=tensor([0.1464, 0.0946, 0.3619, 0.1021, 0.2683, 0.1359, 0.5294, 0.2030], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0216, 0.0249, 0.0307, 0.0301, 0.0249, 0.0272, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:33:01,013 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3629, 1.6334, 1.4984, 1.6004, 1.4090, 1.3559, 1.4160, 1.1074], + device='cuda:2'), covar=tensor([0.1659, 0.1138, 0.0809, 0.1101, 0.3344, 0.1212, 0.1639, 0.1988], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0301, 0.0215, 0.0277, 0.0309, 0.0256, 0.0249, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1467e-04, 1.1947e-04, 8.5294e-05, 1.1004e-04, 1.2548e-04, 1.0165e-04, + 1.0064e-04, 1.0458e-04], device='cuda:2') +2023-04-27 12:33:03,457 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8971, 2.1128, 1.9681, 2.1787, 1.9279, 2.0695, 2.0897, 2.0695], + device='cuda:2'), covar=tensor([0.3999, 0.6056, 0.5521, 0.4522, 0.6072, 0.7358, 0.6252, 0.6096], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0371, 0.0321, 0.0333, 0.0345, 0.0394, 0.0357, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:33:13,826 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4432, 2.8994, 1.0077, 1.4190, 2.2630, 1.4256, 4.2023, 1.8959], + device='cuda:2'), covar=tensor([0.0642, 0.0820, 0.0913, 0.1271, 0.0550, 0.1040, 0.0244, 0.0659], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 12:33:18,033 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4271, 1.6501, 1.8162, 1.9534, 1.8568, 1.8969, 1.9185, 1.8891], + device='cuda:2'), covar=tensor([0.3727, 0.5564, 0.4810, 0.4406, 0.5130, 0.7017, 0.5088, 0.4818], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0371, 0.0321, 0.0332, 0.0345, 0.0393, 0.0356, 0.0327], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:33:18,502 INFO [finetune.py:976] (2/7) Epoch 19, batch 1950, loss[loss=0.1845, simple_loss=0.2592, pruned_loss=0.05491, over 4925.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2489, pruned_loss=0.05283, over 958183.27 frames. ], batch size: 38, lr: 3.30e-03, grad_scale: 64.0 +2023-04-27 12:33:22,765 INFO [optim.py:369] (2/7) 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 12:34:18,468 INFO [finetune.py:976] (2/7) Epoch 19, batch 2000, loss[loss=0.163, simple_loss=0.2318, pruned_loss=0.0471, over 4767.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2465, pruned_loss=0.05174, over 958234.64 frames. ], batch size: 28, lr: 3.29e-03, grad_scale: 64.0 +2023-04-27 12:34:45,389 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.60 vs. limit=5.0 +2023-04-27 12:35:13,450 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2661, 1.6677, 1.6119, 2.0804, 2.2992, 2.0050, 1.8900, 1.6896], + device='cuda:2'), covar=tensor([0.1668, 0.1832, 0.1739, 0.1569, 0.1256, 0.1785, 0.2341, 0.2353], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0312, 0.0349, 0.0289, 0.0327, 0.0307, 0.0300, 0.0368], + device='cuda:2'), out_proj_covar=tensor([6.3071e-05, 6.4724e-05, 7.4066e-05, 5.8689e-05, 6.7872e-05, 6.4597e-05, + 6.3031e-05, 7.8342e-05], device='cuda:2') +2023-04-27 12:35:14,516 INFO [finetune.py:976] (2/7) Epoch 19, batch 2050, loss[loss=0.1969, simple_loss=0.2552, pruned_loss=0.06932, over 4821.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2424, pruned_loss=0.05048, over 958786.34 frames. ], batch size: 38, lr: 3.29e-03, grad_scale: 64.0 +2023-04-27 12:35:18,783 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 19, batch 2100, loss[loss=0.1667, simple_loss=0.2319, pruned_loss=0.05074, over 4697.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.243, pruned_loss=0.05138, over 958605.53 frames. ], batch size: 23, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:36:37,772 INFO [finetune.py:976] (2/7) Epoch 19, batch 2150, loss[loss=0.1912, simple_loss=0.2796, pruned_loss=0.0514, over 4812.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2463, pruned_loss=0.05254, over 959767.90 frames. ], batch size: 45, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:36:48,984 INFO [optim.py:369] (2/7) 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 12:37:36,678 INFO [finetune.py:976] (2/7) Epoch 19, batch 2200, loss[loss=0.1601, simple_loss=0.2263, pruned_loss=0.04697, over 4188.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2485, pruned_loss=0.05316, over 959735.00 frames. ], batch size: 65, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:38:04,759 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0324, 2.4349, 2.0646, 1.8742, 1.6797, 1.6949, 2.1113, 1.7190], + device='cuda:2'), covar=tensor([0.1259, 0.1403, 0.1139, 0.1501, 0.2105, 0.1501, 0.0837, 0.1602], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0213, 0.0170, 0.0205, 0.0200, 0.0186, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 12:38:27,531 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 12:38:49,116 INFO [finetune.py:976] (2/7) Epoch 19, batch 2250, loss[loss=0.166, simple_loss=0.2436, pruned_loss=0.0442, over 4724.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2502, pruned_loss=0.05391, over 956801.67 frames. ], batch size: 54, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:38:49,911 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-27 12:38:59,481 INFO [optim.py:369] (2/7) 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 12:39:34,797 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105388.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:39:53,432 INFO [finetune.py:976] (2/7) Epoch 19, batch 2300, loss[loss=0.1747, simple_loss=0.2468, pruned_loss=0.05128, over 4764.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2505, pruned_loss=0.05371, over 957656.56 frames. ], batch size: 27, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:40:49,815 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3315, 1.7751, 2.1483, 2.7017, 2.1746, 1.7653, 1.5104, 1.9726], + device='cuda:2'), covar=tensor([0.3409, 0.3370, 0.1765, 0.2434, 0.2822, 0.2914, 0.4311, 0.2227], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0247, 0.0228, 0.0315, 0.0218, 0.0232, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 12:40:58,684 INFO [finetune.py:976] (2/7) Epoch 19, batch 2350, loss[loss=0.1535, simple_loss=0.2149, pruned_loss=0.04602, over 4414.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2474, pruned_loss=0.05251, over 956156.11 frames. ], batch size: 19, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:40:58,842 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105449.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 12:41:09,880 INFO [optim.py:369] (2/7) 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 12:41:22,068 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7599, 2.0461, 1.2444, 1.4630, 2.2604, 1.6301, 1.5598, 1.7205], + device='cuda:2'), covar=tensor([0.0495, 0.0347, 0.0294, 0.0560, 0.0232, 0.0519, 0.0477, 0.0567], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 12:41:52,223 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 2400, loss[loss=0.1661, simple_loss=0.2375, pruned_loss=0.04739, over 4907.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2444, pruned_loss=0.05152, over 956350.18 frames. ], batch size: 35, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:42:08,504 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105502.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:42:21,469 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 12:42:32,838 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1257, 2.5896, 2.1862, 2.4535, 1.7100, 2.2693, 2.2199, 1.7823], + device='cuda:2'), covar=tensor([0.2097, 0.1056, 0.0723, 0.1124, 0.3447, 0.0970, 0.1926, 0.2547], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0302, 0.0215, 0.0278, 0.0310, 0.0257, 0.0250, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1476e-04, 1.1982e-04, 8.5403e-05, 1.1053e-04, 1.2588e-04, 1.0182e-04, + 1.0080e-04, 1.0447e-04], device='cuda:2') +2023-04-27 12:42:39,430 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105547.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:42:40,535 INFO [finetune.py:976] (2/7) Epoch 19, batch 2450, loss[loss=0.1433, simple_loss=0.2068, pruned_loss=0.03992, over 4789.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2409, pruned_loss=0.05046, over 952907.04 frames. ], batch size: 26, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:42:45,853 INFO [optim.py:369] (2/7) 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 12:42:50,152 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105563.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:43:05,156 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-04-27 12:43:14,477 INFO [finetune.py:976] (2/7) Epoch 19, batch 2500, loss[loss=0.1536, simple_loss=0.2359, pruned_loss=0.03558, over 4808.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2424, pruned_loss=0.05099, over 952936.85 frames. ], batch size: 45, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:43:47,976 INFO [finetune.py:976] (2/7) Epoch 19, batch 2550, loss[loss=0.1563, simple_loss=0.2379, pruned_loss=0.03734, over 4816.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2465, pruned_loss=0.0519, over 953398.13 frames. ], batch size: 39, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:43:53,308 INFO [optim.py:369] (2/7) 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 12:44:15,066 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3530, 1.2560, 1.3937, 1.5968, 1.6274, 1.2527, 1.0786, 1.4812], + device='cuda:2'), covar=tensor([0.0900, 0.1395, 0.0895, 0.0672, 0.0730, 0.0870, 0.0827, 0.0623], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0199, 0.0181, 0.0171, 0.0176, 0.0180, 0.0150, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 12:44:27,705 INFO [finetune.py:976] (2/7) Epoch 19, batch 2600, loss[loss=0.1563, simple_loss=0.2396, pruned_loss=0.03648, over 4794.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2483, pruned_loss=0.05189, over 955464.83 frames. ], batch size: 29, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:44:57,493 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8918, 2.8028, 2.2337, 3.3084, 2.8335, 2.9075, 1.2048, 2.7710], + device='cuda:2'), covar=tensor([0.2119, 0.1628, 0.3324, 0.3221, 0.2900, 0.2150, 0.5648, 0.3143], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0217, 0.0249, 0.0308, 0.0300, 0.0248, 0.0272, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:44:58,140 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105744.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 12:45:01,137 INFO [finetune.py:976] (2/7) Epoch 19, batch 2650, loss[loss=0.2027, simple_loss=0.2845, pruned_loss=0.06045, over 4098.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2492, pruned_loss=0.05194, over 954036.55 frames. ], batch size: 65, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:45:06,397 INFO [optim.py:369] (2/7) 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 12:45:30,692 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105792.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:45:34,813 INFO [finetune.py:976] (2/7) Epoch 19, batch 2700, loss[loss=0.1421, simple_loss=0.2201, pruned_loss=0.03206, over 4795.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2481, pruned_loss=0.05155, over 955176.13 frames. ], batch size: 25, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:45:40,221 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105799.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:45:41,998 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4309, 1.2611, 1.6229, 1.6086, 1.3013, 1.1800, 1.2880, 0.7580], + device='cuda:2'), covar=tensor([0.0548, 0.0643, 0.0382, 0.0546, 0.0739, 0.1273, 0.0667, 0.0702], + device='cuda:2'), in_proj_covar=tensor([0.0067, 0.0067, 0.0066, 0.0066, 0.0073, 0.0094, 0.0072, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 12:46:35,700 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7630, 3.3744, 2.6827, 3.1090, 2.5028, 2.8864, 2.9522, 2.2515], + device='cuda:2'), covar=tensor([0.1679, 0.1078, 0.0759, 0.1013, 0.2669, 0.1044, 0.1769, 0.2474], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0304, 0.0218, 0.0280, 0.0313, 0.0260, 0.0251, 0.0266], + device='cuda:2'), out_proj_covar=tensor([1.1561e-04, 1.2082e-04, 8.6422e-05, 1.1122e-04, 1.2715e-04, 1.0283e-04, + 1.0145e-04, 1.0562e-04], device='cuda:2') +2023-04-27 12:46:36,275 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105842.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:46:40,480 INFO [finetune.py:976] (2/7) Epoch 19, batch 2750, loss[loss=0.2165, simple_loss=0.2756, pruned_loss=0.07865, over 4758.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.245, pruned_loss=0.05103, over 956091.15 frames. ], batch size: 59, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:46:43,050 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:46:46,595 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105860.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:47:01,294 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4852, 1.0910, 1.2773, 1.2331, 1.6183, 1.3459, 1.1223, 1.2160], + device='cuda:2'), covar=tensor([0.1327, 0.1175, 0.1587, 0.1167, 0.0691, 0.1253, 0.1766, 0.2010], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0315, 0.0354, 0.0294, 0.0331, 0.0312, 0.0304, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4207e-05, 6.5347e-05, 7.5132e-05, 5.9662e-05, 6.8684e-05, 6.5620e-05, + 6.3762e-05, 7.9537e-05], device='cuda:2') +2023-04-27 12:47:23,412 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 12:47:35,994 INFO [finetune.py:976] (2/7) Epoch 19, batch 2800, loss[loss=0.1802, simple_loss=0.2497, pruned_loss=0.0554, over 4869.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2421, pruned_loss=0.05037, over 957796.81 frames. ], batch size: 31, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:47:37,969 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8837, 2.1547, 1.9898, 2.1690, 1.9492, 2.1490, 2.0726, 1.9851], + device='cuda:2'), covar=tensor([0.3885, 0.6126, 0.4970, 0.4603, 0.6231, 0.7271, 0.6438, 0.5971], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0372, 0.0319, 0.0333, 0.0344, 0.0394, 0.0357, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:48:43,335 INFO [finetune.py:976] (2/7) Epoch 19, batch 2850, loss[loss=0.2024, simple_loss=0.2635, pruned_loss=0.07063, over 4761.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2423, pruned_loss=0.05101, over 957928.40 frames. ], batch size: 54, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:48:53,755 INFO [optim.py:369] (2/7) 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 12:48:54,528 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7528, 2.0188, 1.9759, 2.0949, 1.9379, 2.0332, 2.0622, 2.0041], + device='cuda:2'), covar=tensor([0.3839, 0.6288, 0.4753, 0.4638, 0.5819, 0.6920, 0.6283, 0.5403], + device='cuda:2'), in_proj_covar=tensor([0.0334, 0.0370, 0.0318, 0.0332, 0.0342, 0.0392, 0.0355, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:49:19,920 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-04-27 12:49:49,884 INFO [finetune.py:976] (2/7) Epoch 19, batch 2900, loss[loss=0.2136, simple_loss=0.2836, pruned_loss=0.07176, over 4802.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2449, pruned_loss=0.05185, over 958119.42 frames. ], batch size: 45, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:50:40,785 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106044.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:50:43,675 INFO [finetune.py:976] (2/7) Epoch 19, batch 2950, loss[loss=0.1982, simple_loss=0.2664, pruned_loss=0.06498, over 4821.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2476, pruned_loss=0.05244, over 955937.75 frames. ], batch size: 45, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:50:48,563 INFO [optim.py:369] (2/7) 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 12:50:51,652 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2557, 1.7535, 2.0741, 2.4904, 2.1282, 1.6849, 1.3474, 1.9223], + device='cuda:2'), covar=tensor([0.2966, 0.3031, 0.1685, 0.2117, 0.2436, 0.2630, 0.4214, 0.1855], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0245, 0.0226, 0.0314, 0.0217, 0.0230, 0.0226, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 12:50:54,820 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-04-27 12:51:01,762 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106092.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:51:17,520 INFO [finetune.py:976] (2/7) Epoch 19, batch 3000, loss[loss=0.1712, simple_loss=0.2512, pruned_loss=0.0456, over 4894.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2467, pruned_loss=0.05188, over 954681.11 frames. ], batch size: 36, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:51:17,520 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 12:51:22,493 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9772, 1.8853, 1.7420, 1.5713, 1.9195, 1.7189, 2.1720, 1.5606], + device='cuda:2'), covar=tensor([0.2891, 0.1540, 0.3633, 0.2105, 0.1333, 0.1812, 0.1424, 0.3696], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0343, 0.0424, 0.0351, 0.0380, 0.0374, 0.0370, 0.0415], + device='cuda:2'), out_proj_covar=tensor([9.9476e-05, 1.0309e-04, 1.2892e-04, 1.0590e-04, 1.1345e-04, 1.1206e-04, + 1.0897e-04, 1.2569e-04], device='cuda:2') +2023-04-27 12:51:33,614 INFO [finetune.py:1010] (2/7) Epoch 19, validation: loss=0.1523, simple_loss=0.2226, pruned_loss=0.04099, over 2265189.00 frames. +2023-04-27 12:51:33,614 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 12:52:14,879 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2575, 1.6216, 1.5719, 1.9444, 1.7640, 1.9311, 1.5210, 3.8800], + device='cuda:2'), covar=tensor([0.0661, 0.0883, 0.0843, 0.1220, 0.0694, 0.0527, 0.0820, 0.0149], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 12:52:18,497 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106138.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:52:26,618 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106142.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:52:36,649 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106148.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:52:37,147 INFO [finetune.py:976] (2/7) Epoch 19, batch 3050, loss[loss=0.1739, simple_loss=0.2496, pruned_loss=0.04907, over 4780.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2486, pruned_loss=0.05228, over 954482.34 frames. ], batch size: 25, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:52:46,797 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 12:52:49,215 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:53:31,156 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106190.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:53:33,060 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3618, 1.5388, 1.3925, 1.5205, 1.3125, 1.3118, 1.4501, 1.0763], + device='cuda:2'), covar=tensor([0.1606, 0.1223, 0.0899, 0.1302, 0.3350, 0.1208, 0.1597, 0.2151], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0304, 0.0217, 0.0279, 0.0312, 0.0259, 0.0251, 0.0266], + device='cuda:2'), out_proj_covar=tensor([1.1490e-04, 1.2064e-04, 8.6136e-05, 1.1087e-04, 1.2670e-04, 1.0255e-04, + 1.0153e-04, 1.0536e-04], device='cuda:2') +2023-04-27 12:53:42,132 INFO [finetune.py:976] (2/7) Epoch 19, batch 3100, loss[loss=0.2191, simple_loss=0.2803, pruned_loss=0.07892, over 4901.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.247, pruned_loss=0.05178, over 955866.77 frames. ], batch size: 43, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:53:51,783 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:54:35,394 INFO [finetune.py:976] (2/7) Epoch 19, batch 3150, loss[loss=0.1744, simple_loss=0.2437, pruned_loss=0.05259, over 4895.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2449, pruned_loss=0.05163, over 957048.87 frames. ], batch size: 43, lr: 3.29e-03, grad_scale: 32.0 +2023-04-27 12:54:40,719 INFO [optim.py:369] (2/7) 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 12:54:49,457 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 3200, loss[loss=0.1578, simple_loss=0.233, pruned_loss=0.04133, over 4819.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2419, pruned_loss=0.05085, over 957931.02 frames. ], batch size: 30, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:55:47,938 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1486, 2.8367, 1.0920, 1.3447, 2.1840, 1.3856, 3.7152, 1.8331], + device='cuda:2'), covar=tensor([0.0691, 0.0575, 0.0805, 0.1296, 0.0509, 0.0973, 0.0188, 0.0669], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 12:55:49,185 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7285, 2.1161, 1.8951, 1.4755, 1.3079, 1.3112, 1.9621, 1.2534], + device='cuda:2'), covar=tensor([0.1730, 0.1333, 0.1301, 0.1709, 0.2263, 0.1883, 0.0909, 0.2030], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0212, 0.0168, 0.0203, 0.0199, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 12:55:50,415 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:55:54,735 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106336.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:55:55,914 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4460, 3.0264, 0.9529, 1.5519, 2.2757, 1.6902, 4.1804, 2.1125], + device='cuda:2'), covar=tensor([0.0611, 0.0646, 0.0882, 0.1271, 0.0512, 0.0884, 0.0225, 0.0607], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 12:55:55,923 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5037, 1.7357, 1.5430, 1.9098, 1.9749, 2.0842, 1.6395, 4.3516], + device='cuda:2'), covar=tensor([0.0512, 0.0760, 0.0805, 0.1172, 0.0586, 0.0518, 0.0701, 0.0130], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0039, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 12:56:03,601 INFO [finetune.py:976] (2/7) Epoch 19, batch 3250, loss[loss=0.1817, simple_loss=0.2515, pruned_loss=0.05598, over 4723.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2435, pruned_loss=0.05178, over 957782.52 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:56:08,525 INFO [optim.py:369] (2/7) 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 12:56:35,791 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106397.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:56:37,423 INFO [finetune.py:976] (2/7) Epoch 19, batch 3300, loss[loss=0.1825, simple_loss=0.2717, pruned_loss=0.04667, over 4894.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2488, pruned_loss=0.0536, over 957481.36 frames. ], batch size: 43, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:56:44,871 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2878, 4.2784, 2.9650, 4.9408, 4.2433, 4.2848, 1.7913, 4.2243], + device='cuda:2'), covar=tensor([0.1775, 0.1045, 0.3629, 0.0996, 0.2603, 0.1629, 0.5716, 0.2188], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0216, 0.0249, 0.0309, 0.0300, 0.0248, 0.0271, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:56:51,369 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5245, 1.1266, 1.2552, 1.2049, 1.6125, 1.2475, 1.0525, 1.2137], + device='cuda:2'), covar=tensor([0.1542, 0.1340, 0.2006, 0.1383, 0.0874, 0.1565, 0.1869, 0.2233], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0312, 0.0350, 0.0289, 0.0327, 0.0309, 0.0301, 0.0370], + device='cuda:2'), out_proj_covar=tensor([6.3441e-05, 6.4630e-05, 7.4204e-05, 5.8529e-05, 6.7867e-05, 6.4911e-05, + 6.3078e-05, 7.8929e-05], device='cuda:2') +2023-04-27 12:56:55,723 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 12:57:00,457 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106433.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:09,675 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 3350, loss[loss=0.2034, simple_loss=0.2573, pruned_loss=0.0748, over 4730.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.251, pruned_loss=0.05397, over 958768.08 frames. ], batch size: 54, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:57:14,360 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106454.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:14,959 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106496.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:42,288 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 12:57:43,904 INFO [finetune.py:976] (2/7) Epoch 19, batch 3400, loss[loss=0.1595, simple_loss=0.2225, pruned_loss=0.04821, over 4780.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2526, pruned_loss=0.05462, over 958148.04 frames. ], batch size: 27, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:57:45,889 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4193, 3.1156, 0.8153, 1.6988, 1.7676, 2.1200, 1.8581, 0.9544], + device='cuda:2'), covar=tensor([0.1408, 0.0876, 0.2029, 0.1311, 0.1065, 0.1057, 0.1538, 0.1894], + device='cuda:2'), in_proj_covar=tensor([0.0115, 0.0240, 0.0136, 0.0119, 0.0130, 0.0150, 0.0115, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 12:57:46,915 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106503.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:49,920 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6512, 1.4971, 1.9324, 1.9965, 1.4873, 1.3647, 1.5763, 1.0265], + device='cuda:2'), covar=tensor([0.0517, 0.0642, 0.0371, 0.0495, 0.0684, 0.1116, 0.0601, 0.0646], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0075, 0.0095, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 12:57:54,771 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106515.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:57:57,744 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9139, 2.5773, 2.0737, 1.9894, 1.4387, 1.4561, 2.0844, 1.3804], + device='cuda:2'), covar=tensor([0.1894, 0.1343, 0.1403, 0.1663, 0.2404, 0.1981, 0.1013, 0.2174], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0212, 0.0169, 0.0204, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 12:57:58,315 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3772, 3.2781, 2.4585, 3.8438, 3.2618, 3.3331, 1.4484, 3.2497], + device='cuda:2'), covar=tensor([0.1924, 0.1409, 0.3345, 0.2447, 0.2603, 0.1957, 0.5872, 0.2626], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0215, 0.0248, 0.0307, 0.0298, 0.0247, 0.0271, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:58:17,776 INFO [finetune.py:976] (2/7) Epoch 19, batch 3450, loss[loss=0.1878, simple_loss=0.2637, pruned_loss=0.05599, over 4812.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2521, pruned_loss=0.05416, over 958609.47 frames. ], batch size: 39, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:58:20,345 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4943, 1.6744, 1.3404, 1.0064, 1.1567, 1.0748, 1.2373, 1.0582], + device='cuda:2'), covar=tensor([0.1934, 0.1293, 0.1616, 0.1921, 0.2396, 0.2241, 0.1153, 0.2117], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0212, 0.0168, 0.0204, 0.0199, 0.0185, 0.0156, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 12:58:23,137 INFO [optim.py:369] (2/7) 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 12:58:30,242 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6841, 3.6647, 2.8357, 4.2881, 3.5671, 3.6661, 1.6207, 3.6418], + device='cuda:2'), covar=tensor([0.1633, 0.1282, 0.3545, 0.1425, 0.3245, 0.1612, 0.5354, 0.2396], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0216, 0.0248, 0.0308, 0.0298, 0.0247, 0.0271, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 12:58:47,506 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1553, 2.4566, 0.9790, 1.4715, 1.9819, 1.4001, 3.1276, 1.7936], + device='cuda:2'), covar=tensor([0.0632, 0.0554, 0.0715, 0.1238, 0.0444, 0.0909, 0.0214, 0.0636], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 12:58:57,661 INFO [finetune.py:976] (2/7) Epoch 19, batch 3500, loss[loss=0.1608, simple_loss=0.2361, pruned_loss=0.04273, over 4790.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.248, pruned_loss=0.05284, over 957880.91 frames. ], batch size: 29, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:59:14,547 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106624.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 12:59:36,352 INFO [finetune.py:976] (2/7) Epoch 19, batch 3550, loss[loss=0.1819, simple_loss=0.2479, pruned_loss=0.05791, over 4823.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.246, pruned_loss=0.05264, over 958487.88 frames. ], batch size: 38, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 12:59:41,156 INFO [optim.py:369] (2/7) 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 12:59:46,101 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-27 13:00:05,461 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:00:14,735 INFO [finetune.py:976] (2/7) Epoch 19, batch 3600, loss[loss=0.1501, simple_loss=0.2166, pruned_loss=0.04183, over 4790.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2431, pruned_loss=0.05148, over 957633.86 frames. ], batch size: 25, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:00:58,214 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106733.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:01:14,587 INFO [finetune.py:976] (2/7) Epoch 19, batch 3650, loss[loss=0.1955, simple_loss=0.2772, pruned_loss=0.05695, over 4807.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2466, pruned_loss=0.05318, over 956714.99 frames. ], batch size: 41, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:01:19,420 INFO [optim.py:369] (2/7) 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 13:01:29,176 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6394, 1.3476, 4.3777, 4.1174, 3.8132, 4.1404, 4.0269, 3.8859], + device='cuda:2'), covar=tensor([0.6584, 0.5886, 0.1028, 0.1476, 0.1126, 0.1662, 0.1496, 0.1446], + device='cuda:2'), in_proj_covar=tensor([0.0302, 0.0298, 0.0401, 0.0398, 0.0343, 0.0398, 0.0307, 0.0360], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:01:35,793 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106781.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:01:48,135 INFO [finetune.py:976] (2/7) Epoch 19, batch 3700, loss[loss=0.1462, simple_loss=0.2261, pruned_loss=0.03311, over 4798.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2503, pruned_loss=0.05377, over 956486.56 frames. ], batch size: 29, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:01:55,078 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106810.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:02:10,591 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0429, 1.5444, 1.9712, 2.1222, 1.8748, 1.5433, 1.0860, 1.6340], + device='cuda:2'), covar=tensor([0.3222, 0.3005, 0.1643, 0.2367, 0.2430, 0.2488, 0.4393, 0.1939], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0245, 0.0226, 0.0315, 0.0219, 0.0231, 0.0227, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 13:02:18,184 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7561, 1.3599, 1.8804, 2.2568, 1.8551, 1.7403, 1.7983, 1.7539], + device='cuda:2'), covar=tensor([0.4561, 0.6453, 0.6392, 0.5479, 0.5642, 0.7527, 0.8261, 0.8690], + device='cuda:2'), in_proj_covar=tensor([0.0427, 0.0409, 0.0502, 0.0505, 0.0454, 0.0482, 0.0489, 0.0494], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:02:22,251 INFO [finetune.py:976] (2/7) Epoch 19, batch 3750, loss[loss=0.1302, simple_loss=0.1984, pruned_loss=0.03104, over 4030.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2497, pruned_loss=0.0531, over 954882.37 frames. ], batch size: 17, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:02:27,098 INFO [optim.py:369] (2/7) 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 13:02:55,380 INFO [finetune.py:976] (2/7) Epoch 19, batch 3800, loss[loss=0.1641, simple_loss=0.2141, pruned_loss=0.05701, over 4468.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2502, pruned_loss=0.0531, over 954258.36 frames. ], batch size: 19, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:03:02,845 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0338, 1.8934, 2.2092, 2.3693, 2.1045, 1.8956, 2.0422, 2.0237], + device='cuda:2'), covar=tensor([0.4574, 0.6192, 0.6461, 0.5469, 0.5762, 0.8236, 0.8580, 0.8447], + device='cuda:2'), in_proj_covar=tensor([0.0425, 0.0407, 0.0500, 0.0504, 0.0452, 0.0479, 0.0486, 0.0492], + device='cuda:2'), out_proj_covar=tensor([1.0219e-04, 9.9982e-05, 1.1236e-04, 1.2025e-04, 1.0830e-04, 1.1483e-04, + 1.1508e-04, 1.1571e-04], device='cuda:2') +2023-04-27 13:03:10,668 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106924.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:03:20,095 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 13:03:26,501 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3468, 3.3188, 2.4418, 3.8375, 3.2876, 3.3131, 1.4218, 3.2507], + device='cuda:2'), covar=tensor([0.1926, 0.1362, 0.3376, 0.2367, 0.3446, 0.2007, 0.5962, 0.2583], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0215, 0.0248, 0.0307, 0.0297, 0.0247, 0.0270, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:03:27,653 INFO [finetune.py:976] (2/7) Epoch 19, batch 3850, loss[loss=0.1443, simple_loss=0.2105, pruned_loss=0.03906, over 4904.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2484, pruned_loss=0.05257, over 954469.31 frames. ], batch size: 43, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:03:33,087 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106972.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:03:56,025 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106992.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:04:00,616 INFO [finetune.py:976] (2/7) Epoch 19, batch 3900, loss[loss=0.1389, simple_loss=0.2075, pruned_loss=0.03517, over 4822.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2452, pruned_loss=0.05144, over 953511.26 frames. ], batch size: 30, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:04:33,283 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:04:34,528 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2009, 2.6748, 1.1446, 1.5296, 2.2978, 1.4298, 3.6389, 1.9749], + device='cuda:2'), covar=tensor([0.0651, 0.0658, 0.0768, 0.1250, 0.0463, 0.0969, 0.0355, 0.0616], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 13:04:41,523 INFO [finetune.py:976] (2/7) Epoch 19, batch 3950, loss[loss=0.164, simple_loss=0.2296, pruned_loss=0.04918, over 4894.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2416, pruned_loss=0.05033, over 954289.36 frames. ], batch size: 32, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:04:54,082 INFO [optim.py:369] (2/7) 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 13:05:24,400 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.19 vs. limit=5.0 +2023-04-27 13:05:26,564 INFO [finetune.py:976] (2/7) Epoch 19, batch 4000, loss[loss=0.2126, simple_loss=0.2807, pruned_loss=0.07224, over 4814.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2413, pruned_loss=0.05041, over 954053.75 frames. ], batch size: 51, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:05:35,379 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:05:43,835 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.9930, 4.9936, 3.2062, 5.6750, 4.9502, 5.0634, 2.3018, 4.8414], + device='cuda:2'), covar=tensor([0.1488, 0.1043, 0.2902, 0.0919, 0.3768, 0.1351, 0.5253, 0.1938], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0216, 0.0249, 0.0307, 0.0299, 0.0247, 0.0271, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:06:15,748 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107140.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:06:28,916 INFO [finetune.py:976] (2/7) Epoch 19, batch 4050, loss[loss=0.1798, simple_loss=0.2501, pruned_loss=0.05474, over 4775.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2441, pruned_loss=0.05108, over 951325.54 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:06:30,800 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:07:15,574 INFO [finetune.py:976] (2/7) Epoch 19, batch 4100, loss[loss=0.1665, simple_loss=0.2455, pruned_loss=0.04378, over 4705.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.246, pruned_loss=0.0514, over 950322.87 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:07:17,384 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107201.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:07:26,151 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107213.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:07:48,304 INFO [finetune.py:976] (2/7) Epoch 19, batch 4150, loss[loss=0.2191, simple_loss=0.2818, pruned_loss=0.07823, over 4734.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2477, pruned_loss=0.05246, over 950372.67 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:07:54,579 INFO [optim.py:369] (2/7) 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 13:07:55,446 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 13:08:11,093 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2859, 3.1500, 0.9671, 1.7071, 1.6760, 2.2629, 1.8424, 0.9486], + device='cuda:2'), covar=tensor([0.1648, 0.1417, 0.2008, 0.1455, 0.1312, 0.1241, 0.1769, 0.2267], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0242, 0.0138, 0.0119, 0.0132, 0.0152, 0.0116, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:08:22,019 INFO [finetune.py:976] (2/7) Epoch 19, batch 4200, loss[loss=0.1515, simple_loss=0.2292, pruned_loss=0.03692, over 4812.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2486, pruned_loss=0.05224, over 949227.50 frames. ], batch size: 45, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:08:23,866 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6444, 1.3767, 1.8751, 1.9938, 1.7507, 1.6458, 1.6804, 1.7492], + device='cuda:2'), covar=tensor([0.6153, 0.7965, 0.8298, 0.9336, 0.7612, 1.0822, 1.0677, 1.1520], + device='cuda:2'), in_proj_covar=tensor([0.0428, 0.0410, 0.0505, 0.0507, 0.0456, 0.0484, 0.0489, 0.0496], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:08:37,205 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6004, 1.7972, 1.6940, 2.4430, 2.4384, 2.0902, 2.0289, 1.6988], + device='cuda:2'), covar=tensor([0.1465, 0.1783, 0.2059, 0.1434, 0.1471, 0.1677, 0.2016, 0.2313], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0310, 0.0347, 0.0287, 0.0325, 0.0306, 0.0297, 0.0368], + device='cuda:2'), out_proj_covar=tensor([6.2810e-05, 6.4277e-05, 7.3477e-05, 5.8090e-05, 6.7488e-05, 6.4184e-05, + 6.2203e-05, 7.8305e-05], device='cuda:2') +2023-04-27 13:08:48,325 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 13:08:55,959 INFO [finetune.py:976] (2/7) Epoch 19, batch 4250, loss[loss=0.1566, simple_loss=0.2263, pruned_loss=0.04342, over 4809.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2457, pruned_loss=0.05145, over 950498.90 frames. ], batch size: 41, lr: 3.28e-03, grad_scale: 64.0 +2023-04-27 13:09:01,364 INFO [optim.py:369] (2/7) 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 13:09:14,778 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8218, 2.1502, 1.7220, 1.5715, 1.3703, 1.3741, 1.7009, 1.2539], + device='cuda:2'), covar=tensor([0.1778, 0.1233, 0.1441, 0.1734, 0.2341, 0.1952, 0.1118, 0.2122], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0203, 0.0199, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 13:09:17,110 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7039, 3.5649, 0.8907, 1.8567, 2.0489, 2.5901, 2.0151, 1.0325], + device='cuda:2'), covar=tensor([0.1252, 0.0727, 0.2011, 0.1195, 0.0998, 0.0900, 0.1441, 0.2044], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0242, 0.0137, 0.0119, 0.0132, 0.0152, 0.0116, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:09:29,132 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.6677, 4.5561, 3.1325, 5.3997, 4.7341, 4.6657, 2.2355, 4.5984], + device='cuda:2'), covar=tensor([0.1448, 0.1080, 0.3080, 0.0900, 0.2522, 0.1605, 0.5285, 0.1997], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0216, 0.0249, 0.0308, 0.0298, 0.0247, 0.0272, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:09:29,660 INFO [finetune.py:976] (2/7) Epoch 19, batch 4300, loss[loss=0.1646, simple_loss=0.2299, pruned_loss=0.04966, over 4826.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2452, pruned_loss=0.05207, over 953267.83 frames. ], batch size: 30, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:09:42,507 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2979, 1.7185, 2.0864, 2.4439, 2.1047, 1.7181, 1.2250, 1.9039], + device='cuda:2'), covar=tensor([0.3203, 0.3240, 0.1700, 0.2316, 0.2604, 0.2571, 0.4589, 0.1957], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0247, 0.0228, 0.0317, 0.0220, 0.0232, 0.0229, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 13:09:54,397 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6346, 3.5904, 2.6543, 4.2508, 3.6670, 3.6603, 1.5208, 3.5173], + device='cuda:2'), covar=tensor([0.1642, 0.1282, 0.2674, 0.1755, 0.2666, 0.1817, 0.5835, 0.2450], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0216, 0.0249, 0.0308, 0.0298, 0.0247, 0.0272, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:09:55,672 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1424, 1.8655, 2.0977, 2.4700, 2.5188, 2.0603, 1.5984, 2.3164], + device='cuda:2'), covar=tensor([0.0766, 0.1053, 0.0683, 0.0567, 0.0516, 0.0811, 0.0775, 0.0513], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0200, 0.0182, 0.0172, 0.0175, 0.0181, 0.0150, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:10:14,628 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6701, 1.3114, 4.6714, 4.4192, 4.0446, 4.5317, 4.3674, 4.1165], + device='cuda:2'), covar=tensor([0.6603, 0.5959, 0.0868, 0.1323, 0.0912, 0.1389, 0.0908, 0.1222], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0302, 0.0406, 0.0404, 0.0348, 0.0405, 0.0312, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:10:31,468 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-04-27 13:10:31,842 INFO [finetune.py:976] (2/7) Epoch 19, batch 4350, loss[loss=0.1755, simple_loss=0.2378, pruned_loss=0.0566, over 4835.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2412, pruned_loss=0.05082, over 955308.98 frames. ], batch size: 30, lr: 3.28e-03, grad_scale: 32.0 +2023-04-27 13:10:37,327 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107496.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:11:04,779 INFO [finetune.py:976] (2/7) Epoch 19, batch 4400, loss[loss=0.2013, simple_loss=0.2737, pruned_loss=0.06442, over 4827.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2428, pruned_loss=0.05172, over 954030.65 frames. ], batch size: 51, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:11:10,415 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107508.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:11:11,650 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107510.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:11:49,533 INFO [finetune.py:976] (2/7) Epoch 19, batch 4450, loss[loss=0.2005, simple_loss=0.2754, pruned_loss=0.06284, over 4907.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.247, pruned_loss=0.05297, over 954531.63 frames. ], batch size: 35, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:11:59,062 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.54 vs. limit=2.0 +2023-04-27 13:12:00,085 INFO [optim.py:369] (2/7) 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 13:12:12,165 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 13:12:18,630 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107571.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:12:43,863 INFO [finetune.py:976] (2/7) Epoch 19, batch 4500, loss[loss=0.1386, simple_loss=0.2181, pruned_loss=0.02948, over 4772.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2489, pruned_loss=0.05333, over 953759.50 frames. ], batch size: 54, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:13:17,596 INFO [finetune.py:976] (2/7) Epoch 19, batch 4550, loss[loss=0.2196, simple_loss=0.2908, pruned_loss=0.07424, over 4807.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2503, pruned_loss=0.05351, over 954573.83 frames. ], batch size: 45, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:13:20,733 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7206, 1.4416, 1.6678, 2.0510, 2.0355, 1.7204, 1.3793, 1.9127], + device='cuda:2'), covar=tensor([0.0824, 0.1334, 0.0781, 0.0589, 0.0593, 0.0746, 0.0746, 0.0534], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0203, 0.0183, 0.0173, 0.0177, 0.0182, 0.0152, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:13:23,033 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.9875, 4.3356, 0.7608, 2.2566, 2.4563, 2.6184, 2.4661, 0.9602], + device='cuda:2'), covar=tensor([0.1297, 0.0862, 0.2187, 0.1215, 0.1018, 0.1189, 0.1524, 0.2148], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0242, 0.0138, 0.0120, 0.0133, 0.0152, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:13:46,986 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9050, 2.4774, 1.1011, 1.7262, 2.3698, 1.8723, 1.7739, 1.9636], + device='cuda:2'), covar=tensor([0.0457, 0.0309, 0.0288, 0.0504, 0.0220, 0.0465, 0.0458, 0.0502], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 13:13:51,126 INFO [finetune.py:976] (2/7) Epoch 19, batch 4600, loss[loss=0.19, simple_loss=0.2561, pruned_loss=0.06197, over 4812.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2492, pruned_loss=0.0531, over 954439.16 frames. ], batch size: 38, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:13:51,355 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 13:14:04,029 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:14:16,219 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 4650, loss[loss=0.1419, simple_loss=0.2063, pruned_loss=0.03876, over 4832.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2472, pruned_loss=0.05292, over 954118.24 frames. ], batch size: 30, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:14:26,857 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6821, 1.8458, 0.6917, 1.3498, 1.8407, 1.5193, 1.4274, 1.5081], + device='cuda:2'), covar=tensor([0.0500, 0.0356, 0.0376, 0.0553, 0.0276, 0.0523, 0.0506, 0.0590], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0051], + device='cuda:2') +2023-04-27 13:14:29,768 INFO [optim.py:369] (2/7) 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 13:15:06,674 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107796.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 13:15:07,908 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107798.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:15:08,399 INFO [finetune.py:976] (2/7) Epoch 19, batch 4700, loss[loss=0.1707, simple_loss=0.2412, pruned_loss=0.05005, over 4861.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2437, pruned_loss=0.05165, over 956668.24 frames. ], batch size: 31, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:15:19,599 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107808.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:15:22,703 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7372, 2.0920, 1.7286, 1.5659, 1.2872, 1.3167, 1.7941, 1.2313], + device='cuda:2'), covar=tensor([0.1891, 0.1295, 0.1588, 0.1758, 0.2566, 0.2300, 0.1078, 0.2182], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0203, 0.0200, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 13:15:41,404 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6754, 1.2600, 1.3311, 1.3916, 1.8318, 1.4180, 1.1387, 1.3023], + device='cuda:2'), covar=tensor([0.1465, 0.1359, 0.1707, 0.1362, 0.0757, 0.1442, 0.1977, 0.2007], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0315, 0.0352, 0.0293, 0.0332, 0.0310, 0.0302, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4124e-05, 6.5322e-05, 7.4461e-05, 5.9252e-05, 6.8902e-05, 6.5189e-05, + 6.3278e-05, 7.9342e-05], device='cuda:2') +2023-04-27 13:15:41,533 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-27 13:15:48,344 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107844.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:15:52,873 INFO [finetune.py:976] (2/7) Epoch 19, batch 4750, loss[loss=0.1988, simple_loss=0.2718, pruned_loss=0.06285, over 4805.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2418, pruned_loss=0.05123, over 954850.43 frames. ], batch size: 45, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:15:57,097 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 13:16:09,258 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107866.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:16:18,791 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0095, 1.4021, 1.8104, 2.1870, 1.7710, 1.4553, 1.1264, 1.6228], + device='cuda:2'), covar=tensor([0.3307, 0.3570, 0.1865, 0.2109, 0.2532, 0.2638, 0.4250, 0.1997], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0246, 0.0226, 0.0315, 0.0218, 0.0232, 0.0227, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 13:16:53,225 INFO [finetune.py:976] (2/7) Epoch 19, batch 4800, loss[loss=0.1837, simple_loss=0.2581, pruned_loss=0.05465, over 4873.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2444, pruned_loss=0.05218, over 953250.74 frames. ], batch size: 31, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:17:31,834 INFO [finetune.py:976] (2/7) Epoch 19, batch 4850, loss[loss=0.1621, simple_loss=0.2367, pruned_loss=0.04376, over 4862.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2455, pruned_loss=0.05191, over 952607.64 frames. ], batch size: 34, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:17:37,756 INFO [optim.py:369] (2/7) 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 13:17:50,206 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-27 13:18:04,090 INFO [finetune.py:976] (2/7) Epoch 19, batch 4900, loss[loss=0.1387, simple_loss=0.2164, pruned_loss=0.03044, over 4709.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2477, pruned_loss=0.05294, over 951395.72 frames. ], batch size: 23, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:18:15,611 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 13:18:16,702 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108015.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:18:17,499 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-27 13:18:38,326 INFO [finetune.py:976] (2/7) Epoch 19, batch 4950, loss[loss=0.1832, simple_loss=0.2678, pruned_loss=0.04925, over 4847.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.247, pruned_loss=0.05206, over 951552.96 frames. ], batch size: 44, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:18:45,329 INFO [optim.py:369] (2/7) 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 13:19:07,754 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108093.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:19:11,333 INFO [finetune.py:976] (2/7) Epoch 19, batch 5000, loss[loss=0.1662, simple_loss=0.2431, pruned_loss=0.04465, over 4908.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2457, pruned_loss=0.05169, over 950773.73 frames. ], batch size: 36, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:19:44,761 INFO [finetune.py:976] (2/7) Epoch 19, batch 5050, loss[loss=0.1571, simple_loss=0.2211, pruned_loss=0.04657, over 4868.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2428, pruned_loss=0.05069, over 950319.22 frames. ], batch size: 31, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:19:51,166 INFO [optim.py:369] (2/7) 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 13:19:57,093 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:20:29,632 INFO [finetune.py:976] (2/7) Epoch 19, batch 5100, loss[loss=0.1571, simple_loss=0.2111, pruned_loss=0.05149, over 4068.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2406, pruned_loss=0.05018, over 950830.38 frames. ], batch size: 17, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:20:50,602 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108214.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:21:34,995 INFO [finetune.py:976] (2/7) Epoch 19, batch 5150, loss[loss=0.2162, simple_loss=0.2809, pruned_loss=0.07571, over 4726.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2392, pruned_loss=0.04935, over 950081.86 frames. ], batch size: 59, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:21:44,948 INFO [optim.py:369] (2/7) 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 13:22:29,163 INFO [finetune.py:976] (2/7) Epoch 19, batch 5200, loss[loss=0.2127, simple_loss=0.2841, pruned_loss=0.07066, over 4881.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2419, pruned_loss=0.05007, over 950249.65 frames. ], batch size: 32, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:22:41,861 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8690, 2.8990, 2.1918, 3.2891, 2.8557, 2.8453, 1.2394, 2.8299], + device='cuda:2'), covar=tensor([0.2130, 0.1518, 0.3334, 0.2804, 0.3516, 0.1973, 0.5432, 0.2769], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0214, 0.0248, 0.0304, 0.0296, 0.0247, 0.0271, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:22:51,797 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108315.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:23:04,121 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-27 13:23:05,664 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 5250, loss[loss=0.1436, simple_loss=0.2234, pruned_loss=0.03196, over 4376.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2442, pruned_loss=0.05063, over 951320.46 frames. ], batch size: 65, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:23:44,339 INFO [optim.py:369] (2/7) 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 13:23:53,058 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108363.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:19,496 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108386.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:23,676 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108393.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:27,186 INFO [finetune.py:976] (2/7) Epoch 19, batch 5300, loss[loss=0.1388, simple_loss=0.2192, pruned_loss=0.02926, over 4778.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2456, pruned_loss=0.0513, over 951765.00 frames. ], batch size: 26, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:24:29,889 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2382, 2.0322, 1.7203, 1.8142, 2.1351, 1.7097, 2.4926, 1.4802], + device='cuda:2'), covar=tensor([0.3651, 0.1992, 0.4373, 0.2962, 0.1706, 0.2463, 0.1785, 0.4217], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0345, 0.0426, 0.0353, 0.0382, 0.0377, 0.0372, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:24:50,798 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:24:50,830 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3830, 1.7361, 1.8179, 1.8985, 1.7222, 1.7899, 1.8928, 1.8377], + device='cuda:2'), covar=tensor([0.4354, 0.5524, 0.4575, 0.4287, 0.5375, 0.7430, 0.5364, 0.5015], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0372, 0.0321, 0.0334, 0.0345, 0.0394, 0.0357, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:24:56,168 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108441.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:25:01,041 INFO [finetune.py:976] (2/7) Epoch 19, batch 5350, loss[loss=0.1683, simple_loss=0.2457, pruned_loss=0.04546, over 4895.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2465, pruned_loss=0.05166, over 952327.80 frames. ], batch size: 32, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:25:06,499 INFO [optim.py:369] (2/7) 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 13:25:31,256 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108493.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:25:34,730 INFO [finetune.py:976] (2/7) Epoch 19, batch 5400, loss[loss=0.1796, simple_loss=0.2384, pruned_loss=0.06046, over 4239.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2448, pruned_loss=0.05135, over 952717.42 frames. ], batch size: 17, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:26:31,589 INFO [finetune.py:976] (2/7) Epoch 19, batch 5450, loss[loss=0.1321, simple_loss=0.2099, pruned_loss=0.02713, over 4757.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2431, pruned_loss=0.05113, over 954225.53 frames. ], batch size: 27, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:26:42,377 INFO [optim.py:369] (2/7) 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 13:27:06,393 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:27:25,453 INFO [finetune.py:976] (2/7) Epoch 19, batch 5500, loss[loss=0.2176, simple_loss=0.2734, pruned_loss=0.08085, over 4899.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2401, pruned_loss=0.04992, over 956609.83 frames. ], batch size: 32, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:27:51,070 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 19, batch 5550, loss[loss=0.175, simple_loss=0.2451, pruned_loss=0.05248, over 4897.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.242, pruned_loss=0.05075, over 956543.43 frames. ], batch size: 32, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:28:09,999 INFO [optim.py:369] (2/7) 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 13:28:41,414 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108681.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:28:43,771 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1849, 1.4280, 1.2561, 1.6522, 1.4959, 1.9739, 1.2878, 3.3835], + device='cuda:2'), covar=tensor([0.0667, 0.0851, 0.0877, 0.1241, 0.0703, 0.0534, 0.0823, 0.0144], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 13:29:03,284 INFO [finetune.py:976] (2/7) Epoch 19, batch 5600, loss[loss=0.1799, simple_loss=0.2552, pruned_loss=0.05227, over 4839.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2456, pruned_loss=0.05195, over 951073.78 frames. ], batch size: 47, lr: 3.27e-03, grad_scale: 32.0 +2023-04-27 13:29:26,979 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 13:30:00,556 INFO [finetune.py:976] (2/7) Epoch 19, batch 5650, loss[loss=0.1761, simple_loss=0.2618, pruned_loss=0.04522, over 4823.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2479, pruned_loss=0.05206, over 952985.95 frames. ], batch size: 25, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:30:17,473 INFO [optim.py:369] (2/7) 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 13:30:52,614 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108788.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 13:30:54,418 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108791.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:31:01,944 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1597, 2.5883, 1.0353, 1.4140, 1.8086, 1.2889, 3.0225, 1.7217], + device='cuda:2'), covar=tensor([0.0634, 0.0624, 0.0732, 0.1166, 0.0484, 0.0924, 0.0410, 0.0604], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 13:31:04,409 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 13:31:04,843 INFO [finetune.py:976] (2/7) Epoch 19, batch 5700, loss[loss=0.139, simple_loss=0.197, pruned_loss=0.04044, over 4314.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2447, pruned_loss=0.05165, over 933423.95 frames. ], batch size: 18, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:31:40,722 INFO [finetune.py:976] (2/7) Epoch 20, batch 0, loss[loss=0.1602, simple_loss=0.2379, pruned_loss=0.04121, over 4752.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.2379, pruned_loss=0.04121, over 4752.00 frames. ], batch size: 27, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:31:40,722 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 13:31:47,931 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1404, 2.5937, 1.0132, 1.4380, 1.7476, 1.3393, 3.0163, 1.7411], + device='cuda:2'), covar=tensor([0.0666, 0.0628, 0.0726, 0.1151, 0.0481, 0.0888, 0.0278, 0.0544], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 13:31:49,304 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3966, 1.5172, 1.8536, 1.9906, 1.9272, 2.1507, 1.8792, 1.9362], + device='cuda:2'), covar=tensor([0.3793, 0.5177, 0.4232, 0.4397, 0.5351, 0.6523, 0.5097, 0.4559], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0373, 0.0321, 0.0334, 0.0345, 0.0394, 0.0357, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:31:57,317 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6357MB +2023-04-27 13:32:03,724 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1322, 2.6680, 1.0226, 1.4804, 1.9220, 1.3128, 3.4164, 1.8492], + device='cuda:2'), covar=tensor([0.0696, 0.0632, 0.0784, 0.1153, 0.0511, 0.0967, 0.0294, 0.0592], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0066, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 13:32:13,977 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:32:17,454 INFO [optim.py:369] (2/7) 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 13:32:30,439 INFO [finetune.py:976] (2/7) Epoch 20, batch 50, loss[loss=0.1789, simple_loss=0.255, pruned_loss=0.05147, over 4885.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2533, pruned_loss=0.05435, over 217588.76 frames. ], batch size: 35, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:32:59,535 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1459, 2.7247, 2.1907, 2.3403, 1.6174, 1.5661, 2.3408, 1.6093], + device='cuda:2'), covar=tensor([0.1589, 0.1512, 0.1450, 0.1541, 0.2269, 0.1876, 0.0938, 0.1968], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0213, 0.0170, 0.0204, 0.0201, 0.0185, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 13:33:00,086 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2092, 2.9390, 0.9847, 1.5573, 1.6615, 2.1634, 1.7160, 0.9439], + device='cuda:2'), covar=tensor([0.1427, 0.0888, 0.1730, 0.1267, 0.1144, 0.0945, 0.1527, 0.1919], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0242, 0.0138, 0.0119, 0.0132, 0.0152, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:33:03,608 INFO [finetune.py:976] (2/7) Epoch 20, batch 100, loss[loss=0.1499, simple_loss=0.2239, pruned_loss=0.03798, over 4819.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2419, pruned_loss=0.04954, over 381176.05 frames. ], batch size: 33, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:33:08,744 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:33:23,934 INFO [optim.py:369] (2/7) 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 13:33:36,945 INFO [finetune.py:976] (2/7) Epoch 20, batch 150, loss[loss=0.1845, simple_loss=0.2604, pruned_loss=0.05428, over 4831.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2377, pruned_loss=0.04862, over 509830.30 frames. ], batch size: 40, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:33:39,446 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108981.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:33:41,577 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6837, 3.2495, 1.0791, 1.7735, 2.4605, 1.6785, 4.4381, 2.5386], + device='cuda:2'), covar=tensor([0.0633, 0.0724, 0.0866, 0.1212, 0.0520, 0.0977, 0.0251, 0.0508], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 13:34:09,750 INFO [finetune.py:976] (2/7) Epoch 20, batch 200, loss[loss=0.1512, simple_loss=0.2142, pruned_loss=0.04408, over 4903.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2366, pruned_loss=0.04865, over 610469.19 frames. ], batch size: 32, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:34:11,041 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109029.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:34:16,954 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109037.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:34:21,223 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.5252, 1.3976, 1.3581, 1.1744, 1.5135, 1.2603, 1.7409, 1.2610], + device='cuda:2'), covar=tensor([0.3316, 0.1704, 0.4974, 0.2315, 0.1377, 0.1992, 0.1489, 0.4978], + device='cuda:2'), in_proj_covar=tensor([0.0334, 0.0341, 0.0421, 0.0347, 0.0377, 0.0371, 0.0368, 0.0413], + device='cuda:2'), out_proj_covar=tensor([9.9277e-05, 1.0228e-04, 1.2802e-04, 1.0482e-04, 1.1246e-04, 1.1101e-04, + 1.0855e-04, 1.2483e-04], device='cuda:2') +2023-04-27 13:34:29,595 INFO [optim.py:369] (2/7) 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 13:34:42,777 INFO [finetune.py:976] (2/7) Epoch 20, batch 250, loss[loss=0.2115, simple_loss=0.2709, pruned_loss=0.07604, over 4903.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2423, pruned_loss=0.05106, over 686867.17 frames. ], batch size: 35, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:35:02,399 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:35:02,509 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-04-27 13:35:13,174 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109098.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:35:30,782 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6549, 3.6190, 1.0812, 1.8395, 1.9646, 2.4372, 2.0254, 1.0371], + device='cuda:2'), covar=tensor([0.1314, 0.0960, 0.1940, 0.1311, 0.1026, 0.1068, 0.1605, 0.1927], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0241, 0.0138, 0.0119, 0.0132, 0.0152, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:35:31,287 INFO [finetune.py:976] (2/7) Epoch 20, batch 300, loss[loss=0.1571, simple_loss=0.2224, pruned_loss=0.04588, over 4719.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2468, pruned_loss=0.05204, over 745359.22 frames. ], batch size: 23, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:35:49,173 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:35:55,689 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7587, 1.3974, 1.8628, 2.2638, 1.8597, 1.6976, 1.7413, 1.7482], + device='cuda:2'), covar=tensor([0.4471, 0.6545, 0.6667, 0.5595, 0.5647, 0.7973, 0.8206, 0.9447], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0410, 0.0506, 0.0507, 0.0456, 0.0485, 0.0492, 0.0498], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:35:56,827 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109147.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:35:59,522 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 13:36:03,479 INFO [optim.py:369] (2/7) 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 13:36:25,686 INFO [finetune.py:976] (2/7) Epoch 20, batch 350, loss[loss=0.1463, simple_loss=0.2, pruned_loss=0.04628, over 4538.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2481, pruned_loss=0.05303, over 791712.98 frames. ], batch size: 20, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:36:47,071 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109194.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:36:47,721 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6344, 1.9137, 1.9492, 2.0317, 1.8994, 1.9850, 2.0744, 2.0725], + device='cuda:2'), covar=tensor([0.3715, 0.5520, 0.4795, 0.4495, 0.5704, 0.7655, 0.5527, 0.5049], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0373, 0.0323, 0.0336, 0.0347, 0.0396, 0.0358, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:37:21,407 INFO [finetune.py:976] (2/7) Epoch 20, batch 400, loss[loss=0.1245, simple_loss=0.1891, pruned_loss=0.02991, over 4161.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2484, pruned_loss=0.05257, over 827301.34 frames. ], batch size: 17, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:37:31,615 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109233.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:03,868 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 13:38:16,306 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9135, 2.4979, 1.9386, 1.8750, 1.3794, 1.4607, 1.9727, 1.3414], + device='cuda:2'), covar=tensor([0.1723, 0.1428, 0.1426, 0.1613, 0.2441, 0.2004, 0.1047, 0.2155], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0211, 0.0168, 0.0202, 0.0199, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 13:38:22,933 INFO [finetune.py:976] (2/7) Epoch 20, batch 450, loss[loss=0.1444, simple_loss=0.2124, pruned_loss=0.03823, over 4817.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2463, pruned_loss=0.05167, over 856049.42 frames. ], batch size: 38, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:38:24,196 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6373, 3.6870, 2.7112, 4.3397, 3.6600, 3.6517, 1.7535, 3.6500], + device='cuda:2'), covar=tensor([0.1883, 0.1250, 0.3297, 0.1276, 0.3094, 0.1721, 0.5450, 0.2139], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0214, 0.0248, 0.0304, 0.0294, 0.0246, 0.0270, 0.0269], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:38:25,927 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109281.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:34,961 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:50,556 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109318.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:38:55,898 INFO [finetune.py:976] (2/7) Epoch 20, batch 500, loss[loss=0.1972, simple_loss=0.2577, pruned_loss=0.06834, over 4821.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2432, pruned_loss=0.05086, over 879335.06 frames. ], batch size: 40, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:39:16,072 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109355.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:39:17,760 INFO [optim.py:369] (2/7) 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 13:39:24,572 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5445, 3.9603, 0.7652, 1.9528, 2.0657, 2.7492, 2.3060, 0.9056], + device='cuda:2'), covar=tensor([0.1636, 0.1548, 0.2497, 0.1522, 0.1240, 0.1207, 0.1672, 0.2595], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0241, 0.0138, 0.0119, 0.0132, 0.0151, 0.0116, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:39:29,353 INFO [finetune.py:976] (2/7) Epoch 20, batch 550, loss[loss=0.1541, simple_loss=0.2249, pruned_loss=0.04159, over 4827.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2402, pruned_loss=0.05017, over 894454.72 frames. ], batch size: 25, lr: 3.26e-03, grad_scale: 32.0 +2023-04-27 13:39:30,731 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109379.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 13:39:40,650 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109393.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:40:02,927 INFO [finetune.py:976] (2/7) Epoch 20, batch 600, loss[loss=0.2203, simple_loss=0.3045, pruned_loss=0.06805, over 4841.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2421, pruned_loss=0.05122, over 908481.44 frames. ], batch size: 47, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:40:16,606 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109447.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:40:24,267 INFO [optim.py:369] (2/7) 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 13:40:28,569 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6731, 1.3743, 1.3427, 1.3954, 1.8418, 1.4739, 1.1995, 1.2951], + device='cuda:2'), covar=tensor([0.1585, 0.1293, 0.1879, 0.1600, 0.0902, 0.1451, 0.1971, 0.2203], + device='cuda:2'), in_proj_covar=tensor([0.0305, 0.0308, 0.0346, 0.0287, 0.0323, 0.0305, 0.0297, 0.0367], + device='cuda:2'), out_proj_covar=tensor([6.2788e-05, 6.3800e-05, 7.3288e-05, 5.8051e-05, 6.6753e-05, 6.4074e-05, + 6.2323e-05, 7.8065e-05], device='cuda:2') +2023-04-27 13:40:36,278 INFO [finetune.py:976] (2/7) Epoch 20, batch 650, loss[loss=0.2212, simple_loss=0.2835, pruned_loss=0.07945, over 4930.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2458, pruned_loss=0.05224, over 919906.44 frames. ], batch size: 42, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:40:48,774 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:40:54,348 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-27 13:41:10,002 INFO [finetune.py:976] (2/7) Epoch 20, batch 700, loss[loss=0.1626, simple_loss=0.2475, pruned_loss=0.03888, over 4918.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2476, pruned_loss=0.05231, over 927158.37 frames. ], batch size: 38, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:41:25,582 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109550.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:41:30,317 INFO [optim.py:369] (2/7) 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 13:41:43,810 INFO [finetune.py:976] (2/7) Epoch 20, batch 750, loss[loss=0.206, simple_loss=0.2723, pruned_loss=0.06981, over 4815.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2501, pruned_loss=0.05347, over 932222.23 frames. ], batch size: 39, lr: 3.26e-03, grad_scale: 64.0 +2023-04-27 13:41:45,102 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 800, loss[loss=0.1562, simple_loss=0.2288, pruned_loss=0.04174, over 4867.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2481, pruned_loss=0.05204, over 936362.68 frames. ], batch size: 31, lr: 3.25e-03, grad_scale: 64.0 +2023-04-27 13:43:07,639 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109640.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:43:19,834 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:43:27,908 INFO [optim.py:369] (2/7) 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 13:43:28,670 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:43:50,498 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109674.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:43:52,232 INFO [finetune.py:976] (2/7) Epoch 20, batch 850, loss[loss=0.1588, simple_loss=0.2072, pruned_loss=0.05518, over 4274.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2463, pruned_loss=0.05153, over 940333.70 frames. ], batch size: 18, lr: 3.25e-03, grad_scale: 64.0 +2023-04-27 13:44:05,033 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0725, 2.5109, 1.3580, 1.6872, 2.4086, 1.9549, 1.7971, 1.9022], + device='cuda:2'), covar=tensor([0.0455, 0.0317, 0.0269, 0.0521, 0.0211, 0.0459, 0.0488, 0.0519], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0051], + device='cuda:2') +2023-04-27 13:44:08,052 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109693.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:44:26,900 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109720.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:44:31,595 INFO [finetune.py:976] (2/7) Epoch 20, batch 900, loss[loss=0.1613, simple_loss=0.2314, pruned_loss=0.04559, over 4819.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2421, pruned_loss=0.05004, over 944320.62 frames. ], batch size: 41, lr: 3.25e-03, grad_scale: 64.0 +2023-04-27 13:44:37,325 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 13:44:40,145 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109741.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 13:44:45,675 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4071, 2.6472, 2.1285, 2.3199, 2.6288, 2.3221, 3.4813, 2.0706], + device='cuda:2'), covar=tensor([0.3930, 0.1926, 0.4546, 0.3328, 0.1896, 0.2427, 0.1552, 0.4094], + device='cuda:2'), in_proj_covar=tensor([0.0334, 0.0341, 0.0420, 0.0348, 0.0376, 0.0371, 0.0367, 0.0411], + device='cuda:2'), out_proj_covar=tensor([9.9184e-05, 1.0235e-04, 1.2754e-04, 1.0502e-04, 1.1225e-04, 1.1073e-04, + 1.0803e-04, 1.2432e-04], device='cuda:2') +2023-04-27 13:44:50,920 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 20, batch 950, loss[loss=0.2005, simple_loss=0.2691, pruned_loss=0.06596, over 4821.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2405, pruned_loss=0.04995, over 946832.31 frames. ], batch size: 30, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:45:38,010 INFO [finetune.py:976] (2/7) Epoch 20, batch 1000, loss[loss=0.1633, simple_loss=0.2195, pruned_loss=0.05353, over 4725.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2422, pruned_loss=0.05049, over 950535.46 frames. ], batch size: 23, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:45:52,566 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 13:46:03,382 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 13:46:08,146 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3745, 1.3136, 1.6075, 1.6719, 1.2828, 1.2298, 1.3058, 0.9090], + device='cuda:2'), covar=tensor([0.0555, 0.0547, 0.0402, 0.0527, 0.0644, 0.0971, 0.0604, 0.0583], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0074, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:46:11,418 INFO [finetune.py:976] (2/7) Epoch 20, batch 1050, loss[loss=0.1679, simple_loss=0.244, pruned_loss=0.04587, over 4914.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.245, pruned_loss=0.0511, over 953464.31 frames. ], batch size: 36, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:46:16,812 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8296, 2.8077, 2.2292, 3.2490, 2.7947, 2.8205, 1.1846, 2.8102], + device='cuda:2'), covar=tensor([0.2164, 0.1799, 0.3581, 0.3048, 0.3488, 0.2031, 0.5357, 0.2813], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0215, 0.0249, 0.0304, 0.0295, 0.0246, 0.0272, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:46:21,140 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.65 vs. limit=5.0 +2023-04-27 13:46:25,235 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109898.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:46:37,214 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8312, 2.4936, 2.0100, 2.4298, 1.7698, 2.1542, 2.1081, 1.6992], + device='cuda:2'), covar=tensor([0.2269, 0.1039, 0.0833, 0.1198, 0.3241, 0.1109, 0.1915, 0.2482], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0304, 0.0218, 0.0280, 0.0314, 0.0260, 0.0252, 0.0265], + device='cuda:2'), out_proj_covar=tensor([1.1619e-04, 1.2078e-04, 8.6318e-05, 1.1092e-04, 1.2746e-04, 1.0315e-04, + 1.0188e-04, 1.0498e-04], device='cuda:2') +2023-04-27 13:46:43,731 INFO [finetune.py:976] (2/7) Epoch 20, batch 1100, loss[loss=0.1408, simple_loss=0.217, pruned_loss=0.03233, over 4776.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2467, pruned_loss=0.05135, over 955214.75 frames. ], batch size: 26, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:46:50,144 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109935.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:46:59,850 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:05,248 INFO [optim.py:369] (2/7) 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 13:47:07,441 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 13:47:16,128 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:17,879 INFO [finetune.py:976] (2/7) Epoch 20, batch 1150, loss[loss=0.1743, simple_loss=0.2472, pruned_loss=0.05069, over 4870.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2481, pruned_loss=0.05154, over 955169.59 frames. ], batch size: 34, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:47:40,683 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:46,318 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3135, 1.8195, 2.1741, 2.7067, 2.1256, 1.7632, 1.4649, 1.9827], + device='cuda:2'), covar=tensor([0.3467, 0.3169, 0.1764, 0.2625, 0.2965, 0.2799, 0.4284, 0.2170], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0245, 0.0225, 0.0314, 0.0217, 0.0231, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 13:47:52,883 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110015.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:54,548 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110017.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:47:57,538 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110022.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:48:00,998 INFO [finetune.py:976] (2/7) Epoch 20, batch 1200, loss[loss=0.1942, simple_loss=0.2567, pruned_loss=0.06588, over 4865.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2472, pruned_loss=0.05113, over 956278.77 frames. ], batch size: 34, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:48:17,820 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110044.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:48:37,464 INFO [optim.py:369] (2/7) 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 13:49:01,339 INFO [finetune.py:976] (2/7) Epoch 20, batch 1250, loss[loss=0.1749, simple_loss=0.2474, pruned_loss=0.05125, over 4840.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2452, pruned_loss=0.05086, over 958206.65 frames. ], batch size: 30, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:49:02,102 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110078.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:25,661 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110105.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:28,706 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:31,777 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:49:40,407 INFO [finetune.py:976] (2/7) Epoch 20, batch 1300, loss[loss=0.1754, simple_loss=0.2373, pruned_loss=0.05673, over 4830.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2421, pruned_loss=0.0501, over 959549.60 frames. ], batch size: 33, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:50:01,795 INFO [optim.py:369] (2/7) 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 13:50:09,733 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:50:11,550 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 13:50:13,288 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:50:13,762 INFO [finetune.py:976] (2/7) Epoch 20, batch 1350, loss[loss=0.1618, simple_loss=0.2377, pruned_loss=0.0429, over 4743.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2415, pruned_loss=0.04995, over 953758.19 frames. ], batch size: 59, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:50:47,121 INFO [finetune.py:976] (2/7) Epoch 20, batch 1400, loss[loss=0.1419, simple_loss=0.2232, pruned_loss=0.03026, over 4764.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2445, pruned_loss=0.05082, over 952081.98 frames. ], batch size: 27, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:50:52,593 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110235.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:50:53,266 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.98 vs. limit=5.0 +2023-04-27 13:51:09,025 INFO [optim.py:369] (2/7) 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 13:51:12,275 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 13:51:19,999 INFO [finetune.py:976] (2/7) Epoch 20, batch 1450, loss[loss=0.133, simple_loss=0.2018, pruned_loss=0.03208, over 4599.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2456, pruned_loss=0.05051, over 953506.17 frames. ], batch size: 20, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:51:25,116 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:25,157 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:34,614 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4366, 3.4535, 2.6385, 4.0725, 3.4314, 3.4547, 1.7184, 3.4474], + device='cuda:2'), covar=tensor([0.2033, 0.1270, 0.3418, 0.1957, 0.2927, 0.1832, 0.5346, 0.2654], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0215, 0.0249, 0.0304, 0.0295, 0.0246, 0.0272, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:51:38,782 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110303.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:46,142 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-04-27 13:51:46,547 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110315.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:48,407 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110318.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:51:53,750 INFO [finetune.py:976] (2/7) Epoch 20, batch 1500, loss[loss=0.1786, simple_loss=0.2429, pruned_loss=0.05708, over 4901.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2465, pruned_loss=0.05073, over 954691.06 frames. ], batch size: 36, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:52:05,636 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110344.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:52:16,150 INFO [optim.py:369] (2/7) 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 13:52:19,127 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110363.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:52:19,792 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110364.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:52:25,266 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 1550, loss[loss=0.1855, simple_loss=0.2494, pruned_loss=0.06082, over 4929.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2474, pruned_loss=0.05081, over 954751.52 frames. ], batch size: 42, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:52:29,064 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110379.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:53:00,008 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110400.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:53:15,141 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-27 13:53:22,804 INFO [finetune.py:976] (2/7) Epoch 20, batch 1600, loss[loss=0.1712, simple_loss=0.2429, pruned_loss=0.04969, over 4867.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.245, pruned_loss=0.05065, over 956192.42 frames. ], batch size: 31, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:54:08,137 INFO [optim.py:369] (2/7) 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 13:54:17,902 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:54:21,383 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110471.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:54:30,173 INFO [finetune.py:976] (2/7) Epoch 20, batch 1650, loss[loss=0.1385, simple_loss=0.21, pruned_loss=0.03349, over 4922.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2437, pruned_loss=0.05091, over 955098.26 frames. ], batch size: 43, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:54:51,130 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6233, 1.3152, 4.2614, 3.9506, 3.7509, 4.0865, 3.9938, 3.7620], + device='cuda:2'), covar=tensor([0.6637, 0.6018, 0.0975, 0.1739, 0.1069, 0.1610, 0.1649, 0.1497], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0301, 0.0402, 0.0404, 0.0347, 0.0405, 0.0311, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:54:59,476 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-04-27 13:55:07,693 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6136, 2.7513, 2.3521, 2.4946, 2.8157, 2.3687, 3.7591, 2.0074], + device='cuda:2'), covar=tensor([0.3524, 0.2092, 0.3606, 0.3145, 0.1619, 0.2546, 0.1475, 0.3941], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0346, 0.0424, 0.0350, 0.0381, 0.0374, 0.0371, 0.0418], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:55:09,361 INFO [finetune.py:976] (2/7) Epoch 20, batch 1700, loss[loss=0.1653, simple_loss=0.2366, pruned_loss=0.04702, over 4855.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2412, pruned_loss=0.05008, over 956431.15 frames. ], batch size: 49, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:55:25,901 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0202, 3.9476, 2.7551, 4.6564, 3.9779, 4.0380, 1.7497, 3.9330], + device='cuda:2'), covar=tensor([0.1607, 0.1046, 0.3065, 0.1355, 0.3947, 0.1694, 0.5740, 0.2247], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0215, 0.0249, 0.0303, 0.0294, 0.0246, 0.0271, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 13:55:31,158 INFO [optim.py:369] (2/7) 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 13:55:42,006 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0454, 1.3976, 1.2511, 1.6196, 1.5017, 1.8331, 1.3203, 3.4007], + device='cuda:2'), covar=tensor([0.0661, 0.0851, 0.0819, 0.1213, 0.0635, 0.0556, 0.0772, 0.0167], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 13:55:43,125 INFO [finetune.py:976] (2/7) Epoch 20, batch 1750, loss[loss=0.2094, simple_loss=0.2775, pruned_loss=0.07062, over 4852.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2444, pruned_loss=0.05164, over 958404.94 frames. ], batch size: 44, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:56:06,131 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110611.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:17,234 INFO [finetune.py:976] (2/7) Epoch 20, batch 1800, loss[loss=0.161, simple_loss=0.2314, pruned_loss=0.04531, over 4744.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2465, pruned_loss=0.05166, over 959628.16 frames. ], batch size: 28, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:56:17,340 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5200, 1.4731, 0.5858, 1.2322, 1.3679, 1.3796, 1.2944, 1.3477], + device='cuda:2'), covar=tensor([0.0484, 0.0340, 0.0411, 0.0536, 0.0324, 0.0478, 0.0479, 0.0505], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0020, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 13:56:24,706 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110639.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:31,175 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9976, 1.4241, 1.6978, 1.7034, 2.1652, 1.7901, 1.5054, 1.5562], + device='cuda:2'), covar=tensor([0.1630, 0.1608, 0.1773, 0.1281, 0.0854, 0.1454, 0.2072, 0.2246], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0308, 0.0348, 0.0287, 0.0322, 0.0305, 0.0298, 0.0368], + device='cuda:2'), out_proj_covar=tensor([6.2979e-05, 6.3794e-05, 7.3632e-05, 5.8032e-05, 6.6726e-05, 6.4018e-05, + 6.2484e-05, 7.8238e-05], device='cuda:2') +2023-04-27 13:56:38,211 INFO [optim.py:369] (2/7) 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 13:56:38,300 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:47,739 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110672.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:48,304 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110673.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:56:48,910 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 1850, loss[loss=0.1988, simple_loss=0.2659, pruned_loss=0.06586, over 4819.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2483, pruned_loss=0.05231, over 959320.76 frames. ], batch size: 33, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:57:06,453 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110700.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:16,572 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0607, 2.3304, 1.9177, 2.2847, 1.6033, 1.9201, 2.1216, 1.5372], + device='cuda:2'), covar=tensor([0.1900, 0.1220, 0.0976, 0.1240, 0.3384, 0.1320, 0.1909, 0.2610], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0299, 0.0216, 0.0276, 0.0310, 0.0257, 0.0248, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1430e-04, 1.1858e-04, 8.5511e-05, 1.0960e-04, 1.2556e-04, 1.0180e-04, + 1.0029e-04, 1.0362e-04], device='cuda:2') +2023-04-27 13:57:17,242 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-27 13:57:21,095 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110721.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:24,738 INFO [finetune.py:976] (2/7) Epoch 20, batch 1900, loss[loss=0.1866, simple_loss=0.2565, pruned_loss=0.05829, over 4724.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2493, pruned_loss=0.05225, over 958969.51 frames. ], batch size: 54, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:57:30,854 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:38,119 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110748.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:41,157 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 13:57:41,604 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5562, 3.7483, 1.2606, 1.8599, 2.0381, 2.7819, 2.0445, 1.1251], + device='cuda:2'), covar=tensor([0.1533, 0.1019, 0.1852, 0.1383, 0.1114, 0.0942, 0.1614, 0.1995], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0245, 0.0139, 0.0121, 0.0135, 0.0154, 0.0118, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:57:46,353 INFO [optim.py:369] (2/7) 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 13:57:48,391 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4388, 2.4163, 2.5644, 2.9230, 2.9009, 2.4244, 1.9688, 2.5599], + device='cuda:2'), covar=tensor([0.0802, 0.0872, 0.0613, 0.0562, 0.0583, 0.0801, 0.0756, 0.0576], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0201, 0.0184, 0.0174, 0.0179, 0.0182, 0.0153, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 13:57:51,432 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:57:55,050 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 1950, loss[loss=0.154, simple_loss=0.231, pruned_loss=0.03856, over 4827.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2477, pruned_loss=0.05138, over 958741.41 frames. ], batch size: 30, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:58:29,591 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110797.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 13:58:52,669 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110814.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:59:01,944 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110819.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:59:11,701 INFO [finetune.py:976] (2/7) Epoch 20, batch 2000, loss[loss=0.1849, simple_loss=0.2533, pruned_loss=0.05827, over 3993.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2457, pruned_loss=0.05104, over 958448.82 frames. ], batch size: 17, lr: 3.25e-03, grad_scale: 32.0 +2023-04-27 13:59:13,084 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7668, 1.8685, 1.0016, 1.3626, 1.9495, 1.6136, 1.4538, 1.5394], + device='cuda:2'), covar=tensor([0.0485, 0.0353, 0.0321, 0.0560, 0.0260, 0.0505, 0.0472, 0.0562], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 13:59:33,914 INFO [optim.py:369] (2/7) 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 13:59:45,775 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110868.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 13:59:56,922 INFO [finetune.py:976] (2/7) Epoch 20, batch 2050, loss[loss=0.1672, simple_loss=0.2421, pruned_loss=0.0462, over 4905.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2424, pruned_loss=0.05012, over 956644.23 frames. ], batch size: 35, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 13:59:58,162 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2382, 2.5406, 0.7552, 1.4546, 1.5363, 1.8943, 1.6446, 0.8498], + device='cuda:2'), covar=tensor([0.1434, 0.1283, 0.1878, 0.1329, 0.1129, 0.0961, 0.1653, 0.1630], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0243, 0.0138, 0.0120, 0.0134, 0.0153, 0.0118, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 13:59:59,395 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8062, 1.9812, 1.6508, 1.9713, 1.4656, 1.6066, 1.8019, 1.3557], + device='cuda:2'), covar=tensor([0.1539, 0.1167, 0.0940, 0.1115, 0.3695, 0.1136, 0.1568, 0.2292], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0299, 0.0215, 0.0276, 0.0309, 0.0257, 0.0248, 0.0260], + device='cuda:2'), out_proj_covar=tensor([1.1381e-04, 1.1847e-04, 8.5300e-05, 1.0931e-04, 1.2527e-04, 1.0171e-04, + 1.0001e-04, 1.0312e-04], device='cuda:2') +2023-04-27 14:00:53,299 INFO [finetune.py:976] (2/7) Epoch 20, batch 2100, loss[loss=0.1577, simple_loss=0.2254, pruned_loss=0.04498, over 4783.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.243, pruned_loss=0.05088, over 958428.83 frames. ], batch size: 26, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:00:54,679 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:00:58,857 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 14:00:59,936 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9424, 2.3370, 1.9040, 2.2288, 1.6677, 1.9404, 1.9326, 1.5527], + device='cuda:2'), covar=tensor([0.1794, 0.1074, 0.0913, 0.1146, 0.3429, 0.1193, 0.1873, 0.2584], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0301, 0.0217, 0.0278, 0.0310, 0.0258, 0.0249, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1468e-04, 1.1945e-04, 8.5826e-05, 1.1001e-04, 1.2591e-04, 1.0232e-04, + 1.0066e-04, 1.0378e-04], device='cuda:2') +2023-04-27 14:01:01,679 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110939.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:13,888 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110958.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:14,357 INFO [optim.py:369] (2/7) 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 14:01:14,970 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110959.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:20,707 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110967.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:25,025 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:26,822 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 14:01:27,216 INFO [finetune.py:976] (2/7) Epoch 20, batch 2150, loss[loss=0.2107, simple_loss=0.2891, pruned_loss=0.06609, over 4829.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2453, pruned_loss=0.05131, over 956198.64 frames. ], batch size: 47, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:01:34,418 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111007.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:50,284 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 14:01:55,861 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111019.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:01:57,638 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111022.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:02:00,609 INFO [finetune.py:976] (2/7) Epoch 20, batch 2200, loss[loss=0.1338, simple_loss=0.2157, pruned_loss=0.02598, over 4834.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2467, pruned_loss=0.05144, over 955605.50 frames. ], batch size: 25, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:02:21,780 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 14:02:22,144 INFO [optim.py:369] (2/7) 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 14:02:33,622 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 14:02:34,615 INFO [finetune.py:976] (2/7) Epoch 20, batch 2250, loss[loss=0.146, simple_loss=0.2327, pruned_loss=0.02965, over 4830.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.248, pruned_loss=0.05151, over 956851.77 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:02:43,210 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3381, 2.0913, 2.5077, 3.1820, 2.1182, 1.6824, 2.0810, 1.2128], + device='cuda:2'), covar=tensor([0.0542, 0.0707, 0.0406, 0.0455, 0.0655, 0.1466, 0.0705, 0.0945], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0069, 0.0067, 0.0068, 0.0075, 0.0097, 0.0074, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:02:44,861 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111092.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:03:08,137 INFO [finetune.py:976] (2/7) Epoch 20, batch 2300, loss[loss=0.1543, simple_loss=0.2345, pruned_loss=0.03709, over 4756.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2473, pruned_loss=0.05088, over 954014.56 frames. ], batch size: 27, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:03:29,169 INFO [optim.py:369] (2/7) 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 14:03:32,376 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0829, 1.8175, 2.2451, 2.4379, 2.0911, 2.0016, 2.1187, 2.0414], + device='cuda:2'), covar=tensor([0.4924, 0.7200, 0.7767, 0.6304, 0.6097, 0.9144, 0.8642, 1.0385], + device='cuda:2'), in_proj_covar=tensor([0.0428, 0.0410, 0.0504, 0.0508, 0.0456, 0.0484, 0.0490, 0.0496], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:03:35,770 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7684, 1.7336, 2.2823, 2.4263, 1.6759, 1.4621, 1.7338, 0.9707], + device='cuda:2'), covar=tensor([0.0703, 0.0970, 0.0495, 0.0810, 0.0950, 0.1346, 0.1033, 0.0839], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0069, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:03:41,038 INFO [finetune.py:976] (2/7) Epoch 20, batch 2350, loss[loss=0.1571, simple_loss=0.2344, pruned_loss=0.03991, over 4777.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2443, pruned_loss=0.04987, over 954679.37 frames. ], batch size: 28, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:04:27,661 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111224.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:04:28,340 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:04:29,453 INFO [finetune.py:976] (2/7) Epoch 20, batch 2400, loss[loss=0.1735, simple_loss=0.2468, pruned_loss=0.0501, over 4833.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2425, pruned_loss=0.0496, over 954705.53 frames. ], batch size: 30, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:05:02,140 INFO [optim.py:369] (2/7) 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 14:05:07,622 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111267.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:05:19,547 INFO [finetune.py:976] (2/7) Epoch 20, batch 2450, loss[loss=0.1881, simple_loss=0.2684, pruned_loss=0.05393, over 4834.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2407, pruned_loss=0.04955, over 955657.82 frames. ], batch size: 39, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:05:36,589 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111286.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:06:05,493 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111314.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:06:06,092 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111315.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:06:13,837 INFO [finetune.py:976] (2/7) Epoch 20, batch 2500, loss[loss=0.2292, simple_loss=0.3013, pruned_loss=0.07861, over 4805.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2415, pruned_loss=0.0502, over 954346.97 frames. ], batch size: 45, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:06:35,391 INFO [optim.py:369] (2/7) 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 14:06:46,745 INFO [finetune.py:976] (2/7) Epoch 20, batch 2550, loss[loss=0.2483, simple_loss=0.3029, pruned_loss=0.0969, over 4101.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2459, pruned_loss=0.05158, over 953795.97 frames. ], batch size: 65, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:06:57,259 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111392.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:07:00,778 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2629, 1.4655, 1.4592, 1.6834, 1.6354, 1.7930, 1.4030, 3.5240], + device='cuda:2'), covar=tensor([0.0584, 0.0829, 0.0780, 0.1198, 0.0611, 0.0495, 0.0750, 0.0138], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 14:07:20,615 INFO [finetune.py:976] (2/7) Epoch 20, batch 2600, loss[loss=0.1704, simple_loss=0.2417, pruned_loss=0.04957, over 4186.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2481, pruned_loss=0.05246, over 953662.95 frames. ], batch size: 65, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:07:29,129 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111440.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:07:42,989 INFO [optim.py:369] (2/7) 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 14:07:48,661 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111468.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:07:54,482 INFO [finetune.py:976] (2/7) Epoch 20, batch 2650, loss[loss=0.1926, simple_loss=0.2758, pruned_loss=0.05477, over 4822.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2489, pruned_loss=0.05242, over 955021.78 frames. ], batch size: 39, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:08:26,361 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111524.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:08:28,082 INFO [finetune.py:976] (2/7) Epoch 20, batch 2700, loss[loss=0.1532, simple_loss=0.2318, pruned_loss=0.03728, over 4809.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2478, pruned_loss=0.05209, over 954845.02 frames. ], batch size: 51, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:08:29,431 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111529.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:08:49,484 INFO [optim.py:369] (2/7) 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 14:08:58,396 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111572.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:09:01,908 INFO [finetune.py:976] (2/7) Epoch 20, batch 2750, loss[loss=0.1465, simple_loss=0.2194, pruned_loss=0.03682, over 4749.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2443, pruned_loss=0.05103, over 953347.91 frames. ], batch size: 27, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:09:04,329 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111581.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:09:37,583 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111614.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:09:38,681 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3678, 3.3406, 2.4793, 3.9037, 3.4220, 3.3469, 1.5441, 3.2745], + device='cuda:2'), covar=tensor([0.1977, 0.1346, 0.2983, 0.2160, 0.2660, 0.1976, 0.5696, 0.2685], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0214, 0.0249, 0.0302, 0.0295, 0.0247, 0.0272, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:09:57,463 INFO [finetune.py:976] (2/7) Epoch 20, batch 2800, loss[loss=0.1548, simple_loss=0.2218, pruned_loss=0.04386, over 4910.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2418, pruned_loss=0.05067, over 951956.30 frames. ], batch size: 43, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:10:24,458 INFO [optim.py:369] (2/7) 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 14:10:26,838 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111662.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:10:36,931 INFO [finetune.py:976] (2/7) Epoch 20, batch 2850, loss[loss=0.1577, simple_loss=0.2388, pruned_loss=0.03837, over 4763.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2392, pruned_loss=0.04932, over 951053.73 frames. ], batch size: 28, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:10:41,383 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 14:11:20,203 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6063, 1.9126, 2.0001, 2.0767, 1.9083, 2.0127, 2.0392, 2.0171], + device='cuda:2'), covar=tensor([0.3833, 0.5054, 0.4372, 0.4297, 0.5332, 0.6367, 0.5212, 0.4874], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0370, 0.0320, 0.0333, 0.0343, 0.0390, 0.0356, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:11:31,378 INFO [finetune.py:976] (2/7) Epoch 20, batch 2900, loss[loss=0.1143, simple_loss=0.1791, pruned_loss=0.02476, over 4167.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2421, pruned_loss=0.0507, over 951492.61 frames. ], batch size: 17, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:12:12,495 INFO [optim.py:369] (2/7) 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 14:12:36,595 INFO [finetune.py:976] (2/7) Epoch 20, batch 2950, loss[loss=0.1461, simple_loss=0.2272, pruned_loss=0.03251, over 4788.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2447, pruned_loss=0.05126, over 953266.78 frames. ], batch size: 29, lr: 3.24e-03, grad_scale: 64.0 +2023-04-27 14:12:55,725 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4285, 1.1042, 0.3029, 1.1668, 1.1288, 1.3251, 1.2282, 1.2403], + device='cuda:2'), covar=tensor([0.0553, 0.0417, 0.0441, 0.0598, 0.0312, 0.0550, 0.0529, 0.0603], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0049, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 14:12:56,808 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111791.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 14:13:05,057 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4291, 1.8511, 2.2859, 2.7798, 2.2881, 1.8240, 1.6075, 2.0870], + device='cuda:2'), covar=tensor([0.3270, 0.3269, 0.1775, 0.2542, 0.2579, 0.2714, 0.4147, 0.2212], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0245, 0.0226, 0.0313, 0.0217, 0.0232, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 14:13:27,629 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7270, 2.2997, 1.5514, 1.5454, 1.2488, 1.2824, 1.6457, 1.2205], + device='cuda:2'), covar=tensor([0.1865, 0.1253, 0.1769, 0.1895, 0.2592, 0.2301, 0.1141, 0.2195], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0211, 0.0169, 0.0204, 0.0200, 0.0186, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 14:13:29,430 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3118, 1.7024, 2.1565, 2.6824, 2.1430, 1.7127, 1.5253, 2.0391], + device='cuda:2'), covar=tensor([0.3019, 0.3208, 0.1718, 0.2448, 0.2706, 0.2816, 0.4297, 0.2129], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0226, 0.0314, 0.0218, 0.0232, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 14:13:29,950 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 3000, loss[loss=0.1811, simple_loss=0.2556, pruned_loss=0.05327, over 4780.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2473, pruned_loss=0.05259, over 951280.99 frames. ], batch size: 51, lr: 3.24e-03, grad_scale: 64.0 +2023-04-27 14:13:32,106 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 14:13:52,372 INFO [finetune.py:1010] (2/7) Epoch 20, validation: loss=0.1527, simple_loss=0.2229, pruned_loss=0.04123, over 2265189.00 frames. +2023-04-27 14:13:52,373 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 14:14:18,540 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-27 14:14:20,671 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111852.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:14:29,833 INFO [optim.py:369] (2/7) 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 14:14:52,301 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.37 vs. limit=5.0 +2023-04-27 14:14:58,616 INFO [finetune.py:976] (2/7) Epoch 20, batch 3050, loss[loss=0.1933, simple_loss=0.2613, pruned_loss=0.06263, over 4909.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2485, pruned_loss=0.05223, over 949829.64 frames. ], batch size: 37, lr: 3.24e-03, grad_scale: 64.0 +2023-04-27 14:15:01,109 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111881.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:15:48,558 INFO [finetune.py:976] (2/7) Epoch 20, batch 3100, loss[loss=0.1842, simple_loss=0.2434, pruned_loss=0.06247, over 4818.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2473, pruned_loss=0.05193, over 952450.42 frames. ], batch size: 30, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:15:50,328 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:15:59,434 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4217, 1.8725, 2.2734, 2.6120, 2.2987, 1.8254, 1.4461, 2.0218], + device='cuda:2'), covar=tensor([0.3305, 0.3123, 0.1702, 0.2330, 0.2559, 0.2768, 0.4181, 0.2028], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0226, 0.0313, 0.0218, 0.0232, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 14:16:16,415 INFO [optim.py:369] (2/7) 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 14:16:21,338 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8556, 1.9398, 0.9571, 1.5856, 2.1918, 1.7300, 1.6771, 1.6888], + device='cuda:2'), covar=tensor([0.0455, 0.0330, 0.0296, 0.0504, 0.0224, 0.0476, 0.0444, 0.0510], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0049, 0.0050], + device='cuda:2') +2023-04-27 14:16:27,777 INFO [finetune.py:976] (2/7) Epoch 20, batch 3150, loss[loss=0.1489, simple_loss=0.2103, pruned_loss=0.04376, over 3994.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2454, pruned_loss=0.05162, over 951861.87 frames. ], batch size: 17, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:17:01,632 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-04-27 14:17:02,007 INFO [finetune.py:976] (2/7) Epoch 20, batch 3200, loss[loss=0.1725, simple_loss=0.2407, pruned_loss=0.05214, over 4861.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2414, pruned_loss=0.05065, over 952525.02 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 32.0 +2023-04-27 14:17:35,520 INFO [optim.py:369] (2/7) 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 14:17:38,528 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4405, 1.8417, 2.3130, 2.8539, 2.3505, 1.8838, 1.7719, 2.3049], + device='cuda:2'), covar=tensor([0.3046, 0.3115, 0.1598, 0.2245, 0.2438, 0.2457, 0.3668, 0.1850], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0246, 0.0226, 0.0313, 0.0218, 0.0232, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 14:17:39,744 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112066.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:17:45,864 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.91 vs. limit=5.0 +2023-04-27 14:17:46,878 INFO [finetune.py:976] (2/7) Epoch 20, batch 3250, loss[loss=0.197, simple_loss=0.2682, pruned_loss=0.06288, over 4865.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.241, pruned_loss=0.05026, over 950692.38 frames. ], batch size: 31, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:17:57,056 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3720, 1.3715, 1.6842, 1.6544, 1.2803, 1.1418, 1.3999, 0.8777], + device='cuda:2'), covar=tensor([0.0564, 0.0586, 0.0375, 0.0662, 0.0789, 0.1087, 0.0510, 0.0564], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0069, 0.0068, 0.0068, 0.0076, 0.0097, 0.0074, 0.0066], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:18:02,876 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0562, 2.3956, 1.0544, 1.3296, 1.8930, 1.2934, 3.2583, 1.7184], + device='cuda:2'), covar=tensor([0.0666, 0.0575, 0.0708, 0.1296, 0.0480, 0.0998, 0.0245, 0.0654], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0046, 0.0049, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 14:18:09,605 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7776, 1.4456, 1.3785, 1.6703, 2.0170, 1.6378, 1.4007, 1.3322], + device='cuda:2'), covar=tensor([0.1523, 0.1330, 0.2026, 0.1136, 0.0844, 0.1430, 0.1856, 0.2323], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0313, 0.0353, 0.0290, 0.0328, 0.0309, 0.0303, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4225e-05, 6.4711e-05, 7.4793e-05, 5.8602e-05, 6.7735e-05, 6.4856e-05, + 6.3492e-05, 7.9287e-05], device='cuda:2') +2023-04-27 14:18:18,410 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:18:20,135 INFO [finetune.py:976] (2/7) Epoch 20, batch 3300, loss[loss=0.2078, simple_loss=0.285, pruned_loss=0.06535, over 4803.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2446, pruned_loss=0.05135, over 951884.02 frames. ], batch size: 51, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:18:20,278 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112127.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:18:44,844 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112147.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:18:52,672 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) Epoch 20, batch 3350, loss[loss=0.1423, simple_loss=0.2302, pruned_loss=0.02722, over 4882.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2474, pruned_loss=0.05194, over 952303.21 frames. ], batch size: 35, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:19:37,509 INFO [finetune.py:976] (2/7) Epoch 20, batch 3400, loss[loss=0.1987, simple_loss=0.2408, pruned_loss=0.0783, over 4042.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2483, pruned_loss=0.05257, over 950372.96 frames. ], batch size: 17, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:20:15,437 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 20, batch 3450, loss[loss=0.1693, simple_loss=0.2456, pruned_loss=0.04646, over 4885.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2481, pruned_loss=0.05207, over 950117.76 frames. ], batch size: 35, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:20:50,671 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8033, 2.0694, 0.9342, 1.5782, 2.3349, 1.6581, 1.6158, 1.6995], + device='cuda:2'), covar=tensor([0.0462, 0.0311, 0.0289, 0.0519, 0.0207, 0.0455, 0.0452, 0.0489], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0051, 0.0038, 0.0049, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 14:21:31,591 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8675, 2.8022, 2.2714, 3.2909, 2.8437, 2.8225, 1.2587, 2.8030], + device='cuda:2'), covar=tensor([0.2586, 0.1978, 0.3249, 0.2944, 0.3320, 0.2454, 0.6172, 0.2826], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0213, 0.0249, 0.0301, 0.0294, 0.0245, 0.0272, 0.0268], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:21:35,777 INFO [finetune.py:976] (2/7) Epoch 20, batch 3500, loss[loss=0.1605, simple_loss=0.2358, pruned_loss=0.04265, over 4901.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2454, pruned_loss=0.05124, over 952570.36 frames. ], batch size: 36, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:21:41,793 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6970, 1.5662, 1.7041, 2.0702, 2.1421, 1.6514, 1.3138, 1.7723], + device='cuda:2'), covar=tensor([0.0757, 0.1159, 0.0775, 0.0560, 0.0494, 0.0788, 0.0827, 0.0589], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0200, 0.0182, 0.0171, 0.0177, 0.0180, 0.0152, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:21:43,682 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8594, 1.3793, 1.9287, 2.3592, 1.9518, 1.8267, 1.8954, 1.7904], + device='cuda:2'), covar=tensor([0.4602, 0.6714, 0.6461, 0.5869, 0.5901, 0.7866, 0.8182, 0.8938], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0411, 0.0506, 0.0506, 0.0456, 0.0485, 0.0492, 0.0497], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:21:59,355 INFO [optim.py:369] (2/7) 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 14:22:09,766 INFO [finetune.py:976] (2/7) Epoch 20, batch 3550, loss[loss=0.1202, simple_loss=0.1975, pruned_loss=0.02146, over 4759.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2419, pruned_loss=0.04978, over 953402.19 frames. ], batch size: 28, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:22:18,187 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112389.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:22:40,996 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112422.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:22:44,032 INFO [finetune.py:976] (2/7) Epoch 20, batch 3600, loss[loss=0.1324, simple_loss=0.1932, pruned_loss=0.03577, over 4720.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2397, pruned_loss=0.04951, over 951150.93 frames. ], batch size: 23, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:22:55,010 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7033, 1.6118, 1.9434, 2.1308, 1.5810, 1.3950, 1.7148, 1.0356], + device='cuda:2'), covar=tensor([0.0657, 0.0616, 0.0518, 0.0702, 0.0813, 0.1040, 0.0637, 0.0715], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0069, 0.0068, 0.0069, 0.0076, 0.0098, 0.0074, 0.0066], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:22:57,509 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112447.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 14:22:59,366 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112450.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:22:59,383 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6125, 2.1805, 2.4832, 3.1178, 2.4994, 2.1665, 2.0777, 2.3128], + device='cuda:2'), covar=tensor([0.2978, 0.2873, 0.1433, 0.2015, 0.2397, 0.2345, 0.3371, 0.1934], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0246, 0.0226, 0.0313, 0.0219, 0.0233, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 14:23:06,296 INFO [optim.py:369] (2/7) 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 14:23:18,445 INFO [finetune.py:976] (2/7) Epoch 20, batch 3650, loss[loss=0.1761, simple_loss=0.2442, pruned_loss=0.05395, over 4790.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.241, pruned_loss=0.04998, over 951850.45 frames. ], batch size: 26, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:23:30,563 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112495.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:23:58,095 INFO [finetune.py:976] (2/7) Epoch 20, batch 3700, loss[loss=0.1964, simple_loss=0.2483, pruned_loss=0.07226, over 3845.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2453, pruned_loss=0.0515, over 949784.07 frames. ], batch size: 16, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:23:58,370 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-04-27 14:24:09,191 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7829, 2.3849, 0.9711, 1.2152, 1.5954, 1.0875, 2.8767, 1.3993], + device='cuda:2'), covar=tensor([0.0910, 0.0844, 0.0877, 0.1604, 0.0633, 0.1277, 0.0342, 0.0852], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0048, 0.0047, 0.0050, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 14:24:29,867 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1898, 2.5648, 0.9464, 1.4514, 1.5311, 1.8873, 1.6357, 0.8506], + device='cuda:2'), covar=tensor([0.1559, 0.1276, 0.1670, 0.1425, 0.1190, 0.0975, 0.1632, 0.1779], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0243, 0.0138, 0.0121, 0.0134, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:24:37,756 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8726, 1.3822, 1.9942, 2.4191, 2.0016, 1.8867, 1.9699, 1.9307], + device='cuda:2'), covar=tensor([0.4570, 0.6665, 0.6040, 0.5441, 0.5932, 0.7712, 0.7842, 0.8021], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0411, 0.0506, 0.0507, 0.0456, 0.0486, 0.0492, 0.0497], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:24:40,763 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 14:24:41,784 INFO [optim.py:369] (2/7) 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 14:24:43,161 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112562.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:25:05,134 INFO [finetune.py:976] (2/7) Epoch 20, batch 3750, loss[loss=0.2183, simple_loss=0.2777, pruned_loss=0.07942, over 4821.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2469, pruned_loss=0.05208, over 950257.20 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:25:05,264 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:26:00,865 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:26:10,705 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112623.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:26:19,268 INFO [finetune.py:976] (2/7) Epoch 20, batch 3800, loss[loss=0.1885, simple_loss=0.2676, pruned_loss=0.05467, over 4718.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2477, pruned_loss=0.05234, over 950798.11 frames. ], batch size: 54, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:26:31,081 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112638.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 14:27:02,950 INFO [optim.py:369] (2/7) 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 14:27:07,984 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-27 14:27:25,229 INFO [finetune.py:976] (2/7) Epoch 20, batch 3850, loss[loss=0.1772, simple_loss=0.2498, pruned_loss=0.0523, over 4772.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2459, pruned_loss=0.05118, over 951185.50 frames. ], batch size: 28, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:27:25,950 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112678.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 14:27:27,273 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 14:27:38,714 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7349, 2.8263, 2.4166, 2.6813, 2.9623, 2.3668, 3.8070, 2.2571], + device='cuda:2'), covar=tensor([0.3274, 0.1858, 0.3324, 0.2824, 0.1386, 0.2362, 0.1152, 0.3416], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0349, 0.0425, 0.0354, 0.0380, 0.0374, 0.0372, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:28:09,558 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112722.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:28:09,595 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5242, 1.7902, 1.7020, 2.4659, 2.5928, 2.0927, 2.1021, 1.7959], + device='cuda:2'), covar=tensor([0.2022, 0.1947, 0.2307, 0.1832, 0.1354, 0.2035, 0.2210, 0.2386], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0313, 0.0351, 0.0290, 0.0328, 0.0308, 0.0302, 0.0374], + device='cuda:2'), 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:2') +2023-04-27 14:28:13,024 INFO [finetune.py:976] (2/7) Epoch 20, batch 3900, loss[loss=0.173, simple_loss=0.2409, pruned_loss=0.05255, over 4923.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2432, pruned_loss=0.05081, over 951607.04 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:28:24,999 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112745.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:28:27,504 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5749, 1.8235, 1.7885, 2.5645, 2.6852, 2.1285, 2.0842, 1.8517], + device='cuda:2'), covar=tensor([0.1931, 0.1872, 0.2097, 0.1594, 0.1258, 0.1913, 0.2565, 0.2475], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0313, 0.0351, 0.0289, 0.0327, 0.0308, 0.0302, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.4201e-05, 6.4808e-05, 7.4137e-05, 5.8491e-05, 6.7675e-05, 6.4605e-05, + 6.3245e-05, 7.9438e-05], device='cuda:2') +2023-04-27 14:28:34,456 INFO [optim.py:369] (2/7) 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 14:28:41,692 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 3950, loss[loss=0.1647, simple_loss=0.2322, pruned_loss=0.04866, over 4806.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2403, pruned_loss=0.04974, over 952503.43 frames. ], batch size: 25, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:29:10,113 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7477, 1.3685, 1.3433, 1.5980, 1.9645, 1.5765, 1.3818, 1.2494], + device='cuda:2'), covar=tensor([0.1440, 0.1420, 0.1713, 0.1445, 0.0763, 0.1614, 0.1875, 0.2009], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0310, 0.0348, 0.0287, 0.0324, 0.0306, 0.0299, 0.0370], + device='cuda:2'), out_proj_covar=tensor([6.3561e-05, 6.4313e-05, 7.3525e-05, 5.7976e-05, 6.7009e-05, 6.4064e-05, + 6.2721e-05, 7.8694e-05], device='cuda:2') +2023-04-27 14:29:20,148 INFO [finetune.py:976] (2/7) Epoch 20, batch 4000, loss[loss=0.1842, simple_loss=0.258, pruned_loss=0.05524, over 4909.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2395, pruned_loss=0.04909, over 951998.13 frames. ], batch size: 32, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:29:42,040 INFO [optim.py:369] (2/7) 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 14:29:53,759 INFO [finetune.py:976] (2/7) Epoch 20, batch 4050, loss[loss=0.1922, simple_loss=0.2658, pruned_loss=0.05931, over 4807.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2426, pruned_loss=0.0498, over 953229.23 frames. ], batch size: 38, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:30:21,648 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112918.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:30:23,594 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 4100, loss[loss=0.1582, simple_loss=0.2289, pruned_loss=0.04374, over 4810.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.245, pruned_loss=0.05036, over 952628.59 frames. ], batch size: 25, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:30:31,350 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112933.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 14:31:11,352 INFO [optim.py:369] (2/7) 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 14:31:11,458 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4353, 3.4193, 0.7804, 1.8039, 1.8519, 2.3769, 1.8529, 1.0125], + device='cuda:2'), covar=tensor([0.1393, 0.0819, 0.2177, 0.1204, 0.1151, 0.1025, 0.1517, 0.1849], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0119, 0.0133, 0.0152, 0.0116, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:31:25,620 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 4150, loss[loss=0.2291, simple_loss=0.2829, pruned_loss=0.08769, over 4140.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2469, pruned_loss=0.05155, over 951180.31 frames. ], batch size: 66, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:31:42,875 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112982.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:31:55,942 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6101, 1.3199, 1.2796, 1.4878, 1.8423, 1.5013, 1.2559, 1.2240], + device='cuda:2'), covar=tensor([0.1547, 0.1435, 0.1645, 0.1380, 0.0840, 0.1337, 0.1835, 0.2156], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0312, 0.0350, 0.0289, 0.0326, 0.0306, 0.0300, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.4001e-05, 6.4570e-05, 7.3941e-05, 5.8499e-05, 6.7430e-05, 6.4243e-05, + 6.2915e-05, 7.9096e-05], device='cuda:2') +2023-04-27 14:32:29,695 INFO [finetune.py:976] (2/7) Epoch 20, batch 4200, loss[loss=0.1743, simple_loss=0.2529, pruned_loss=0.0478, over 4894.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2474, pruned_loss=0.05131, over 952336.79 frames. ], batch size: 46, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:32:41,754 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113045.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:32:42,894 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6458, 0.6978, 1.4926, 2.0183, 1.7114, 1.5299, 1.5305, 1.5371], + device='cuda:2'), covar=tensor([0.4195, 0.6339, 0.5876, 0.5684, 0.5675, 0.7010, 0.7082, 0.7678], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0410, 0.0505, 0.0505, 0.0457, 0.0485, 0.0494, 0.0498], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:32:57,027 INFO [optim.py:369] (2/7) 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 14:33:18,849 INFO [finetune.py:976] (2/7) Epoch 20, batch 4250, loss[loss=0.1801, simple_loss=0.2352, pruned_loss=0.06252, over 4828.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2455, pruned_loss=0.05114, over 953161.92 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:33:31,121 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 14:33:40,239 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113093.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:33:43,327 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1094, 1.7267, 1.9312, 2.3774, 2.4040, 2.0374, 1.6790, 2.1808], + device='cuda:2'), covar=tensor([0.0689, 0.1074, 0.0694, 0.0488, 0.0517, 0.0717, 0.0742, 0.0482], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0203, 0.0185, 0.0172, 0.0178, 0.0182, 0.0154, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:33:47,740 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9351, 1.9677, 1.7621, 1.5746, 2.0360, 1.6123, 2.4534, 1.5561], + device='cuda:2'), covar=tensor([0.3589, 0.1677, 0.4553, 0.2968, 0.1695, 0.2285, 0.1422, 0.4117], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0346, 0.0422, 0.0352, 0.0378, 0.0372, 0.0369, 0.0414], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:34:04,771 INFO [finetune.py:976] (2/7) Epoch 20, batch 4300, loss[loss=0.1405, simple_loss=0.218, pruned_loss=0.03146, over 4824.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2434, pruned_loss=0.05066, over 954972.78 frames. ], batch size: 38, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:34:09,887 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 14:34:24,444 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 14:34:26,917 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-04-27 14:34:27,839 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 20, batch 4350, loss[loss=0.1467, simple_loss=0.2103, pruned_loss=0.04156, over 4778.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2399, pruned_loss=0.04913, over 956001.58 frames. ], batch size: 28, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:35:05,057 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4484, 1.2794, 4.0253, 3.5163, 3.6044, 3.8311, 3.7235, 3.3215], + device='cuda:2'), covar=tensor([0.9828, 0.8740, 0.1757, 0.3867, 0.2469, 0.4839, 0.2678, 0.3659], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0307, 0.0407, 0.0405, 0.0350, 0.0410, 0.0315, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:35:06,250 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113218.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:11,655 INFO [finetune.py:976] (2/7) Epoch 20, batch 4400, loss[loss=0.1862, simple_loss=0.2714, pruned_loss=0.05056, over 4842.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.241, pruned_loss=0.04988, over 953938.61 frames. ], batch size: 49, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:35:15,441 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113233.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:16,022 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113234.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:29,209 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.55 vs. limit=5.0 +2023-04-27 14:35:34,722 INFO [optim.py:369] (2/7) 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 14:35:38,475 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113266.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:42,676 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113273.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 14:35:43,499 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-04-27 14:35:45,062 INFO [finetune.py:976] (2/7) Epoch 20, batch 4450, loss[loss=0.2092, simple_loss=0.2852, pruned_loss=0.0666, over 4820.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2456, pruned_loss=0.05163, over 955805.45 frames. ], batch size: 39, lr: 3.23e-03, grad_scale: 32.0 +2023-04-27 14:35:45,129 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113277.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:47,574 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113281.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:35:56,988 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113295.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:36:15,056 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 4500, loss[loss=0.1611, simple_loss=0.2369, pruned_loss=0.0427, over 4793.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.248, pruned_loss=0.05238, over 955310.01 frames. ], batch size: 25, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:37:03,577 INFO [optim.py:369] (2/7) 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 14:37:25,452 INFO [finetune.py:976] (2/7) Epoch 20, batch 4550, loss[loss=0.1689, simple_loss=0.2478, pruned_loss=0.04502, over 4900.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2499, pruned_loss=0.05311, over 956734.78 frames. ], batch size: 37, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:37:38,026 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5175, 1.9377, 2.2911, 2.8990, 2.3006, 1.8446, 1.9285, 2.2181], + device='cuda:2'), covar=tensor([0.2906, 0.3154, 0.1507, 0.2276, 0.2612, 0.2710, 0.3572, 0.2148], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0226, 0.0313, 0.0218, 0.0232, 0.0226, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 14:37:46,239 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5895, 1.4041, 1.2550, 1.5465, 1.8034, 1.5877, 1.3866, 1.2334], + device='cuda:2'), covar=tensor([0.1297, 0.1158, 0.1340, 0.1045, 0.0678, 0.1110, 0.1391, 0.1633], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0315, 0.0353, 0.0291, 0.0328, 0.0309, 0.0303, 0.0376], + device='cuda:2'), out_proj_covar=tensor([6.4496e-05, 6.5233e-05, 7.4585e-05, 5.8767e-05, 6.7902e-05, 6.4748e-05, + 6.3491e-05, 7.9880e-05], device='cuda:2') +2023-04-27 14:37:46,286 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-04-27 14:38:20,443 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3930, 1.6127, 1.8393, 1.9421, 1.7451, 1.8453, 1.8811, 1.8629], + device='cuda:2'), covar=tensor([0.3719, 0.5385, 0.4026, 0.3917, 0.5279, 0.6479, 0.4631, 0.4364], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0372, 0.0319, 0.0334, 0.0344, 0.0392, 0.0354, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:38:24,622 INFO [finetune.py:976] (2/7) Epoch 20, batch 4600, loss[loss=0.1754, simple_loss=0.2478, pruned_loss=0.05151, over 4872.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2492, pruned_loss=0.05281, over 955660.10 frames. ], batch size: 34, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:39:00,865 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 20, batch 4650, loss[loss=0.1743, simple_loss=0.2402, pruned_loss=0.05415, over 4887.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2468, pruned_loss=0.05223, over 957103.53 frames. ], batch size: 32, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:39:13,918 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2258, 1.5999, 2.0808, 2.3582, 2.0601, 1.5744, 1.2761, 1.8130], + device='cuda:2'), covar=tensor([0.2717, 0.2878, 0.1501, 0.2060, 0.2194, 0.2383, 0.3911, 0.1778], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0226, 0.0312, 0.0218, 0.0232, 0.0226, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 14:39:14,586 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 14:39:18,985 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-04-27 14:39:47,027 INFO [finetune.py:976] (2/7) Epoch 20, batch 4700, loss[loss=0.1164, simple_loss=0.1989, pruned_loss=0.01699, over 4832.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2436, pruned_loss=0.05137, over 956881.22 frames. ], batch size: 30, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:40:08,112 INFO [optim.py:369] (2/7) 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 14:40:19,997 INFO [finetune.py:976] (2/7) Epoch 20, batch 4750, loss[loss=0.1474, simple_loss=0.2307, pruned_loss=0.03207, over 4816.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2411, pruned_loss=0.04993, over 957325.54 frames. ], batch size: 39, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:40:20,084 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:29,014 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113590.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:35,206 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6482, 1.3942, 1.2304, 1.5994, 1.8575, 1.5532, 1.4140, 1.1417], + device='cuda:2'), covar=tensor([0.1776, 0.1696, 0.2013, 0.1352, 0.0980, 0.1742, 0.2098, 0.2341], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0312, 0.0350, 0.0289, 0.0325, 0.0307, 0.0301, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.3879e-05, 6.4733e-05, 7.4047e-05, 5.8369e-05, 6.7188e-05, 6.4281e-05, + 6.2968e-05, 7.9177e-05], device='cuda:2') +2023-04-27 14:40:46,845 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113617.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:52,169 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:40:53,787 INFO [finetune.py:976] (2/7) Epoch 20, batch 4800, loss[loss=0.182, simple_loss=0.2658, pruned_loss=0.04911, over 4901.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2439, pruned_loss=0.05109, over 955513.60 frames. ], batch size: 43, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:41:06,498 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113646.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:41:13,302 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 14:41:15,402 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 20, batch 4850, loss[loss=0.1458, simple_loss=0.2235, pruned_loss=0.03401, over 4815.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2466, pruned_loss=0.05151, over 953380.91 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:41:27,805 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113678.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:41:42,788 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2089, 1.5454, 1.3492, 1.7796, 1.6396, 1.7765, 1.3830, 3.3910], + device='cuda:2'), covar=tensor([0.0604, 0.0778, 0.0824, 0.1126, 0.0600, 0.0518, 0.0742, 0.0146], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 14:41:46,442 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:41:50,523 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5464, 1.5414, 4.4067, 4.1631, 3.8267, 4.2283, 4.0096, 3.8375], + device='cuda:2'), covar=tensor([0.7288, 0.5396, 0.0989, 0.1678, 0.1071, 0.1547, 0.1514, 0.1671], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0307, 0.0407, 0.0406, 0.0350, 0.0411, 0.0315, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:41:58,721 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2293, 1.5387, 4.3733, 4.1707, 3.9219, 4.0416, 3.8556, 3.8903], + device='cuda:2'), covar=tensor([0.5873, 0.5328, 0.1008, 0.1522, 0.0976, 0.1528, 0.3647, 0.1414], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0308, 0.0408, 0.0407, 0.0350, 0.0411, 0.0315, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:42:00,310 INFO [finetune.py:976] (2/7) Epoch 20, batch 4900, loss[loss=0.2025, simple_loss=0.2733, pruned_loss=0.06586, over 4177.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2482, pruned_loss=0.05242, over 951950.10 frames. ], batch size: 65, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:42:02,864 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2645, 2.9049, 2.2402, 2.3847, 1.6198, 1.5635, 2.4269, 1.6284], + device='cuda:2'), covar=tensor([0.1595, 0.1369, 0.1282, 0.1525, 0.2254, 0.1820, 0.0909, 0.1910], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0211, 0.0168, 0.0203, 0.0198, 0.0185, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 14:42:26,158 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 14:42:26,595 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8730, 3.7362, 2.6589, 4.4532, 3.8674, 3.8598, 1.6593, 3.8145], + device='cuda:2'), covar=tensor([0.1525, 0.1317, 0.3697, 0.1394, 0.2710, 0.1663, 0.5636, 0.2276], + device='cuda:2'), in_proj_covar=tensor([0.0248, 0.0216, 0.0253, 0.0308, 0.0298, 0.0249, 0.0276, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:42:27,118 INFO [optim.py:369] (2/7) 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 14:42:38,977 INFO [finetune.py:976] (2/7) Epoch 20, batch 4950, loss[loss=0.1848, simple_loss=0.2625, pruned_loss=0.0536, over 4892.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2489, pruned_loss=0.05231, over 953549.74 frames. ], batch size: 37, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:43:28,111 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3130, 3.2479, 2.4374, 3.8340, 3.3176, 3.2590, 1.2962, 3.2510], + device='cuda:2'), covar=tensor([0.2266, 0.1406, 0.3508, 0.2434, 0.2889, 0.2027, 0.5816, 0.2784], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0215, 0.0253, 0.0306, 0.0297, 0.0248, 0.0275, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:43:30,853 INFO [finetune.py:976] (2/7) Epoch 20, batch 5000, loss[loss=0.1905, simple_loss=0.2578, pruned_loss=0.06156, over 4210.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2468, pruned_loss=0.05125, over 952921.92 frames. ], batch size: 65, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:44:14,003 INFO [optim.py:369] (2/7) 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 14:44:16,043 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9596, 2.2424, 2.1578, 2.3834, 2.1481, 2.2764, 2.2276, 2.2339], + device='cuda:2'), covar=tensor([0.3893, 0.6214, 0.4894, 0.4451, 0.5805, 0.7065, 0.5982, 0.5791], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0374, 0.0321, 0.0336, 0.0345, 0.0395, 0.0357, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:44:32,033 INFO [finetune.py:976] (2/7) Epoch 20, batch 5050, loss[loss=0.1483, simple_loss=0.2155, pruned_loss=0.04053, over 4768.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2438, pruned_loss=0.05038, over 952984.90 frames. ], batch size: 26, lr: 3.22e-03, grad_scale: 32.0 +2023-04-27 14:44:41,935 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9129, 2.2878, 0.8967, 1.2573, 1.4728, 1.1505, 2.4473, 1.2988], + device='cuda:2'), covar=tensor([0.0647, 0.0539, 0.0620, 0.1185, 0.0462, 0.0971, 0.0288, 0.0687], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 14:44:52,888 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113890.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:45:34,209 INFO [finetune.py:976] (2/7) Epoch 20, batch 5100, loss[loss=0.1748, simple_loss=0.2396, pruned_loss=0.05497, over 4813.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2413, pruned_loss=0.04998, over 953424.83 frames. ], batch size: 25, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:45:42,068 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113938.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:45:45,611 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6846, 1.6035, 1.9610, 2.1889, 1.5675, 1.3464, 1.7679, 1.0001], + device='cuda:2'), covar=tensor([0.0710, 0.0575, 0.0578, 0.0625, 0.0806, 0.1140, 0.0615, 0.0773], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0069, 0.0068, 0.0068, 0.0076, 0.0097, 0.0074, 0.0066], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:45:54,368 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([5.0509, 5.0657, 3.4706, 5.7469, 5.1017, 4.9883, 2.5237, 5.0449], + device='cuda:2'), covar=tensor([0.1754, 0.1052, 0.2949, 0.0896, 0.3405, 0.1651, 0.5507, 0.1990], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0216, 0.0253, 0.0307, 0.0297, 0.0248, 0.0276, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:45:57,339 INFO [optim.py:369] (2/7) 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 14:46:05,378 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 20, batch 5150, loss[loss=0.1551, simple_loss=0.219, pruned_loss=0.04562, over 4706.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2408, pruned_loss=0.05022, over 951772.55 frames. ], batch size: 23, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:46:27,889 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114002.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:46:31,599 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3264, 3.2878, 2.5440, 3.8575, 3.3610, 3.2699, 1.3943, 3.2672], + device='cuda:2'), covar=tensor([0.2022, 0.1346, 0.3325, 0.2403, 0.3206, 0.2121, 0.6394, 0.2750], + device='cuda:2'), in_proj_covar=tensor([0.0249, 0.0217, 0.0254, 0.0309, 0.0298, 0.0250, 0.0277, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:46:43,539 INFO [finetune.py:976] (2/7) Epoch 20, batch 5200, loss[loss=0.1847, simple_loss=0.2636, pruned_loss=0.05292, over 4149.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2456, pruned_loss=0.0521, over 952605.14 frames. ], batch size: 65, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:47:06,606 INFO [optim.py:369] (2/7) 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 14:47:16,853 INFO [finetune.py:976] (2/7) Epoch 20, batch 5250, loss[loss=0.1784, simple_loss=0.2478, pruned_loss=0.05455, over 4819.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2473, pruned_loss=0.05193, over 953579.37 frames. ], batch size: 30, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:47:50,695 INFO [finetune.py:976] (2/7) Epoch 20, batch 5300, loss[loss=0.1492, simple_loss=0.2185, pruned_loss=0.0399, over 4775.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2489, pruned_loss=0.05217, over 954773.15 frames. ], batch size: 26, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:47:50,797 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114127.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:47:59,390 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 14:48:01,094 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7859, 1.2501, 1.8210, 2.2276, 1.8726, 1.7265, 1.7965, 1.7672], + device='cuda:2'), covar=tensor([0.4388, 0.7190, 0.6378, 0.5656, 0.6050, 0.8073, 0.7788, 0.8859], + device='cuda:2'), in_proj_covar=tensor([0.0430, 0.0412, 0.0506, 0.0507, 0.0458, 0.0487, 0.0495, 0.0499], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:48:13,280 INFO [optim.py:369] (2/7) 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 14:48:24,237 INFO [finetune.py:976] (2/7) Epoch 20, batch 5350, loss[loss=0.1959, simple_loss=0.2553, pruned_loss=0.06828, over 4857.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2476, pruned_loss=0.05118, over 954751.55 frames. ], batch size: 31, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:48:31,432 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114188.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:48:33,239 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9785, 2.5312, 2.2347, 2.2878, 1.6799, 2.0784, 2.1932, 1.6406], + device='cuda:2'), covar=tensor([0.1920, 0.0971, 0.0694, 0.1150, 0.3099, 0.1151, 0.1694, 0.2585], + device='cuda:2'), in_proj_covar=tensor([0.0288, 0.0303, 0.0218, 0.0278, 0.0314, 0.0258, 0.0250, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1546e-04, 1.2019e-04, 8.6144e-05, 1.1009e-04, 1.2739e-04, 1.0229e-04, + 1.0126e-04, 1.0430e-04], device='cuda:2') +2023-04-27 14:48:58,063 INFO [finetune.py:976] (2/7) Epoch 20, batch 5400, loss[loss=0.1589, simple_loss=0.2242, pruned_loss=0.04683, over 4738.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2445, pruned_loss=0.04995, over 953925.14 frames. ], batch size: 23, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:49:33,868 INFO [optim.py:369] (2/7) 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 14:49:33,961 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9354, 1.1735, 4.8045, 4.4933, 4.1793, 4.5295, 4.3185, 4.1476], + device='cuda:2'), covar=tensor([0.6961, 0.6365, 0.1127, 0.1796, 0.0943, 0.1303, 0.1394, 0.1476], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0305, 0.0404, 0.0402, 0.0347, 0.0408, 0.0312, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:49:44,713 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-27 14:49:53,954 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:49:56,285 INFO [finetune.py:976] (2/7) Epoch 20, batch 5450, loss[loss=0.1479, simple_loss=0.2157, pruned_loss=0.04006, over 4913.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2419, pruned_loss=0.04955, over 954891.36 frames. ], batch size: 32, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:50:28,226 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114302.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:50:58,434 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:50:59,247 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 14:51:02,103 INFO [finetune.py:976] (2/7) Epoch 20, batch 5500, loss[loss=0.1931, simple_loss=0.2449, pruned_loss=0.07071, over 4836.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2383, pruned_loss=0.04824, over 952334.30 frames. ], batch size: 30, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:51:22,116 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 14:51:34,724 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4181, 1.2992, 1.6264, 1.5950, 1.3246, 1.2321, 1.3421, 0.9564], + device='cuda:2'), covar=tensor([0.0512, 0.0624, 0.0399, 0.0571, 0.0788, 0.1137, 0.0549, 0.0557], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0069, 0.0067, 0.0068, 0.0076, 0.0097, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:51:40,049 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.21 vs. limit=5.0 +2023-04-27 14:51:41,081 INFO [finetune.py:976] (2/7) Epoch 20, batch 5550, loss[loss=0.2263, simple_loss=0.3008, pruned_loss=0.07587, over 4821.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2396, pruned_loss=0.04907, over 954439.98 frames. ], batch size: 51, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:52:13,272 INFO [finetune.py:976] (2/7) Epoch 20, batch 5600, loss[loss=0.1836, simple_loss=0.2668, pruned_loss=0.0502, over 4730.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2436, pruned_loss=0.04978, over 955894.79 frames. ], batch size: 59, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:52:15,077 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3683, 1.3646, 1.2738, 1.6999, 1.5099, 1.5887, 1.2977, 3.1142], + device='cuda:2'), covar=tensor([0.0629, 0.0990, 0.0980, 0.1322, 0.0766, 0.0546, 0.0906, 0.0195], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0040, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 14:52:23,319 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5568, 1.8247, 1.9108, 2.0297, 1.8510, 1.9246, 2.0417, 1.9574], + device='cuda:2'), covar=tensor([0.3968, 0.5306, 0.4380, 0.4153, 0.5638, 0.7120, 0.5089, 0.4807], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0377, 0.0323, 0.0338, 0.0348, 0.0397, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:52:29,720 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7136, 1.2724, 1.7928, 2.2317, 1.8449, 1.7115, 1.7451, 1.7382], + device='cuda:2'), covar=tensor([0.4414, 0.6850, 0.6078, 0.5404, 0.5473, 0.7566, 0.7957, 0.8659], + device='cuda:2'), in_proj_covar=tensor([0.0428, 0.0411, 0.0504, 0.0505, 0.0456, 0.0485, 0.0494, 0.0497], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:52:32,512 INFO [optim.py:369] (2/7) 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 14:52:42,453 INFO [finetune.py:976] (2/7) Epoch 20, batch 5650, loss[loss=0.1618, simple_loss=0.2374, pruned_loss=0.04308, over 4815.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2461, pruned_loss=0.05017, over 954648.38 frames. ], batch size: 33, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:52:46,426 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:52:46,490 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0917, 2.0096, 2.6637, 2.8859, 2.0599, 1.8521, 2.2195, 1.2029], + device='cuda:2'), covar=tensor([0.0524, 0.0706, 0.0300, 0.0573, 0.0721, 0.1028, 0.0678, 0.0752], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0069, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 14:52:58,919 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0544, 1.4925, 2.0552, 2.4605, 2.1289, 1.9067, 1.9851, 1.9783], + device='cuda:2'), covar=tensor([0.4534, 0.6163, 0.5735, 0.5541, 0.5647, 0.7562, 0.7836, 0.6616], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0412, 0.0506, 0.0506, 0.0458, 0.0486, 0.0496, 0.0499], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:53:08,917 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([5.0415, 5.0055, 3.7599, 5.8128, 5.1227, 5.0737, 2.7339, 5.0814], + device='cuda:2'), covar=tensor([0.1532, 0.0799, 0.2065, 0.0787, 0.2647, 0.1424, 0.4789, 0.1658], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0215, 0.0251, 0.0305, 0.0296, 0.0246, 0.0273, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:53:12,740 INFO [finetune.py:976] (2/7) Epoch 20, batch 5700, loss[loss=0.1468, simple_loss=0.2059, pruned_loss=0.04383, over 4186.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.243, pruned_loss=0.04964, over 938985.26 frames. ], batch size: 18, lr: 3.22e-03, grad_scale: 64.0 +2023-04-27 14:53:39,072 INFO [finetune.py:976] (2/7) Epoch 21, batch 0, loss[loss=0.1801, simple_loss=0.2417, pruned_loss=0.05921, over 4858.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2417, pruned_loss=0.05921, over 4858.00 frames. ], batch size: 31, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:53:39,072 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 14:53:52,792 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5737, 1.5810, 3.7670, 3.5416, 3.3939, 3.5133, 3.6728, 3.3259], + device='cuda:2'), covar=tensor([0.5896, 0.4326, 0.1076, 0.1606, 0.1045, 0.1344, 0.0776, 0.1354], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0306, 0.0403, 0.0403, 0.0347, 0.0407, 0.0311, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:53:56,090 INFO [finetune.py:1010] (2/7) Epoch 21, validation: loss=0.1544, simple_loss=0.2245, pruned_loss=0.04212, over 2265189.00 frames. +2023-04-27 14:53:56,090 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 14:54:02,564 INFO [optim.py:369] (2/7) 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 14:54:24,913 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9898, 1.5948, 2.0981, 2.4975, 2.0991, 1.9453, 2.0352, 1.9784], + device='cuda:2'), covar=tensor([0.4580, 0.6779, 0.6579, 0.5786, 0.6025, 0.7946, 0.7802, 0.8952], + device='cuda:2'), in_proj_covar=tensor([0.0427, 0.0410, 0.0504, 0.0503, 0.0455, 0.0483, 0.0493, 0.0496], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:54:37,746 INFO [finetune.py:976] (2/7) Epoch 21, batch 50, loss[loss=0.1366, simple_loss=0.2177, pruned_loss=0.02772, over 4757.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2505, pruned_loss=0.05422, over 215149.84 frames. ], batch size: 28, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:55:05,759 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4880, 1.8162, 1.8970, 2.0297, 1.8914, 1.9408, 1.9806, 1.9932], + device='cuda:2'), covar=tensor([0.4237, 0.5405, 0.4514, 0.4349, 0.5428, 0.7023, 0.5236, 0.4830], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0376, 0.0322, 0.0337, 0.0348, 0.0395, 0.0358, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 14:55:26,965 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 14:55:37,894 INFO [finetune.py:976] (2/7) Epoch 21, batch 100, loss[loss=0.1477, simple_loss=0.2201, pruned_loss=0.03765, over 4822.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2409, pruned_loss=0.04984, over 377090.31 frames. ], batch size: 51, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:55:46,616 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 14:55:48,245 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 21, batch 150, loss[loss=0.138, simple_loss=0.2122, pruned_loss=0.03193, over 4807.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2379, pruned_loss=0.05002, over 505831.81 frames. ], batch size: 25, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:56:57,803 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114716.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:56:58,596 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 14:57:12,223 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8799, 2.0108, 1.8238, 1.5917, 2.0023, 1.6603, 2.5122, 1.6674], + device='cuda:2'), covar=tensor([0.3792, 0.1700, 0.4369, 0.2847, 0.1691, 0.2378, 0.1434, 0.3902], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0350, 0.0427, 0.0354, 0.0383, 0.0374, 0.0374, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 14:57:22,715 INFO [finetune.py:976] (2/7) Epoch 21, batch 200, loss[loss=0.1768, simple_loss=0.2442, pruned_loss=0.05474, over 4817.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2357, pruned_loss=0.04873, over 606382.47 frames. ], batch size: 40, lr: 3.21e-03, grad_scale: 64.0 +2023-04-27 14:57:26,739 INFO [optim.py:369] (2/7) 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 14:57:38,721 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114777.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:57:42,352 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114783.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:57:52,626 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 14:57:56,200 INFO [finetune.py:976] (2/7) Epoch 21, batch 250, loss[loss=0.1798, simple_loss=0.2374, pruned_loss=0.06107, over 4743.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2403, pruned_loss=0.05051, over 684647.80 frames. ], batch size: 23, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:58:15,122 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114831.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:58:30,014 INFO [finetune.py:976] (2/7) Epoch 21, batch 300, loss[loss=0.172, simple_loss=0.2501, pruned_loss=0.04701, over 4919.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2438, pruned_loss=0.0508, over 745779.21 frames. ], batch size: 36, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:58:34,660 INFO [optim.py:369] (2/7) 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 14:58:42,255 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114871.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:59:03,087 INFO [finetune.py:976] (2/7) Epoch 21, batch 350, loss[loss=0.1659, simple_loss=0.2431, pruned_loss=0.04435, over 4831.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2477, pruned_loss=0.05256, over 793395.16 frames. ], batch size: 49, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:59:22,181 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114932.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 14:59:36,224 INFO [finetune.py:976] (2/7) Epoch 21, batch 400, loss[loss=0.1662, simple_loss=0.2402, pruned_loss=0.04613, over 4905.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2473, pruned_loss=0.05188, over 828356.20 frames. ], batch size: 37, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 14:59:40,904 INFO [optim.py:369] (2/7) 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 15:00:10,133 INFO [finetune.py:976] (2/7) Epoch 21, batch 450, loss[loss=0.2013, simple_loss=0.2593, pruned_loss=0.07162, over 4762.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2448, pruned_loss=0.05056, over 855816.95 frames. ], batch size: 28, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:00:46,048 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6183, 1.4643, 4.5246, 4.3081, 3.9101, 4.2133, 4.1224, 3.9300], + device='cuda:2'), covar=tensor([0.6923, 0.5933, 0.1066, 0.1629, 0.1247, 0.1810, 0.1408, 0.1782], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0307, 0.0405, 0.0404, 0.0348, 0.0409, 0.0311, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:00:49,806 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-04-27 15:00:59,032 INFO [finetune.py:976] (2/7) Epoch 21, batch 500, loss[loss=0.1576, simple_loss=0.2386, pruned_loss=0.03826, over 4794.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2419, pruned_loss=0.04926, over 878307.54 frames. ], batch size: 29, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:01:09,614 INFO [optim.py:369] (2/7) 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 15:01:16,894 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115072.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:01:32,721 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115086.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:01:54,523 INFO [finetune.py:976] (2/7) Epoch 21, batch 550, loss[loss=0.234, simple_loss=0.2966, pruned_loss=0.08567, over 4913.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2411, pruned_loss=0.04978, over 894578.14 frames. ], batch size: 43, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:02:09,100 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8414, 3.3954, 2.7221, 3.0886, 2.1853, 2.1149, 2.8506, 2.2767], + device='cuda:2'), covar=tensor([0.1278, 0.1185, 0.1165, 0.1147, 0.1975, 0.1624, 0.0847, 0.1642], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0211, 0.0169, 0.0203, 0.0199, 0.0185, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:02:53,401 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115147.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:03:02,662 INFO [finetune.py:976] (2/7) Epoch 21, batch 600, loss[loss=0.1815, simple_loss=0.2623, pruned_loss=0.05039, over 4900.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2401, pruned_loss=0.04924, over 906411.99 frames. ], batch size: 43, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:03:03,473 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 15:03:12,536 INFO [optim.py:369] (2/7) 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 15:03:26,517 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-27 15:03:46,920 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1574, 1.7395, 1.9556, 2.4683, 2.4666, 1.9882, 1.5353, 2.2454], + device='cuda:2'), covar=tensor([0.0812, 0.1235, 0.0850, 0.0563, 0.0581, 0.0856, 0.0835, 0.0573], + device='cuda:2'), in_proj_covar=tensor([0.0189, 0.0203, 0.0184, 0.0172, 0.0178, 0.0181, 0.0153, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:04:02,673 INFO [finetune.py:976] (2/7) Epoch 21, batch 650, loss[loss=0.1807, simple_loss=0.2578, pruned_loss=0.05179, over 4918.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2442, pruned_loss=0.05072, over 914947.40 frames. ], batch size: 43, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:04:12,864 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1211, 1.8163, 2.1000, 2.4897, 2.4777, 2.0096, 1.6283, 2.2524], + device='cuda:2'), covar=tensor([0.0894, 0.1175, 0.0727, 0.0557, 0.0628, 0.0924, 0.0813, 0.0548], + device='cuda:2'), in_proj_covar=tensor([0.0189, 0.0203, 0.0185, 0.0172, 0.0178, 0.0181, 0.0153, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:04:17,776 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 700, loss[loss=0.1626, simple_loss=0.2429, pruned_loss=0.04114, over 4777.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2457, pruned_loss=0.05056, over 925141.12 frames. ], batch size: 29, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:04:40,862 INFO [optim.py:369] (2/7) 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 15:05:06,406 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4499, 1.4638, 1.8703, 1.9063, 1.4087, 1.3125, 1.5225, 0.8946], + device='cuda:2'), covar=tensor([0.0638, 0.0728, 0.0401, 0.0587, 0.0821, 0.1114, 0.0668, 0.0709], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0067, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:05:10,550 INFO [finetune.py:976] (2/7) Epoch 21, batch 750, loss[loss=0.147, simple_loss=0.2154, pruned_loss=0.0393, over 4712.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2464, pruned_loss=0.05081, over 929199.82 frames. ], batch size: 23, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:05:44,407 INFO [finetune.py:976] (2/7) Epoch 21, batch 800, loss[loss=0.184, simple_loss=0.2516, pruned_loss=0.05816, over 4788.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2474, pruned_loss=0.05108, over 933434.35 frames. ], batch size: 51, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:05:48,601 INFO [optim.py:369] (2/7) 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 15:05:54,623 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115370.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:05:55,235 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115372.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:01,786 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9556, 2.4761, 2.1743, 2.3994, 1.6430, 2.2105, 2.0380, 1.5460], + device='cuda:2'), covar=tensor([0.2049, 0.1095, 0.0762, 0.1208, 0.3478, 0.1024, 0.1950, 0.2861], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0300, 0.0216, 0.0276, 0.0310, 0.0254, 0.0248, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1425e-04, 1.1930e-04, 8.5454e-05, 1.0903e-04, 1.2583e-04, 1.0051e-04, + 1.0015e-04, 1.0403e-04], device='cuda:2') +2023-04-27 15:06:18,145 INFO [finetune.py:976] (2/7) Epoch 21, batch 850, loss[loss=0.1733, simple_loss=0.2554, pruned_loss=0.04561, over 4761.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2458, pruned_loss=0.05053, over 939807.65 frames. ], batch size: 26, lr: 3.21e-03, grad_scale: 32.0 +2023-04-27 15:06:27,930 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115420.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:35,551 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115431.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:36,161 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:06:40,888 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8636, 1.8529, 1.8315, 1.5607, 2.0130, 1.5972, 2.6162, 1.5976], + device='cuda:2'), covar=tensor([0.3620, 0.1907, 0.4319, 0.2780, 0.1626, 0.2358, 0.1270, 0.4189], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0353, 0.0432, 0.0357, 0.0384, 0.0378, 0.0376, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:06:43,132 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115442.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:07:01,736 INFO [finetune.py:976] (2/7) Epoch 21, batch 900, loss[loss=0.1706, simple_loss=0.2357, pruned_loss=0.05274, over 4752.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.243, pruned_loss=0.04936, over 944894.99 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:07:06,009 INFO [optim.py:369] (2/7) 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 15:07:06,744 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1837, 2.8555, 2.2059, 2.2905, 1.7444, 1.6825, 2.3179, 1.7705], + device='cuda:2'), covar=tensor([0.1244, 0.1274, 0.1256, 0.1530, 0.1937, 0.1536, 0.0868, 0.1662], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0202, 0.0198, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:07:34,516 INFO [finetune.py:976] (2/7) Epoch 21, batch 950, loss[loss=0.1649, simple_loss=0.244, pruned_loss=0.0429, over 4869.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.242, pruned_loss=0.05007, over 947744.62 frames. ], batch size: 31, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:08:06,324 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115527.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:08:28,096 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 15:08:36,292 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 15:08:39,545 INFO [finetune.py:976] (2/7) Epoch 21, batch 1000, loss[loss=0.1705, simple_loss=0.2346, pruned_loss=0.05319, over 4421.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2421, pruned_loss=0.05001, over 947726.16 frames. ], batch size: 19, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:08:47,519 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-04-27 15:08:49,083 INFO [optim.py:369] (2/7) 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 15:09:08,300 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 1050, loss[loss=0.1884, simple_loss=0.2631, pruned_loss=0.05682, over 4752.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2455, pruned_loss=0.05002, over 951540.67 frames. ], batch size: 27, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:09:54,544 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7333, 2.0533, 2.0011, 2.1609, 1.8970, 2.0302, 2.0151, 2.0512], + device='cuda:2'), covar=tensor([0.4229, 0.6398, 0.5167, 0.4764, 0.6188, 0.7785, 0.6808, 0.5735], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0373, 0.0320, 0.0334, 0.0344, 0.0393, 0.0354, 0.0326], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 15:10:13,351 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 15:10:18,959 INFO [finetune.py:976] (2/7) Epoch 21, batch 1100, loss[loss=0.1818, simple_loss=0.2473, pruned_loss=0.05819, over 4713.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2474, pruned_loss=0.05065, over 952644.99 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:10:20,459 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 15:10:23,752 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 21, batch 1150, loss[loss=0.1705, simple_loss=0.2518, pruned_loss=0.04457, over 4897.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2467, pruned_loss=0.0505, over 950146.62 frames. ], batch size: 43, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:11:07,919 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115726.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:11:08,533 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115742.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:11:24,223 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.29 vs. limit=5.0 +2023-04-27 15:11:25,942 INFO [finetune.py:976] (2/7) Epoch 21, batch 1200, loss[loss=0.1891, simple_loss=0.2618, pruned_loss=0.05816, over 4867.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2463, pruned_loss=0.05034, over 952080.93 frames. ], batch size: 34, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:11:28,490 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-04-27 15:11:31,122 INFO [optim.py:369] (2/7) 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 15:11:39,270 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3220, 1.5620, 1.3679, 1.5658, 1.2868, 1.3587, 1.4232, 1.0405], + device='cuda:2'), covar=tensor([0.1620, 0.1206, 0.0903, 0.1171, 0.3499, 0.1147, 0.1649, 0.2050], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0299, 0.0215, 0.0275, 0.0309, 0.0253, 0.0247, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1395e-04, 1.1855e-04, 8.4949e-05, 1.0900e-04, 1.2520e-04, 1.0008e-04, + 9.9799e-05, 1.0348e-04], device='cuda:2') +2023-04-27 15:11:50,226 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115790.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:11:59,754 INFO [finetune.py:976] (2/7) Epoch 21, batch 1250, loss[loss=0.1551, simple_loss=0.216, pruned_loss=0.04708, over 4454.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2451, pruned_loss=0.05065, over 954076.11 frames. ], batch size: 19, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:12:55,528 INFO [finetune.py:976] (2/7) Epoch 21, batch 1300, loss[loss=0.1695, simple_loss=0.2414, pruned_loss=0.04879, over 4945.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2422, pruned_loss=0.04984, over 953970.56 frames. ], batch size: 33, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:12:59,791 INFO [optim.py:369] (2/7) 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 15:13:17,679 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 15:13:21,243 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115892.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:13:21,875 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0566, 1.9210, 2.5726, 2.7480, 1.8254, 1.7588, 2.0871, 1.0932], + device='cuda:2'), covar=tensor([0.0619, 0.0811, 0.0320, 0.0686, 0.0725, 0.1053, 0.0647, 0.0753], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0067, 0.0066, 0.0067, 0.0074, 0.0095, 0.0072, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:13:29,450 INFO [finetune.py:976] (2/7) Epoch 21, batch 1350, loss[loss=0.1973, simple_loss=0.2797, pruned_loss=0.05749, over 4910.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2418, pruned_loss=0.04965, over 954512.05 frames. ], batch size: 43, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:13:47,854 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0169, 2.5896, 1.0529, 1.3359, 1.9476, 1.1951, 3.4833, 1.7130], + device='cuda:2'), covar=tensor([0.0761, 0.0748, 0.0812, 0.1254, 0.0511, 0.1089, 0.0320, 0.0683], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0065, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 15:14:34,415 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115953.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:14:34,901 INFO [finetune.py:976] (2/7) Epoch 21, batch 1400, loss[loss=0.1764, simple_loss=0.2547, pruned_loss=0.04904, over 4740.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2452, pruned_loss=0.05049, over 953763.37 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:14:44,917 INFO [optim.py:369] (2/7) 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 15:14:52,698 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9657, 1.5387, 2.0532, 2.3651, 2.0096, 1.9502, 2.0332, 2.0066], + device='cuda:2'), covar=tensor([0.4301, 0.5995, 0.5624, 0.5607, 0.5622, 0.7494, 0.7250, 0.7263], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0413, 0.0506, 0.0506, 0.0458, 0.0487, 0.0496, 0.0500], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:14:56,841 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-04-27 15:15:07,198 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-04-27 15:15:22,117 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9068, 2.5248, 2.0189, 2.0012, 1.5595, 1.5266, 2.0803, 1.5087], + device='cuda:2'), covar=tensor([0.1358, 0.1221, 0.1286, 0.1493, 0.1917, 0.1563, 0.0821, 0.1731], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0203, 0.0199, 0.0184, 0.0155, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:15:26,319 INFO [finetune.py:976] (2/7) Epoch 21, batch 1450, loss[loss=0.16, simple_loss=0.2389, pruned_loss=0.04055, over 4898.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2461, pruned_loss=0.05058, over 952018.45 frames. ], batch size: 43, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:15:33,444 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116014.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:15:42,218 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116026.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:15:42,818 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 1500, loss[loss=0.229, simple_loss=0.2918, pruned_loss=0.08314, over 4810.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2481, pruned_loss=0.05204, over 951409.79 frames. ], batch size: 39, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:16:05,187 INFO [optim.py:369] (2/7) 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 15:16:08,376 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7449, 1.7691, 0.8972, 1.3845, 1.9489, 1.5708, 1.4501, 1.5263], + device='cuda:2'), covar=tensor([0.0472, 0.0362, 0.0329, 0.0541, 0.0243, 0.0492, 0.0479, 0.0556], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0051, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 15:16:13,682 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116074.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:14,790 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116075.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:14,857 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116075.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 15:16:33,606 INFO [finetune.py:976] (2/7) Epoch 21, batch 1550, loss[loss=0.1865, simple_loss=0.2504, pruned_loss=0.06132, over 4795.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2481, pruned_loss=0.05185, over 952813.17 frames. ], batch size: 45, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:16:34,358 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116105.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:41,987 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116116.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:16:46,386 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-04-27 15:16:59,564 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9316, 2.5144, 1.9999, 1.8370, 1.4294, 1.4503, 2.0766, 1.4101], + device='cuda:2'), covar=tensor([0.1486, 0.1258, 0.1348, 0.1600, 0.2269, 0.1835, 0.0881, 0.1910], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0169, 0.0204, 0.0199, 0.0184, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:17:06,680 INFO [finetune.py:976] (2/7) Epoch 21, batch 1600, loss[loss=0.1683, simple_loss=0.238, pruned_loss=0.04929, over 4821.00 frames. ], tot_loss[loss=0.173, simple_loss=0.245, pruned_loss=0.05046, over 953038.47 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:17:10,637 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-04-27 15:17:10,955 INFO [optim.py:369] (2/7) 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 15:17:14,619 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116166.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:17:21,319 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116177.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:17:39,788 INFO [finetune.py:976] (2/7) Epoch 21, batch 1650, loss[loss=0.1684, simple_loss=0.2441, pruned_loss=0.04634, over 4905.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2426, pruned_loss=0.04995, over 954446.47 frames. ], batch size: 37, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:18:04,933 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8389, 2.4878, 1.8686, 1.8292, 1.3523, 1.3834, 1.9613, 1.2990], + device='cuda:2'), covar=tensor([0.1473, 0.1241, 0.1391, 0.1631, 0.2155, 0.1858, 0.0904, 0.1939], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0212, 0.0170, 0.0205, 0.0200, 0.0184, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:18:14,848 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116248.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 15:18:18,417 INFO [finetune.py:976] (2/7) Epoch 21, batch 1700, loss[loss=0.1685, simple_loss=0.236, pruned_loss=0.05052, over 4841.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2404, pruned_loss=0.04909, over 955772.52 frames. ], batch size: 47, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:18:28,119 INFO [optim.py:369] (2/7) 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 15:19:10,034 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.91 vs. limit=5.0 +2023-04-27 15:19:31,551 INFO [finetune.py:976] (2/7) Epoch 21, batch 1750, loss[loss=0.1734, simple_loss=0.2511, pruned_loss=0.04782, over 4915.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2425, pruned_loss=0.05022, over 955925.77 frames. ], batch size: 37, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:19:55,826 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 1800, loss[loss=0.1962, simple_loss=0.274, pruned_loss=0.05915, over 4751.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2447, pruned_loss=0.05033, over 955822.16 frames. ], batch size: 59, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:20:38,070 INFO [optim.py:369] (2/7) 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 15:20:43,065 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1137, 2.5208, 2.0541, 2.3663, 1.7091, 2.1823, 2.1532, 1.7049], + device='cuda:2'), covar=tensor([0.1701, 0.1112, 0.0766, 0.1066, 0.3241, 0.1121, 0.1810, 0.2576], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0299, 0.0214, 0.0276, 0.0309, 0.0252, 0.0247, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1382e-04, 1.1849e-04, 8.4650e-05, 1.0924e-04, 1.2538e-04, 9.9786e-05, + 9.9643e-05, 1.0337e-04], device='cuda:2') +2023-04-27 15:20:43,633 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116370.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:20:46,056 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4092, 1.9353, 2.3003, 2.8168, 2.3187, 1.8198, 1.6420, 2.2537], + device='cuda:2'), covar=tensor([0.3308, 0.3100, 0.1632, 0.2264, 0.2502, 0.2674, 0.3898, 0.1789], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0243, 0.0225, 0.0311, 0.0217, 0.0230, 0.0226, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 15:20:52,649 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5656, 1.3669, 0.6132, 1.2305, 1.4831, 1.4281, 1.3005, 1.3422], + device='cuda:2'), covar=tensor([0.0506, 0.0415, 0.0392, 0.0589, 0.0292, 0.0527, 0.0546, 0.0574], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 15:20:53,264 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 1850, loss[loss=0.1709, simple_loss=0.2382, pruned_loss=0.05181, over 4764.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2461, pruned_loss=0.05119, over 954422.83 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:21:35,292 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116446.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:21:35,933 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1616, 1.7765, 2.0340, 2.4677, 2.4781, 1.9920, 1.6437, 2.1625], + device='cuda:2'), covar=tensor([0.0888, 0.1099, 0.0720, 0.0562, 0.0662, 0.0964, 0.0805, 0.0631], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0201, 0.0184, 0.0172, 0.0177, 0.0180, 0.0152, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:21:39,701 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-27 15:21:40,090 INFO [finetune.py:976] (2/7) Epoch 21, batch 1900, loss[loss=0.1573, simple_loss=0.2275, pruned_loss=0.0435, over 4760.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2466, pruned_loss=0.05058, over 955468.69 frames. ], batch size: 28, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:21:45,215 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116461.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:21:52,091 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 1950, loss[loss=0.1442, simple_loss=0.2208, pruned_loss=0.03383, over 4746.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2455, pruned_loss=0.05023, over 953380.59 frames. ], batch size: 27, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:22:14,297 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116505.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:22:16,017 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116507.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 15:22:43,311 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116548.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:22:46,839 INFO [finetune.py:976] (2/7) Epoch 21, batch 2000, loss[loss=0.1285, simple_loss=0.2057, pruned_loss=0.02567, over 4796.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2431, pruned_loss=0.04951, over 953131.12 frames. ], batch size: 29, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:22:48,168 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7443, 2.0518, 1.6801, 1.4055, 1.3177, 1.3273, 1.7397, 1.2365], + device='cuda:2'), covar=tensor([0.1586, 0.1237, 0.1439, 0.1697, 0.2211, 0.1938, 0.0987, 0.1988], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0210, 0.0169, 0.0203, 0.0199, 0.0185, 0.0155, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:22:51,556 INFO [optim.py:369] (2/7) 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 15:22:55,205 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116566.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:23:14,413 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116596.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:23:18,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6139, 1.3372, 4.5454, 4.2278, 3.9296, 4.3639, 4.2432, 4.0172], + device='cuda:2'), covar=tensor([0.6916, 0.6379, 0.0936, 0.1733, 0.1095, 0.1915, 0.1018, 0.1250], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0307, 0.0403, 0.0404, 0.0347, 0.0408, 0.0311, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:23:20,569 INFO [finetune.py:976] (2/7) Epoch 21, batch 2050, loss[loss=0.2263, simple_loss=0.2697, pruned_loss=0.09149, over 4427.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2405, pruned_loss=0.04909, over 954397.97 frames. ], batch size: 19, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:23:30,133 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116618.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:23:50,736 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5133, 3.4396, 2.5029, 3.9215, 3.4300, 3.3920, 1.5143, 3.4274], + device='cuda:2'), covar=tensor([0.1724, 0.1431, 0.3051, 0.2075, 0.4502, 0.1940, 0.5199, 0.2491], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0216, 0.0254, 0.0308, 0.0300, 0.0248, 0.0277, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 15:23:59,069 INFO [finetune.py:976] (2/7) Epoch 21, batch 2100, loss[loss=0.15, simple_loss=0.2101, pruned_loss=0.04496, over 4713.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2405, pruned_loss=0.0497, over 955543.66 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:24:03,925 INFO [optim.py:369] (2/7) 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 15:24:21,167 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116670.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 15:24:32,887 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:24:32,930 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:25:00,268 INFO [finetune.py:976] (2/7) Epoch 21, batch 2150, loss[loss=0.1986, simple_loss=0.2668, pruned_loss=0.06524, over 4827.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2437, pruned_loss=0.05128, over 954955.67 frames. ], batch size: 30, lr: 3.20e-03, grad_scale: 32.0 +2023-04-27 15:25:05,392 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:25:07,109 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0832, 2.6495, 3.0025, 3.4014, 3.1870, 2.8082, 2.3087, 3.0214], + device='cuda:2'), covar=tensor([0.0607, 0.0858, 0.0497, 0.0520, 0.0529, 0.0732, 0.0698, 0.0463], + device='cuda:2'), in_proj_covar=tensor([0.0186, 0.0200, 0.0183, 0.0172, 0.0176, 0.0179, 0.0151, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:25:21,295 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:25:27,191 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-27 15:25:37,855 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9123, 1.5620, 1.7910, 2.1628, 2.2754, 1.7200, 1.4048, 1.8886], + device='cuda:2'), covar=tensor([0.0841, 0.1189, 0.0805, 0.0687, 0.0665, 0.0858, 0.0827, 0.0620], + device='cuda:2'), in_proj_covar=tensor([0.0186, 0.0199, 0.0183, 0.0172, 0.0176, 0.0179, 0.0151, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:26:05,620 INFO [finetune.py:976] (2/7) Epoch 21, batch 2200, loss[loss=0.1277, simple_loss=0.1976, pruned_loss=0.02896, over 4771.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.246, pruned_loss=0.05205, over 956032.89 frames. ], batch size: 26, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:26:10,848 INFO [optim.py:369] (2/7) 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 15:26:10,964 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116761.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:13,909 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:18,724 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116772.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:18,881 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-27 15:26:38,509 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116802.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 15:26:38,537 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 2250, loss[loss=0.1333, simple_loss=0.197, pruned_loss=0.03477, over 3998.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2475, pruned_loss=0.0524, over 955328.79 frames. ], batch size: 17, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:26:42,837 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116809.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:26:50,605 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:27:13,183 INFO [finetune.py:976] (2/7) Epoch 21, batch 2300, loss[loss=0.1339, simple_loss=0.2126, pruned_loss=0.02757, over 4891.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2472, pruned_loss=0.05156, over 956189.83 frames. ], batch size: 32, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:27:17,444 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116861.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:27:18,812 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116863.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:27:24,041 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 15:27:43,652 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9359, 2.3086, 1.9811, 2.2422, 1.7063, 1.9615, 1.9065, 1.5507], + device='cuda:2'), covar=tensor([0.1775, 0.1459, 0.0757, 0.1186, 0.3249, 0.1103, 0.1926, 0.2580], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0300, 0.0216, 0.0277, 0.0311, 0.0254, 0.0247, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1444e-04, 1.1908e-04, 8.5442e-05, 1.0963e-04, 1.2603e-04, 1.0050e-04, + 9.9664e-05, 1.0365e-04], device='cuda:2') +2023-04-27 15:27:46,961 INFO [finetune.py:976] (2/7) Epoch 21, batch 2350, loss[loss=0.175, simple_loss=0.2491, pruned_loss=0.05039, over 4754.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2454, pruned_loss=0.05079, over 956764.74 frames. ], batch size: 59, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:27:54,442 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5039, 1.4366, 1.8574, 1.8126, 1.4290, 1.2366, 1.4818, 0.8622], + device='cuda:2'), covar=tensor([0.0614, 0.0551, 0.0372, 0.0498, 0.0684, 0.1121, 0.0563, 0.0667], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0068, 0.0066, 0.0067, 0.0074, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:28:20,896 INFO [finetune.py:976] (2/7) Epoch 21, batch 2400, loss[loss=0.1591, simple_loss=0.2276, pruned_loss=0.04531, over 4829.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2435, pruned_loss=0.05045, over 956750.44 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:28:25,119 INFO [optim.py:369] (2/7) 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 15:28:34,740 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:28:37,807 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116979.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:28:45,775 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 15:28:54,617 INFO [finetune.py:976] (2/7) Epoch 21, batch 2450, loss[loss=0.1847, simple_loss=0.2435, pruned_loss=0.063, over 4898.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2405, pruned_loss=0.04935, over 955817.22 frames. ], batch size: 35, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:28:57,625 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6072, 1.2935, 4.5491, 4.2562, 3.9505, 4.2684, 4.1791, 4.0009], + device='cuda:2'), covar=tensor([0.7346, 0.6474, 0.1270, 0.2114, 0.1233, 0.2442, 0.1372, 0.1594], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0306, 0.0403, 0.0401, 0.0346, 0.0406, 0.0309, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:29:05,098 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3618, 1.8490, 2.2769, 2.7165, 2.2758, 1.8267, 1.4771, 1.9908], + device='cuda:2'), covar=tensor([0.3126, 0.2976, 0.1506, 0.2090, 0.2337, 0.2462, 0.3990, 0.2037], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0245, 0.0227, 0.0314, 0.0219, 0.0232, 0.0229, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 15:29:10,687 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:29:13,195 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5558, 0.6877, 1.5316, 1.9665, 1.6463, 1.5094, 1.5380, 1.5445], + device='cuda:2'), covar=tensor([0.3989, 0.6127, 0.5573, 0.5354, 0.5233, 0.6729, 0.6721, 0.7560], + device='cuda:2'), in_proj_covar=tensor([0.0428, 0.0410, 0.0504, 0.0504, 0.0455, 0.0485, 0.0493, 0.0498], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:29:28,081 INFO [finetune.py:976] (2/7) Epoch 21, batch 2500, loss[loss=0.1632, simple_loss=0.2415, pruned_loss=0.0424, over 4809.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2419, pruned_loss=0.05029, over 953560.75 frames. ], batch size: 45, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:29:32,764 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:30:15,615 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8004, 2.2318, 2.1078, 2.8228, 3.0003, 2.5247, 2.4237, 1.9626], + device='cuda:2'), covar=tensor([0.1374, 0.1467, 0.1540, 0.1836, 0.0863, 0.1560, 0.2064, 0.1971], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0313, 0.0352, 0.0290, 0.0327, 0.0311, 0.0304, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4001e-05, 6.4776e-05, 7.4238e-05, 5.8537e-05, 6.7556e-05, 6.5136e-05, + 6.3807e-05, 7.9243e-05], device='cuda:2') +2023-04-27 15:30:29,012 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117102.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:30:30,116 INFO [finetune.py:976] (2/7) Epoch 21, batch 2550, loss[loss=0.2197, simple_loss=0.285, pruned_loss=0.07717, over 4856.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.246, pruned_loss=0.05122, over 954675.58 frames. ], batch size: 44, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:31:21,834 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-27 15:31:33,695 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117150.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:31:41,894 INFO [finetune.py:976] (2/7) Epoch 21, batch 2600, loss[loss=0.1869, simple_loss=0.2659, pruned_loss=0.05399, over 4905.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2467, pruned_loss=0.05153, over 955236.93 frames. ], batch size: 43, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:31:44,414 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117158.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:31:52,558 INFO [optim.py:369] (2/7) 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 15:31:52,651 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117161.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:32:09,289 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-27 15:32:24,662 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9264, 2.4520, 2.0640, 2.3030, 1.7970, 2.0707, 2.0135, 1.6114], + device='cuda:2'), covar=tensor([0.1710, 0.1041, 0.0763, 0.1201, 0.3106, 0.1026, 0.1765, 0.2364], + device='cuda:2'), in_proj_covar=tensor([0.0287, 0.0300, 0.0216, 0.0278, 0.0311, 0.0253, 0.0247, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1490e-04, 1.1882e-04, 8.5663e-05, 1.0981e-04, 1.2604e-04, 1.0042e-04, + 9.9591e-05, 1.0399e-04], device='cuda:2') +2023-04-27 15:32:27,548 INFO [finetune.py:976] (2/7) Epoch 21, batch 2650, loss[loss=0.1698, simple_loss=0.2505, pruned_loss=0.04452, over 4903.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2477, pruned_loss=0.05194, over 955064.37 frames. ], batch size: 37, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:32:30,679 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117209.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:32:35,555 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1808, 1.5909, 2.0492, 2.6211, 2.0535, 1.5766, 1.4459, 1.8690], + device='cuda:2'), covar=tensor([0.3232, 0.3293, 0.1744, 0.2204, 0.2742, 0.2729, 0.4125, 0.2159], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0242, 0.0224, 0.0310, 0.0217, 0.0230, 0.0226, 0.0182], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 15:32:39,139 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 2700, loss[loss=0.1945, simple_loss=0.2618, pruned_loss=0.06357, over 4876.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2471, pruned_loss=0.05158, over 951067.11 frames. ], batch size: 31, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:33:06,002 INFO [optim.py:369] (2/7) 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 15:33:10,259 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4115, 1.2887, 1.6232, 1.5947, 1.3687, 1.2503, 1.2771, 0.7438], + device='cuda:2'), covar=tensor([0.0562, 0.0618, 0.0425, 0.0519, 0.0727, 0.1160, 0.0564, 0.0658], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0067, 0.0066, 0.0067, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:33:15,075 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117274.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:33:19,979 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117282.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:33:20,612 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6409, 0.7098, 1.5095, 2.0226, 1.7069, 1.5363, 1.5305, 1.5592], + device='cuda:2'), covar=tensor([0.4388, 0.6062, 0.5444, 0.5486, 0.5388, 0.6894, 0.7202, 0.7405], + device='cuda:2'), in_proj_covar=tensor([0.0426, 0.0410, 0.0503, 0.0502, 0.0455, 0.0484, 0.0492, 0.0498], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:33:35,171 INFO [finetune.py:976] (2/7) Epoch 21, batch 2750, loss[loss=0.1544, simple_loss=0.2145, pruned_loss=0.04716, over 4845.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2448, pruned_loss=0.05112, over 950388.33 frames. ], batch size: 49, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:33:37,022 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0543, 1.5163, 1.5721, 1.8005, 2.1847, 1.7940, 1.5041, 1.4333], + device='cuda:2'), covar=tensor([0.1451, 0.1750, 0.2110, 0.1340, 0.0891, 0.1682, 0.2185, 0.2572], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0313, 0.0352, 0.0290, 0.0327, 0.0310, 0.0304, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.4018e-05, 6.4740e-05, 7.4286e-05, 5.8644e-05, 6.7532e-05, 6.4973e-05, + 6.3785e-05, 7.9050e-05], device='cuda:2') +2023-04-27 15:33:42,707 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 15:33:43,104 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8617, 1.2133, 3.2824, 3.0831, 2.8935, 3.1761, 3.1373, 2.9054], + device='cuda:2'), covar=tensor([0.6954, 0.5044, 0.1283, 0.1905, 0.1365, 0.1750, 0.2243, 0.1501], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0307, 0.0403, 0.0404, 0.0346, 0.0407, 0.0310, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:33:47,709 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117322.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:33:52,058 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 2800, loss[loss=0.1632, simple_loss=0.2137, pruned_loss=0.05639, over 4050.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2406, pruned_loss=0.04927, over 950892.64 frames. ], batch size: 17, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:34:12,916 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:34:24,905 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4196, 1.7303, 1.7281, 1.7838, 1.7108, 1.8468, 1.8634, 1.7627], + device='cuda:2'), covar=tensor([0.3608, 0.5036, 0.4505, 0.4438, 0.5788, 0.7334, 0.4834, 0.4785], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0373, 0.0322, 0.0335, 0.0345, 0.0393, 0.0355, 0.0327], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 15:34:29,155 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 15:34:31,566 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:34:41,632 INFO [finetune.py:976] (2/7) Epoch 21, batch 2850, loss[loss=0.1459, simple_loss=0.2291, pruned_loss=0.03138, over 4840.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2385, pruned_loss=0.04828, over 952483.51 frames. ], batch size: 30, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:34:41,832 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 15:34:44,560 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117408.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:34:48,768 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-04-27 15:34:54,758 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8985, 2.4838, 1.8915, 1.9365, 1.4428, 1.4710, 2.0244, 1.3046], + device='cuda:2'), covar=tensor([0.1495, 0.1295, 0.1381, 0.1538, 0.2165, 0.1825, 0.0943, 0.1958], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0212, 0.0170, 0.0205, 0.0201, 0.0186, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:35:14,767 INFO [finetune.py:976] (2/7) Epoch 21, batch 2900, loss[loss=0.1373, simple_loss=0.2238, pruned_loss=0.02538, over 4823.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2425, pruned_loss=0.04997, over 950481.68 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:35:17,822 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 15:35:54,310 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8194, 1.5088, 1.6934, 2.1088, 2.1136, 1.6798, 1.3676, 1.8022], + device='cuda:2'), covar=tensor([0.0755, 0.1274, 0.0890, 0.0512, 0.0551, 0.0817, 0.0762, 0.0594], + device='cuda:2'), in_proj_covar=tensor([0.0189, 0.0203, 0.0186, 0.0175, 0.0178, 0.0182, 0.0153, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:36:16,927 INFO [finetune.py:976] (2/7) Epoch 21, batch 2950, loss[loss=0.2297, simple_loss=0.2805, pruned_loss=0.08943, over 4817.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2468, pruned_loss=0.05166, over 952069.86 frames. ], batch size: 30, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:36:18,221 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117506.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:37:23,420 INFO [finetune.py:976] (2/7) Epoch 21, batch 3000, loss[loss=0.1486, simple_loss=0.2375, pruned_loss=0.02986, over 4821.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2472, pruned_loss=0.0513, over 953327.41 frames. ], batch size: 39, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:37:23,420 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 15:37:43,669 INFO [finetune.py:1010] (2/7) Epoch 21, validation: loss=0.1531, simple_loss=0.2228, pruned_loss=0.04164, over 2265189.00 frames. +2023-04-27 15:37:43,670 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 15:37:54,336 INFO [optim.py:369] (2/7) 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 15:38:15,507 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117577.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:38:47,576 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3348, 1.3513, 1.4008, 1.6559, 1.6355, 1.2495, 0.9669, 1.4357], + device='cuda:2'), covar=tensor([0.0950, 0.1362, 0.0969, 0.0671, 0.0767, 0.0918, 0.0963, 0.0678], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0202, 0.0185, 0.0174, 0.0178, 0.0181, 0.0153, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:38:49,964 INFO [finetune.py:976] (2/7) Epoch 21, batch 3050, loss[loss=0.1753, simple_loss=0.2466, pruned_loss=0.05199, over 4739.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.248, pruned_loss=0.05118, over 953347.69 frames. ], batch size: 54, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:39:08,806 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1179, 2.4348, 2.4736, 2.7319, 2.6845, 2.5514, 2.4076, 4.9594], + device='cuda:2'), covar=tensor([0.0462, 0.0657, 0.0619, 0.0994, 0.0478, 0.0576, 0.0598, 0.0165], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0056], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 15:39:24,972 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3446, 1.7353, 2.2092, 2.6236, 2.1839, 1.7247, 1.5655, 2.0166], + device='cuda:2'), covar=tensor([0.2918, 0.3003, 0.1547, 0.2266, 0.2497, 0.2592, 0.3873, 0.1966], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0244, 0.0226, 0.0313, 0.0218, 0.0231, 0.0227, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 15:39:28,407 INFO [finetune.py:976] (2/7) Epoch 21, batch 3100, loss[loss=0.1709, simple_loss=0.2482, pruned_loss=0.0468, over 4817.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2452, pruned_loss=0.04993, over 953146.98 frames. ], batch size: 38, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:39:33,614 INFO [optim.py:369] (2/7) 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 15:39:49,470 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:40:01,887 INFO [finetune.py:976] (2/7) Epoch 21, batch 3150, loss[loss=0.1662, simple_loss=0.2349, pruned_loss=0.04873, over 4907.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2424, pruned_loss=0.04943, over 954016.29 frames. ], batch size: 36, lr: 3.19e-03, grad_scale: 64.0 +2023-04-27 15:40:14,504 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 15:40:21,796 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-04-27 15:40:33,784 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6359, 1.4153, 1.9486, 2.0077, 1.5766, 1.3059, 1.5969, 0.9412], + device='cuda:2'), covar=tensor([0.0466, 0.0642, 0.0386, 0.0468, 0.0619, 0.1063, 0.0586, 0.0736], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0067, 0.0066, 0.0067, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:40:34,883 INFO [finetune.py:976] (2/7) Epoch 21, batch 3200, loss[loss=0.1553, simple_loss=0.2263, pruned_loss=0.04214, over 4911.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2395, pruned_loss=0.04865, over 953775.88 frames. ], batch size: 36, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:40:40,630 INFO [optim.py:369] (2/7) 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 15:41:31,202 INFO [finetune.py:976] (2/7) Epoch 21, batch 3250, loss[loss=0.1445, simple_loss=0.2277, pruned_loss=0.03062, over 4937.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2388, pruned_loss=0.04857, over 951921.26 frames. ], batch size: 38, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:41:39,635 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 15:42:33,317 INFO [finetune.py:976] (2/7) Epoch 21, batch 3300, loss[loss=0.185, simple_loss=0.2624, pruned_loss=0.05377, over 4898.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2422, pruned_loss=0.04916, over 953969.38 frames. ], batch size: 35, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:42:43,084 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 15:42:45,024 INFO [optim.py:369] (2/7) 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 15:42:52,595 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8485, 1.7243, 1.4118, 1.6984, 2.0815, 1.7160, 1.4433, 1.2847], + device='cuda:2'), covar=tensor([0.1395, 0.1117, 0.1630, 0.1322, 0.0712, 0.1401, 0.1842, 0.2072], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0312, 0.0351, 0.0289, 0.0326, 0.0309, 0.0303, 0.0371], + device='cuda:2'), out_proj_covar=tensor([6.3528e-05, 6.4567e-05, 7.4050e-05, 5.8239e-05, 6.7293e-05, 6.4814e-05, + 6.3536e-05, 7.8859e-05], device='cuda:2') +2023-04-27 15:43:01,358 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117877.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:43:12,918 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-27 15:43:18,017 INFO [finetune.py:976] (2/7) Epoch 21, batch 3350, loss[loss=0.2216, simple_loss=0.2822, pruned_loss=0.08052, over 4802.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2438, pruned_loss=0.04993, over 954021.88 frames. ], batch size: 45, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:43:34,888 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9201, 2.2887, 2.0050, 2.2551, 1.7279, 2.0647, 2.0129, 1.5388], + device='cuda:2'), covar=tensor([0.1811, 0.1236, 0.0868, 0.1163, 0.3211, 0.1041, 0.1755, 0.2710], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0299, 0.0214, 0.0277, 0.0310, 0.0252, 0.0245, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1388e-04, 1.1830e-04, 8.4852e-05, 1.0955e-04, 1.2567e-04, 9.9786e-05, + 9.9055e-05, 1.0410e-04], device='cuda:2') +2023-04-27 15:43:36,180 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.46 vs. limit=5.0 +2023-04-27 15:43:37,870 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:43:55,692 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0590, 1.9267, 2.4254, 2.7198, 1.9772, 1.7942, 1.9960, 1.0178], + device='cuda:2'), covar=tensor([0.0519, 0.0794, 0.0397, 0.0587, 0.0718, 0.0964, 0.0727, 0.0756], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0068, 0.0067, 0.0067, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:44:02,872 INFO [finetune.py:976] (2/7) Epoch 21, batch 3400, loss[loss=0.1695, simple_loss=0.2536, pruned_loss=0.04272, over 4818.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2453, pruned_loss=0.05022, over 954762.60 frames. ], batch size: 45, lr: 3.19e-03, grad_scale: 32.0 +2023-04-27 15:44:13,573 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117985.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:44:57,989 INFO [finetune.py:976] (2/7) Epoch 21, batch 3450, loss[loss=0.1967, simple_loss=0.2556, pruned_loss=0.0689, over 4758.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2451, pruned_loss=0.04984, over 955713.23 frames. ], batch size: 28, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:45:11,144 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3531, 1.5530, 1.4177, 1.6149, 1.4094, 1.4181, 1.4197, 1.1309], + device='cuda:2'), covar=tensor([0.1569, 0.1256, 0.0855, 0.1136, 0.3328, 0.1096, 0.1691, 0.2297], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0298, 0.0214, 0.0276, 0.0309, 0.0251, 0.0244, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1346e-04, 1.1794e-04, 8.4689e-05, 1.0931e-04, 1.2520e-04, 9.9293e-05, + 9.8641e-05, 1.0392e-04], device='cuda:2') +2023-04-27 15:45:17,644 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118033.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:45:22,436 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:45:31,328 INFO [finetune.py:976] (2/7) Epoch 21, batch 3500, loss[loss=0.1525, simple_loss=0.2179, pruned_loss=0.04351, over 4217.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2439, pruned_loss=0.05001, over 955875.18 frames. ], batch size: 18, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:45:36,183 INFO [optim.py:369] (2/7) 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 15:45:39,286 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-04-27 15:45:51,833 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.65 vs. limit=5.0 +2023-04-27 15:46:03,439 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 3550, loss[loss=0.1853, simple_loss=0.2504, pruned_loss=0.06011, over 4907.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.241, pruned_loss=0.04903, over 953893.28 frames. ], batch size: 35, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:46:19,149 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4366, 1.9116, 1.7263, 2.3499, 2.4926, 2.1004, 2.0127, 1.8824], + device='cuda:2'), covar=tensor([0.1880, 0.1617, 0.2022, 0.1355, 0.1320, 0.1796, 0.2316, 0.2309], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0311, 0.0350, 0.0287, 0.0325, 0.0308, 0.0301, 0.0370], + device='cuda:2'), out_proj_covar=tensor([6.3514e-05, 6.4299e-05, 7.3854e-05, 5.8001e-05, 6.7082e-05, 6.4491e-05, + 6.3200e-05, 7.8559e-05], device='cuda:2') +2023-04-27 15:46:39,040 INFO [finetune.py:976] (2/7) Epoch 21, batch 3600, loss[loss=0.1641, simple_loss=0.2253, pruned_loss=0.05142, over 4761.00 frames. ], tot_loss[loss=0.167, simple_loss=0.238, pruned_loss=0.04798, over 956097.21 frames. ], batch size: 26, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:46:43,924 INFO [optim.py:369] (2/7) 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 15:46:47,629 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118168.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:46:58,025 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4429, 1.9228, 2.2963, 2.7664, 2.3654, 1.7933, 1.6322, 2.2243], + device='cuda:2'), covar=tensor([0.3085, 0.2919, 0.1571, 0.2174, 0.2445, 0.2658, 0.3916, 0.1717], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0245, 0.0227, 0.0315, 0.0219, 0.0232, 0.0228, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 15:47:39,451 INFO [finetune.py:976] (2/7) Epoch 21, batch 3650, loss[loss=0.1792, simple_loss=0.2575, pruned_loss=0.05043, over 4757.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2399, pruned_loss=0.04871, over 955643.39 frames. ], batch size: 59, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:48:09,464 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118229.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 15:48:43,767 INFO [finetune.py:976] (2/7) Epoch 21, batch 3700, loss[loss=0.1449, simple_loss=0.23, pruned_loss=0.02993, over 4820.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2433, pruned_loss=0.04997, over 954966.69 frames. ], batch size: 38, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:48:54,005 INFO [optim.py:369] (2/7) 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 15:48:58,390 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118269.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:49:16,501 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:49:27,410 INFO [finetune.py:976] (2/7) Epoch 21, batch 3750, loss[loss=0.1693, simple_loss=0.2519, pruned_loss=0.04339, over 4894.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2452, pruned_loss=0.0504, over 954211.55 frames. ], batch size: 37, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:49:34,388 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 15:49:43,198 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:49:43,811 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118330.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:49:44,408 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0815, 2.4116, 1.1016, 1.4240, 1.9241, 1.3737, 2.8537, 1.5677], + device='cuda:2'), covar=tensor([0.0586, 0.0707, 0.0751, 0.1005, 0.0424, 0.0822, 0.0215, 0.0591], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 15:50:10,284 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 3800, loss[loss=0.2022, simple_loss=0.2714, pruned_loss=0.06651, over 4844.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2458, pruned_loss=0.05004, over 954553.25 frames. ], batch size: 30, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:50:23,124 INFO [optim.py:369] (2/7) 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 15:50:47,954 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6175, 1.3965, 1.3124, 1.4633, 1.8572, 1.5287, 1.2775, 1.2465], + device='cuda:2'), covar=tensor([0.1600, 0.1198, 0.1549, 0.1259, 0.0698, 0.1416, 0.1913, 0.2197], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0312, 0.0351, 0.0290, 0.0327, 0.0309, 0.0302, 0.0371], + device='cuda:2'), out_proj_covar=tensor([6.3775e-05, 6.4531e-05, 7.4231e-05, 5.8488e-05, 6.7393e-05, 6.4818e-05, + 6.3317e-05, 7.8813e-05], device='cuda:2') +2023-04-27 15:50:52,590 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:50:56,189 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:51:02,177 INFO [finetune.py:976] (2/7) Epoch 21, batch 3850, loss[loss=0.2022, simple_loss=0.2733, pruned_loss=0.06556, over 4891.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2443, pruned_loss=0.04922, over 955309.77 frames. ], batch size: 43, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:51:35,149 INFO [finetune.py:976] (2/7) Epoch 21, batch 3900, loss[loss=0.1571, simple_loss=0.2242, pruned_loss=0.04501, over 4848.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2423, pruned_loss=0.04893, over 957865.01 frames. ], batch size: 49, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:51:40,429 INFO [optim.py:369] (2/7) 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 15:52:07,430 INFO [finetune.py:976] (2/7) Epoch 21, batch 3950, loss[loss=0.1486, simple_loss=0.2241, pruned_loss=0.03658, over 4876.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2397, pruned_loss=0.04802, over 957792.16 frames. ], batch size: 31, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:52:21,114 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118524.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 15:52:40,579 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-27 15:52:40,872 INFO [finetune.py:976] (2/7) Epoch 21, batch 4000, loss[loss=0.2105, simple_loss=0.2764, pruned_loss=0.07225, over 4823.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2395, pruned_loss=0.04865, over 956850.94 frames. ], batch size: 39, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:52:47,270 INFO [optim.py:369] (2/7) 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 15:53:30,868 INFO [finetune.py:976] (2/7) Epoch 21, batch 4050, loss[loss=0.2053, simple_loss=0.2775, pruned_loss=0.0665, over 4828.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2421, pruned_loss=0.0499, over 955596.85 frames. ], batch size: 30, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:53:44,367 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-04-27 15:53:55,021 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9753, 1.1233, 1.5075, 1.6563, 1.6065, 1.7122, 1.5545, 1.5350], + device='cuda:2'), covar=tensor([0.3344, 0.4808, 0.3802, 0.3990, 0.4874, 0.6588, 0.4322, 0.4287], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0374, 0.0324, 0.0337, 0.0348, 0.0395, 0.0357, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 15:53:56,771 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118625.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:54:06,602 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118631.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:54:26,648 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 4100, loss[loss=0.1531, simple_loss=0.2317, pruned_loss=0.03725, over 4737.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2454, pruned_loss=0.05081, over 956869.96 frames. ], batch size: 59, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:54:38,505 INFO [optim.py:369] (2/7) 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 15:54:40,439 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2142, 1.7755, 1.7900, 1.9123, 2.3070, 1.9462, 1.7212, 1.7262], + device='cuda:2'), covar=tensor([0.1564, 0.1506, 0.2066, 0.1394, 0.0922, 0.1750, 0.2017, 0.2360], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0310, 0.0349, 0.0287, 0.0324, 0.0306, 0.0299, 0.0367], + device='cuda:2'), out_proj_covar=tensor([6.3374e-05, 6.4156e-05, 7.3638e-05, 5.7974e-05, 6.6827e-05, 6.4148e-05, + 6.2656e-05, 7.7938e-05], device='cuda:2') +2023-04-27 15:54:54,352 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:54:58,631 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:55:00,959 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118696.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:55:06,698 INFO [finetune.py:976] (2/7) Epoch 21, batch 4150, loss[loss=0.1926, simple_loss=0.2772, pruned_loss=0.05407, over 4806.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2465, pruned_loss=0.05119, over 955166.50 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:55:07,872 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9762, 1.7915, 1.9492, 2.2392, 2.2827, 1.8313, 1.4724, 2.0960], + device='cuda:2'), covar=tensor([0.0803, 0.1040, 0.0695, 0.0599, 0.0566, 0.0817, 0.0835, 0.0543], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0202, 0.0184, 0.0174, 0.0178, 0.0181, 0.0152, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:55:56,016 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118744.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:56:07,490 INFO [finetune.py:976] (2/7) Epoch 21, batch 4200, loss[loss=0.1316, simple_loss=0.2129, pruned_loss=0.02518, over 4761.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2476, pruned_loss=0.05097, over 955901.39 frames. ], batch size: 26, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:56:19,454 INFO [optim.py:369] (2/7) 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 15:56:58,125 INFO [finetune.py:976] (2/7) Epoch 21, batch 4250, loss[loss=0.1534, simple_loss=0.2248, pruned_loss=0.04097, over 4811.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2456, pruned_loss=0.05031, over 956278.31 frames. ], batch size: 25, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:57:08,845 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-27 15:57:12,950 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 15:57:13,419 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118824.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:57:32,154 INFO [finetune.py:976] (2/7) Epoch 21, batch 4300, loss[loss=0.1522, simple_loss=0.2169, pruned_loss=0.04381, over 4911.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2435, pruned_loss=0.05014, over 957005.82 frames. ], batch size: 46, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:57:37,516 INFO [optim.py:369] (2/7) 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 15:57:44,035 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4780, 3.8487, 0.8223, 2.0027, 2.1524, 2.6501, 2.2438, 1.0095], + device='cuda:2'), covar=tensor([0.1504, 0.0997, 0.2160, 0.1307, 0.1146, 0.1033, 0.1482, 0.2122], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0137, 0.0120, 0.0133, 0.0152, 0.0116, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:57:44,623 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 4350, loss[loss=0.1693, simple_loss=0.2426, pruned_loss=0.04803, over 4918.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2402, pruned_loss=0.04889, over 957007.13 frames. ], batch size: 35, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:58:07,420 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0689, 1.8727, 2.0942, 2.3808, 2.3739, 1.9066, 1.5164, 2.1554], + device='cuda:2'), covar=tensor([0.0774, 0.1034, 0.0619, 0.0586, 0.0564, 0.0900, 0.0886, 0.0564], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0202, 0.0184, 0.0174, 0.0179, 0.0181, 0.0153, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 15:58:20,451 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118925.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:58:29,836 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9019, 2.5436, 1.8969, 1.9058, 1.3191, 1.3356, 2.0189, 1.3223], + device='cuda:2'), covar=tensor([0.1717, 0.1255, 0.1445, 0.1694, 0.2372, 0.1969, 0.1011, 0.2088], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0211, 0.0170, 0.0204, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 15:58:37,344 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 15:58:39,669 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:58:51,396 INFO [finetune.py:976] (2/7) Epoch 21, batch 4400, loss[loss=0.1992, simple_loss=0.2829, pruned_loss=0.05778, over 4817.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2408, pruned_loss=0.0493, over 956643.09 frames. ], batch size: 38, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 15:59:00,776 INFO [optim.py:369] (2/7) 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 15:59:14,268 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118973.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:24,168 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3393, 1.2927, 1.5642, 1.5686, 1.2687, 1.0921, 1.3723, 0.9302], + device='cuda:2'), covar=tensor([0.0551, 0.0493, 0.0398, 0.0476, 0.0604, 0.0952, 0.0466, 0.0562], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 15:59:35,128 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118985.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:36,314 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118987.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:45,374 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118993.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 15:59:52,143 INFO [finetune.py:976] (2/7) Epoch 21, batch 4450, loss[loss=0.2051, simple_loss=0.2869, pruned_loss=0.06161, over 4775.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2457, pruned_loss=0.05005, over 956479.79 frames. ], batch size: 54, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:00:11,985 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119033.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:00:20,307 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4613, 1.8438, 1.8753, 2.0335, 1.7685, 1.8910, 1.9313, 1.9202], + device='cuda:2'), covar=tensor([0.4567, 0.5503, 0.4380, 0.3874, 0.5699, 0.7475, 0.5320, 0.4645], + device='cuda:2'), in_proj_covar=tensor([0.0333, 0.0369, 0.0319, 0.0333, 0.0343, 0.0391, 0.0353, 0.0324], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:00:25,718 INFO [finetune.py:976] (2/7) Epoch 21, batch 4500, loss[loss=0.184, simple_loss=0.2596, pruned_loss=0.05418, over 4813.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2486, pruned_loss=0.05163, over 958408.04 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:00:30,592 INFO [optim.py:369] (2/7) 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 16:00:39,563 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119075.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:00:58,674 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 4550, loss[loss=0.1627, simple_loss=0.2252, pruned_loss=0.05006, over 4287.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2499, pruned_loss=0.05162, over 957336.90 frames. ], batch size: 18, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:01:29,165 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.05 vs. limit=5.0 +2023-04-27 16:01:59,522 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:02:16,873 INFO [finetune.py:976] (2/7) Epoch 21, batch 4600, loss[loss=0.2186, simple_loss=0.2729, pruned_loss=0.08212, over 4891.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2487, pruned_loss=0.05109, over 956777.91 frames. ], batch size: 32, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:02:16,985 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119154.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 16:02:21,814 INFO [optim.py:369] (2/7) 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 16:02:21,942 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2874, 2.1880, 1.8200, 1.7447, 2.2305, 1.8141, 2.5859, 1.5192], + device='cuda:2'), covar=tensor([0.3441, 0.1885, 0.4288, 0.3117, 0.1767, 0.2514, 0.1982, 0.4505], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0352, 0.0428, 0.0357, 0.0384, 0.0377, 0.0373, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:02:49,849 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1324, 1.5679, 1.9268, 2.1513, 1.9977, 1.5679, 1.1284, 1.6513], + device='cuda:2'), covar=tensor([0.3053, 0.3126, 0.1610, 0.2113, 0.2359, 0.2552, 0.4184, 0.2016], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0246, 0.0228, 0.0316, 0.0220, 0.0234, 0.0229, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 16:02:50,904 INFO [finetune.py:976] (2/7) Epoch 21, batch 4650, loss[loss=0.1608, simple_loss=0.2351, pruned_loss=0.04325, over 4746.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2465, pruned_loss=0.05094, over 955687.05 frames. ], batch size: 27, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:03:24,687 INFO [finetune.py:976] (2/7) Epoch 21, batch 4700, loss[loss=0.2042, simple_loss=0.2688, pruned_loss=0.06979, over 4828.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2436, pruned_loss=0.04993, over 956669.23 frames. ], batch size: 30, lr: 3.18e-03, grad_scale: 32.0 +2023-04-27 16:03:29,613 INFO [optim.py:369] (2/7) 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 16:03:46,000 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 4750, loss[loss=0.1577, simple_loss=0.2255, pruned_loss=0.04499, over 4763.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2416, pruned_loss=0.049, over 957994.65 frames. ], batch size: 27, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:04:39,497 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119335.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:04:54,473 INFO [finetune.py:976] (2/7) Epoch 21, batch 4800, loss[loss=0.188, simple_loss=0.2665, pruned_loss=0.05476, over 4903.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2432, pruned_loss=0.0498, over 957734.93 frames. ], batch size: 35, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:04:59,371 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 21, batch 4850, loss[loss=0.1633, simple_loss=0.2282, pruned_loss=0.04921, over 4803.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2451, pruned_loss=0.05003, over 956662.60 frames. ], batch size: 25, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:05:43,554 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119431.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:05:55,519 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119449.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:05:58,993 INFO [finetune.py:976] (2/7) Epoch 21, batch 4900, loss[loss=0.1602, simple_loss=0.2268, pruned_loss=0.0468, over 4754.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2472, pruned_loss=0.05068, over 957201.82 frames. ], batch size: 26, lr: 3.17e-03, grad_scale: 32.0 +2023-04-27 16:06:04,806 INFO [optim.py:369] (2/7) 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 16:06:25,484 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2811, 3.1088, 1.0134, 1.7919, 1.6989, 2.3755, 1.7082, 1.0865], + device='cuda:2'), covar=tensor([0.1454, 0.0907, 0.1826, 0.1174, 0.1148, 0.0909, 0.1508, 0.1945], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0119, 0.0132, 0.0151, 0.0115, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:06:35,841 INFO [finetune.py:976] (2/7) Epoch 21, batch 4950, loss[loss=0.1377, simple_loss=0.1809, pruned_loss=0.04724, over 4302.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2474, pruned_loss=0.05089, over 956590.78 frames. ], batch size: 18, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:06:49,614 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119514.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:07:20,080 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7233, 2.1287, 1.6762, 1.4711, 1.2456, 1.2751, 1.7223, 1.2407], + device='cuda:2'), covar=tensor([0.1667, 0.1237, 0.1396, 0.1697, 0.2427, 0.1893, 0.1011, 0.2024], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0211, 0.0168, 0.0203, 0.0199, 0.0185, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 16:07:42,892 INFO [finetune.py:976] (2/7) Epoch 21, batch 5000, loss[loss=0.1595, simple_loss=0.23, pruned_loss=0.04449, over 4807.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2457, pruned_loss=0.05023, over 956315.49 frames. ], batch size: 25, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:07:56,236 INFO [optim.py:369] (2/7) 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 16:08:02,509 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.64 vs. limit=5.0 +2023-04-27 16:08:15,219 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119575.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:08:44,470 INFO [finetune.py:976] (2/7) Epoch 21, batch 5050, loss[loss=0.1913, simple_loss=0.2441, pruned_loss=0.06919, over 4839.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2429, pruned_loss=0.04931, over 957448.15 frames. ], batch size: 30, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:08:56,817 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6265, 2.6155, 1.9544, 2.2188, 2.6052, 2.0964, 3.2754, 1.8404], + device='cuda:2'), covar=tensor([0.3588, 0.2224, 0.4141, 0.3432, 0.1731, 0.2834, 0.1456, 0.4073], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0347, 0.0421, 0.0351, 0.0379, 0.0373, 0.0366, 0.0416], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:09:03,017 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 16:09:28,917 INFO [finetune.py:976] (2/7) Epoch 21, batch 5100, loss[loss=0.1223, simple_loss=0.1949, pruned_loss=0.02486, over 4824.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2399, pruned_loss=0.0486, over 957838.14 frames. ], batch size: 30, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:09:41,676 INFO [optim.py:369] (2/7) 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 16:09:43,054 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119665.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:10:16,801 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 16:10:22,195 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-04-27 16:10:31,241 INFO [finetune.py:976] (2/7) Epoch 21, batch 5150, loss[loss=0.1742, simple_loss=0.2369, pruned_loss=0.05578, over 4772.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2397, pruned_loss=0.04864, over 957004.51 frames. ], batch size: 26, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:10:45,709 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8502, 3.7253, 2.7940, 4.4448, 3.7890, 3.8374, 1.4848, 3.8288], + device='cuda:2'), covar=tensor([0.1898, 0.1203, 0.3531, 0.1529, 0.4120, 0.1823, 0.6376, 0.2658], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0215, 0.0251, 0.0304, 0.0295, 0.0245, 0.0275, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:11:05,042 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119726.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:11:08,012 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119731.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:11:30,492 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 5200, loss[loss=0.2065, simple_loss=0.2802, pruned_loss=0.06637, over 4719.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2434, pruned_loss=0.04995, over 956247.33 frames. ], batch size: 59, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:11:39,368 INFO [optim.py:369] (2/7) 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 16:11:51,106 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:12:06,569 INFO [finetune.py:976] (2/7) Epoch 21, batch 5250, loss[loss=0.1656, simple_loss=0.2486, pruned_loss=0.04131, over 4146.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2454, pruned_loss=0.05, over 956893.04 frames. ], batch size: 65, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:12:48,888 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1808, 1.4354, 1.7004, 1.7744, 1.6719, 1.7699, 1.7431, 1.7163], + device='cuda:2'), covar=tensor([0.4253, 0.5135, 0.4227, 0.4198, 0.5155, 0.7123, 0.4956, 0.4430], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0373, 0.0323, 0.0337, 0.0347, 0.0395, 0.0357, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:12:51,200 INFO [finetune.py:976] (2/7) Epoch 21, batch 5300, loss[loss=0.1836, simple_loss=0.2557, pruned_loss=0.05577, over 4866.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2479, pruned_loss=0.051, over 957910.92 frames. ], batch size: 34, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:13:02,329 INFO [optim.py:369] (2/7) 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 16:13:04,782 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6484, 1.7539, 0.9854, 1.3113, 1.8191, 1.4923, 1.4293, 1.4900], + device='cuda:2'), covar=tensor([0.0506, 0.0367, 0.0313, 0.0553, 0.0270, 0.0498, 0.0491, 0.0589], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 16:13:13,803 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119870.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:13:34,784 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7608, 1.0075, 1.7249, 2.1683, 1.8141, 1.6689, 1.7154, 1.7078], + device='cuda:2'), covar=tensor([0.4615, 0.6877, 0.6381, 0.6182, 0.6665, 0.8018, 0.7877, 0.8737], + device='cuda:2'), in_proj_covar=tensor([0.0428, 0.0411, 0.0502, 0.0502, 0.0456, 0.0485, 0.0493, 0.0500], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:13:38,550 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 16:13:53,929 INFO [finetune.py:976] (2/7) Epoch 21, batch 5350, loss[loss=0.1686, simple_loss=0.242, pruned_loss=0.04761, over 4921.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2471, pruned_loss=0.05081, over 956424.80 frames. ], batch size: 38, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:14:43,494 INFO [finetune.py:976] (2/7) Epoch 21, batch 5400, loss[loss=0.126, simple_loss=0.2004, pruned_loss=0.02584, over 4745.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2448, pruned_loss=0.05028, over 956899.31 frames. ], batch size: 23, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:14:48,954 INFO [optim.py:369] (2/7) 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 16:14:55,599 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3796, 3.3004, 0.9258, 1.8488, 1.7656, 2.4393, 1.9446, 1.0195], + device='cuda:2'), covar=tensor([0.1360, 0.0856, 0.1936, 0.1136, 0.1076, 0.0949, 0.1294, 0.2063], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0239, 0.0136, 0.0119, 0.0132, 0.0151, 0.0115, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:15:18,156 INFO [finetune.py:976] (2/7) Epoch 21, batch 5450, loss[loss=0.1477, simple_loss=0.2167, pruned_loss=0.03932, over 4813.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.242, pruned_loss=0.04964, over 957215.96 frames. ], batch size: 25, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:15:34,283 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=120021.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:15:48,345 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-04-27 16:15:49,992 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5757, 1.2714, 0.4319, 1.3244, 1.2780, 1.4449, 1.4051, 1.3669], + device='cuda:2'), covar=tensor([0.0498, 0.0399, 0.0426, 0.0583, 0.0288, 0.0506, 0.0504, 0.0576], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0029, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 16:16:03,188 INFO [finetune.py:976] (2/7) Epoch 21, batch 5500, loss[loss=0.1733, simple_loss=0.2385, pruned_loss=0.05402, over 4825.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2382, pruned_loss=0.04801, over 958081.10 frames. ], batch size: 30, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:16:14,181 INFO [optim.py:369] (2/7) 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 16:16:34,227 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1390, 4.4647, 1.2521, 2.4301, 2.5614, 3.2088, 2.5914, 0.9736], + device='cuda:2'), covar=tensor([0.1341, 0.1183, 0.1787, 0.1242, 0.1013, 0.0988, 0.1589, 0.2119], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0119, 0.0133, 0.0152, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:17:07,731 INFO [finetune.py:976] (2/7) Epoch 21, batch 5550, loss[loss=0.2246, simple_loss=0.2907, pruned_loss=0.07929, over 4748.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2405, pruned_loss=0.04911, over 958276.42 frames. ], batch size: 59, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:18:05,777 INFO [finetune.py:976] (2/7) Epoch 21, batch 5600, loss[loss=0.1654, simple_loss=0.24, pruned_loss=0.04538, over 4864.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2448, pruned_loss=0.05009, over 957979.66 frames. ], batch size: 44, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:18:11,018 INFO [optim.py:369] (2/7) 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 16:18:14,826 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-04-27 16:18:15,208 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 21, batch 5650, loss[loss=0.156, simple_loss=0.2225, pruned_loss=0.04477, over 4537.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2466, pruned_loss=0.04995, over 957359.26 frames. ], batch size: 19, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:18:45,336 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=120218.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:19:30,424 INFO [finetune.py:976] (2/7) Epoch 21, batch 5700, loss[loss=0.1528, simple_loss=0.2128, pruned_loss=0.04638, over 3531.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2427, pruned_loss=0.04981, over 936733.37 frames. ], batch size: 15, lr: 3.17e-03, grad_scale: 16.0 +2023-04-27 16:19:41,631 INFO [optim.py:369] (2/7) 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 16:20:17,302 INFO [finetune.py:976] (2/7) Epoch 22, batch 0, loss[loss=0.1876, simple_loss=0.2528, pruned_loss=0.06114, over 4818.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2528, pruned_loss=0.06114, over 4818.00 frames. ], batch size: 30, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:20:17,302 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 16:20:33,672 INFO [finetune.py:1010] (2/7) Epoch 22, validation: loss=0.1546, simple_loss=0.2251, pruned_loss=0.04204, over 2265189.00 frames. +2023-04-27 16:20:33,673 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 16:20:37,879 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1906, 1.8130, 2.0361, 2.5642, 2.5391, 2.0873, 1.8236, 2.3425], + device='cuda:2'), covar=tensor([0.0852, 0.1209, 0.0761, 0.0597, 0.0631, 0.0912, 0.0797, 0.0567], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0201, 0.0183, 0.0174, 0.0177, 0.0180, 0.0152, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:20:59,176 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=120321.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:21:04,988 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-27 16:21:06,358 INFO [finetune.py:976] (2/7) Epoch 22, batch 50, loss[loss=0.1565, simple_loss=0.2365, pruned_loss=0.03825, over 4770.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2513, pruned_loss=0.05447, over 216875.59 frames. ], batch size: 28, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:21:21,927 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1080, 1.8048, 2.2539, 2.5462, 2.1441, 2.0429, 2.1509, 2.1416], + device='cuda:2'), covar=tensor([0.4819, 0.7439, 0.7279, 0.5829, 0.6245, 0.9041, 0.8819, 0.9087], + device='cuda:2'), in_proj_covar=tensor([0.0429, 0.0412, 0.0505, 0.0504, 0.0457, 0.0487, 0.0495, 0.0502], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:21:23,747 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4613, 1.6399, 1.8415, 1.9483, 1.7347, 1.9287, 1.9190, 1.8967], + device='cuda:2'), covar=tensor([0.4212, 0.5867, 0.4583, 0.4470, 0.5836, 0.7322, 0.5231, 0.4840], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0375, 0.0323, 0.0338, 0.0348, 0.0395, 0.0359, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:21:24,325 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5203, 3.3548, 1.0121, 1.9217, 1.8050, 2.4755, 2.0116, 0.9727], + device='cuda:2'), covar=tensor([0.1376, 0.0996, 0.1883, 0.1239, 0.1148, 0.1004, 0.1374, 0.2026], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0238, 0.0136, 0.0118, 0.0132, 0.0151, 0.0115, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:21:27,311 INFO [optim.py:369] (2/7) 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 16:21:31,107 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 100, loss[loss=0.1726, simple_loss=0.2476, pruned_loss=0.04876, over 4815.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2426, pruned_loss=0.04995, over 382223.38 frames. ], batch size: 41, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:22:48,987 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 16:22:51,079 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8862, 1.9995, 1.0043, 1.5795, 2.0665, 1.7181, 1.6474, 1.7898], + device='cuda:2'), covar=tensor([0.0469, 0.0355, 0.0297, 0.0530, 0.0241, 0.0480, 0.0467, 0.0534], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 16:22:55,500 INFO [finetune.py:976] (2/7) Epoch 22, batch 150, loss[loss=0.154, simple_loss=0.2274, pruned_loss=0.04031, over 4829.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2372, pruned_loss=0.04894, over 511554.72 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:23:25,574 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 16:23:29,005 INFO [optim.py:369] (2/7) 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 16:23:58,242 INFO [finetune.py:976] (2/7) Epoch 22, batch 200, loss[loss=0.1574, simple_loss=0.2307, pruned_loss=0.04208, over 4860.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2358, pruned_loss=0.04782, over 608195.17 frames. ], batch size: 44, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:25:06,005 INFO [finetune.py:976] (2/7) Epoch 22, batch 250, loss[loss=0.1733, simple_loss=0.2466, pruned_loss=0.05004, over 4814.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2375, pruned_loss=0.04802, over 683759.91 frames. ], batch size: 39, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:25:37,607 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 16:25:49,441 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 22, batch 300, loss[loss=0.1799, simple_loss=0.264, pruned_loss=0.04791, over 4806.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2431, pruned_loss=0.04986, over 742406.34 frames. ], batch size: 45, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:27:06,864 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7613, 1.7739, 1.6508, 1.2752, 1.8087, 1.5190, 2.1950, 1.4710], + device='cuda:2'), covar=tensor([0.3216, 0.1686, 0.4266, 0.2530, 0.1326, 0.1938, 0.1388, 0.4138], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0351, 0.0427, 0.0354, 0.0382, 0.0374, 0.0369, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:27:19,464 INFO [finetune.py:976] (2/7) Epoch 22, batch 350, loss[loss=0.1981, simple_loss=0.2708, pruned_loss=0.06267, over 4789.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2459, pruned_loss=0.0512, over 789371.39 frames. ], batch size: 45, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:28:01,875 INFO [optim.py:369] (2/7) 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 16:28:20,622 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 400, loss[loss=0.1669, simple_loss=0.2488, pruned_loss=0.04247, over 4901.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.248, pruned_loss=0.05162, over 826692.07 frames. ], batch size: 35, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:28:44,122 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9009, 2.4330, 1.9723, 1.8013, 1.3565, 1.4407, 1.9918, 1.3680], + device='cuda:2'), covar=tensor([0.1625, 0.1449, 0.1428, 0.1695, 0.2306, 0.1959, 0.0966, 0.2021], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0209, 0.0168, 0.0202, 0.0198, 0.0184, 0.0154, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 16:29:32,132 INFO [finetune.py:976] (2/7) Epoch 22, batch 450, loss[loss=0.1946, simple_loss=0.2626, pruned_loss=0.06325, over 4902.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2469, pruned_loss=0.05156, over 854768.92 frames. ], batch size: 36, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:29:40,493 INFO [zipformer.py:1188] (2/7) 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] (2/7) attn_weights_entropy = tensor([2.1113, 1.8851, 2.1824, 2.4716, 2.4657, 2.0769, 1.6982, 2.2245], + device='cuda:2'), covar=tensor([0.0817, 0.1082, 0.0646, 0.0595, 0.0571, 0.0872, 0.0778, 0.0580], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0198, 0.0181, 0.0172, 0.0175, 0.0177, 0.0149, 0.0175], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:30:03,215 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5156, 1.7791, 1.9383, 2.0265, 1.8562, 1.9324, 1.9903, 1.9687], + device='cuda:2'), covar=tensor([0.4088, 0.5139, 0.4116, 0.4294, 0.5486, 0.7009, 0.4970, 0.4473], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0372, 0.0323, 0.0337, 0.0347, 0.0394, 0.0356, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:30:03,661 INFO [optim.py:369] (2/7) 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 16:30:15,272 INFO [finetune.py:976] (2/7) Epoch 22, batch 500, loss[loss=0.1261, simple_loss=0.2, pruned_loss=0.02614, over 4765.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2447, pruned_loss=0.05092, over 878269.50 frames. ], batch size: 28, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:30:22,584 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1607, 1.5683, 1.4896, 2.0472, 2.2146, 1.8631, 1.7481, 1.6321], + device='cuda:2'), covar=tensor([0.1905, 0.1991, 0.1910, 0.1694, 0.1242, 0.2468, 0.2437, 0.2315], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0310, 0.0349, 0.0287, 0.0324, 0.0308, 0.0300, 0.0370], + device='cuda:2'), 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:2') +2023-04-27 16:30:28,258 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-04-27 16:30:49,310 INFO [finetune.py:976] (2/7) Epoch 22, batch 550, loss[loss=0.1717, simple_loss=0.2538, pruned_loss=0.04479, over 4814.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2422, pruned_loss=0.05032, over 894019.31 frames. ], batch size: 40, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:30:59,025 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7905, 1.3746, 4.6340, 4.3722, 3.9229, 4.3591, 4.2020, 4.0031], + device='cuda:2'), covar=tensor([0.6827, 0.5923, 0.0924, 0.1400, 0.1011, 0.1812, 0.1143, 0.1511], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0307, 0.0409, 0.0408, 0.0350, 0.0412, 0.0314, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:31:15,724 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 22, batch 600, loss[loss=0.1566, simple_loss=0.2253, pruned_loss=0.04391, over 4704.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2411, pruned_loss=0.05025, over 906913.04 frames. ], batch size: 23, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:32:45,426 INFO [finetune.py:976] (2/7) Epoch 22, batch 650, loss[loss=0.1947, simple_loss=0.2685, pruned_loss=0.06043, over 4839.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2427, pruned_loss=0.04984, over 916908.47 frames. ], batch size: 47, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:33:26,711 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 22, batch 700, loss[loss=0.167, simple_loss=0.2516, pruned_loss=0.04119, over 4916.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.245, pruned_loss=0.05065, over 923568.98 frames. ], batch size: 37, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:34:21,029 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121003.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:34:33,518 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7176, 1.6468, 2.1513, 2.2412, 1.5963, 1.4676, 1.8063, 0.9347], + device='cuda:2'), covar=tensor([0.0640, 0.0655, 0.0416, 0.0711, 0.0756, 0.1162, 0.0626, 0.0757], + device='cuda:2'), in_proj_covar=tensor([0.0068, 0.0067, 0.0066, 0.0067, 0.0074, 0.0096, 0.0072, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:34:45,578 INFO [finetune.py:976] (2/7) Epoch 22, batch 750, loss[loss=0.1495, simple_loss=0.2295, pruned_loss=0.03475, over 4836.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2462, pruned_loss=0.05112, over 931476.62 frames. ], batch size: 49, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:34:45,650 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:34:48,766 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7665, 3.7321, 2.8736, 4.3779, 3.7473, 3.8155, 1.5200, 3.7503], + device='cuda:2'), covar=tensor([0.1679, 0.1200, 0.3271, 0.1437, 0.2960, 0.1627, 0.5757, 0.2359], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0214, 0.0249, 0.0302, 0.0293, 0.0244, 0.0272, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:34:54,309 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-27 16:35:04,850 INFO [optim.py:369] (2/7) 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 16:35:05,575 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121064.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:35:19,165 INFO [finetune.py:976] (2/7) Epoch 22, batch 800, loss[loss=0.1702, simple_loss=0.2384, pruned_loss=0.05103, over 4883.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2445, pruned_loss=0.0504, over 936676.92 frames. ], batch size: 32, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:35:39,996 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-04-27 16:35:52,464 INFO [finetune.py:976] (2/7) Epoch 22, batch 850, loss[loss=0.2387, simple_loss=0.2772, pruned_loss=0.1001, over 4824.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2428, pruned_loss=0.04974, over 939954.23 frames. ], batch size: 38, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:35:59,261 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9156, 2.5524, 1.9547, 2.0245, 1.5475, 1.5243, 2.0689, 1.4793], + device='cuda:2'), covar=tensor([0.1340, 0.1325, 0.1271, 0.1471, 0.1946, 0.1584, 0.0798, 0.1736], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0208, 0.0167, 0.0201, 0.0197, 0.0183, 0.0154, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 16:36:06,431 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3943, 3.3256, 0.9222, 1.8216, 1.7960, 2.4394, 1.9253, 1.0246], + device='cuda:2'), covar=tensor([0.1529, 0.1351, 0.2282, 0.1410, 0.1226, 0.1198, 0.1558, 0.2148], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0119, 0.0132, 0.0153, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:36:11,705 INFO [optim.py:369] (2/7) 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 16:36:12,458 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121164.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:36:25,801 INFO [finetune.py:976] (2/7) Epoch 22, batch 900, loss[loss=0.1763, simple_loss=0.2445, pruned_loss=0.05409, over 4869.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2416, pruned_loss=0.04947, over 943944.82 frames. ], batch size: 34, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:36:52,981 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121225.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:36:53,008 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8319, 1.4001, 1.8792, 2.2511, 1.9081, 1.7414, 1.8116, 1.7554], + device='cuda:2'), covar=tensor([0.4577, 0.7100, 0.6292, 0.5537, 0.5823, 0.7816, 0.8081, 0.9642], + device='cuda:2'), in_proj_covar=tensor([0.0428, 0.0412, 0.0503, 0.0504, 0.0458, 0.0487, 0.0494, 0.0501], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:36:59,093 INFO [finetune.py:976] (2/7) Epoch 22, batch 950, loss[loss=0.1697, simple_loss=0.2478, pruned_loss=0.04577, over 4900.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2401, pruned_loss=0.04905, over 945819.80 frames. ], batch size: 35, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:37:01,045 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6868, 1.4214, 1.3684, 1.4104, 1.8579, 1.5374, 1.2291, 1.3145], + device='cuda:2'), covar=tensor([0.1524, 0.1246, 0.2041, 0.1435, 0.0751, 0.1405, 0.1928, 0.2353], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0313, 0.0352, 0.0289, 0.0326, 0.0311, 0.0303, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.4514e-05, 6.4714e-05, 7.4352e-05, 5.8078e-05, 6.7245e-05, 6.5096e-05, + 6.3454e-05, 7.9554e-05], device='cuda:2') +2023-04-27 16:37:35,643 INFO [optim.py:369] (2/7) 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 16:38:01,692 INFO [finetune.py:976] (2/7) Epoch 22, batch 1000, loss[loss=0.1865, simple_loss=0.2618, pruned_loss=0.05559, over 4211.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2432, pruned_loss=0.05037, over 945006.57 frames. ], batch size: 65, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:38:30,405 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7461, 2.2099, 1.9565, 1.6953, 1.3177, 1.3537, 2.0506, 1.2652], + device='cuda:2'), covar=tensor([0.1793, 0.1592, 0.1401, 0.1782, 0.2388, 0.2049, 0.0863, 0.2120], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0209, 0.0168, 0.0202, 0.0198, 0.0184, 0.0154, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 16:38:41,544 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121311.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 16:39:01,159 INFO [finetune.py:976] (2/7) Epoch 22, batch 1050, loss[loss=0.171, simple_loss=0.2498, pruned_loss=0.04606, over 4871.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2449, pruned_loss=0.05016, over 949215.06 frames. ], batch size: 31, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:39:01,251 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121332.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:39:18,517 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121359.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:39:20,862 INFO [optim.py:369] (2/7) 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 16:39:26,948 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121372.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 16:39:31,738 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:39:33,387 INFO [finetune.py:976] (2/7) Epoch 22, batch 1100, loss[loss=0.1763, simple_loss=0.2477, pruned_loss=0.05243, over 4922.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2463, pruned_loss=0.05104, over 948944.52 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:39:41,064 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 16:39:50,907 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 1150, loss[loss=0.1697, simple_loss=0.2565, pruned_loss=0.0415, over 4922.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2468, pruned_loss=0.05092, over 950877.90 frames. ], batch size: 42, lr: 3.16e-03, grad_scale: 16.0 +2023-04-27 16:40:19,826 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9920, 1.7746, 2.0030, 2.3100, 2.4231, 1.9155, 1.7014, 2.0442], + device='cuda:2'), covar=tensor([0.0775, 0.0993, 0.0634, 0.0559, 0.0557, 0.0849, 0.0788, 0.0550], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0199, 0.0182, 0.0172, 0.0176, 0.0178, 0.0151, 0.0175], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:40:27,723 INFO [optim.py:369] (2/7) 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 16:40:30,880 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 1200, loss[loss=0.1449, simple_loss=0.2154, pruned_loss=0.03716, over 4904.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2442, pruned_loss=0.0495, over 951360.52 frames. ], batch size: 35, lr: 3.15e-03, grad_scale: 16.0 +2023-04-27 16:41:05,075 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:41:12,819 INFO [finetune.py:976] (2/7) Epoch 22, batch 1250, loss[loss=0.1737, simple_loss=0.2331, pruned_loss=0.05717, over 4749.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2425, pruned_loss=0.04937, over 952906.90 frames. ], batch size: 28, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:41:34,045 INFO [optim.py:369] (2/7) 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 16:41:46,168 INFO [finetune.py:976] (2/7) Epoch 22, batch 1300, loss[loss=0.1489, simple_loss=0.2186, pruned_loss=0.03955, over 4790.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2411, pruned_loss=0.04952, over 954644.63 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:41:52,328 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 16:42:19,072 INFO [finetune.py:976] (2/7) Epoch 22, batch 1350, loss[loss=0.2061, simple_loss=0.2701, pruned_loss=0.07104, over 4901.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2407, pruned_loss=0.04943, over 954969.17 frames. ], batch size: 43, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:42:32,455 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7769, 3.6986, 2.8723, 4.3371, 3.6894, 3.7165, 1.6449, 3.7480], + device='cuda:2'), covar=tensor([0.1650, 0.1349, 0.3260, 0.1361, 0.3079, 0.1617, 0.5431, 0.2053], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0214, 0.0249, 0.0303, 0.0292, 0.0244, 0.0272, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:42:55,952 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121659.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:42:58,249 INFO [optim.py:369] (2/7) 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 16:42:58,611 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-27 16:43:04,731 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121667.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 16:43:19,351 INFO [finetune.py:976] (2/7) Epoch 22, batch 1400, loss[loss=0.2216, simple_loss=0.298, pruned_loss=0.07253, over 4922.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2444, pruned_loss=0.05056, over 955161.33 frames. ], batch size: 42, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:43:37,357 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1327, 1.3649, 1.3548, 1.7399, 1.4883, 1.7222, 1.3150, 3.1175], + device='cuda:2'), covar=tensor([0.0677, 0.0951, 0.0908, 0.1252, 0.0759, 0.0512, 0.0882, 0.0195], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 16:43:59,847 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121707.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:44:03,648 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3124, 1.9250, 2.3253, 2.6405, 2.3752, 2.2526, 2.3076, 2.2181], + device='cuda:2'), covar=tensor([0.3856, 0.5252, 0.5338, 0.4810, 0.4565, 0.6569, 0.6571, 0.7183], + device='cuda:2'), in_proj_covar=tensor([0.0430, 0.0413, 0.0505, 0.0505, 0.0457, 0.0489, 0.0494, 0.0502], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:44:24,392 INFO [finetune.py:976] (2/7) Epoch 22, batch 1450, loss[loss=0.1401, simple_loss=0.2111, pruned_loss=0.03454, over 4776.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2468, pruned_loss=0.0513, over 954226.74 frames. ], batch size: 26, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:44:44,669 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1630, 2.1532, 1.7513, 1.7501, 2.1962, 1.7678, 2.6242, 1.5466], + device='cuda:2'), covar=tensor([0.3391, 0.1748, 0.4361, 0.2816, 0.1416, 0.2307, 0.1289, 0.4288], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0348, 0.0423, 0.0350, 0.0379, 0.0373, 0.0365, 0.0417], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:45:04,295 INFO [optim.py:369] (2/7) 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 16:45:04,383 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121763.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:45:22,128 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5220, 1.3100, 3.8476, 3.2931, 3.3932, 3.5840, 3.5311, 3.1814], + device='cuda:2'), covar=tensor([0.9553, 0.8004, 0.1782, 0.3154, 0.2204, 0.3083, 0.3790, 0.3479], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0306, 0.0405, 0.0405, 0.0346, 0.0408, 0.0310, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:45:27,551 INFO [finetune.py:976] (2/7) Epoch 22, batch 1500, loss[loss=0.2336, simple_loss=0.2996, pruned_loss=0.08379, over 4739.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2485, pruned_loss=0.05184, over 954601.22 frames. ], batch size: 59, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:45:43,057 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9402, 2.5415, 1.9554, 2.0549, 1.5445, 1.4668, 2.0331, 1.5055], + device='cuda:2'), covar=tensor([0.1243, 0.1180, 0.1165, 0.1346, 0.1829, 0.1640, 0.0806, 0.1637], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0211, 0.0169, 0.0203, 0.0200, 0.0185, 0.0155, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 16:45:58,963 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:46:06,250 INFO [finetune.py:976] (2/7) Epoch 22, batch 1550, loss[loss=0.1567, simple_loss=0.2329, pruned_loss=0.04022, over 4794.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.249, pruned_loss=0.0523, over 954000.08 frames. ], batch size: 25, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:46:10,078 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-04-27 16:46:28,046 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) Epoch 22, batch 1600, loss[loss=0.1424, simple_loss=0.2134, pruned_loss=0.03565, over 4769.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2466, pruned_loss=0.05167, over 955006.41 frames. ], batch size: 26, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:47:13,762 INFO [finetune.py:976] (2/7) Epoch 22, batch 1650, loss[loss=0.1617, simple_loss=0.2297, pruned_loss=0.04685, over 4898.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2438, pruned_loss=0.05057, over 955973.35 frames. ], batch size: 46, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:47:21,345 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 16:47:24,847 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121950.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:47:33,997 INFO [optim.py:369] (2/7) 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 16:47:38,055 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121967.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 16:47:47,186 INFO [finetune.py:976] (2/7) Epoch 22, batch 1700, loss[loss=0.1346, simple_loss=0.2148, pruned_loss=0.02725, over 4749.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2402, pruned_loss=0.04918, over 956081.16 frames. ], batch size: 59, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:47:47,330 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3582, 1.8030, 2.2202, 2.6004, 2.2022, 1.7772, 1.4521, 2.0108], + device='cuda:2'), covar=tensor([0.3129, 0.3106, 0.1742, 0.2357, 0.2626, 0.2654, 0.4096, 0.1818], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0243, 0.0226, 0.0313, 0.0220, 0.0232, 0.0227, 0.0183], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 16:48:06,896 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122015.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 16:48:37,958 INFO [finetune.py:976] (2/7) Epoch 22, batch 1750, loss[loss=0.1377, simple_loss=0.2126, pruned_loss=0.03143, over 4784.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2415, pruned_loss=0.04928, over 958061.20 frames. ], batch size: 28, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:48:52,471 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3025, 3.1567, 0.9431, 1.6082, 1.6300, 2.2750, 1.7888, 0.9462], + device='cuda:2'), covar=tensor([0.1399, 0.0962, 0.1926, 0.1337, 0.1222, 0.0992, 0.1379, 0.2021], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0138, 0.0120, 0.0133, 0.0152, 0.0116, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:49:11,592 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9513, 1.1934, 1.0990, 1.5467, 1.2787, 1.4582, 1.1818, 2.4811], + device='cuda:2'), covar=tensor([0.0756, 0.1157, 0.1119, 0.1409, 0.0879, 0.0666, 0.1064, 0.0316], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 16:49:13,876 INFO [optim.py:369] (2/7) 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 16:49:13,979 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122063.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:49:14,068 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.11 vs. limit=5.0 +2023-04-27 16:49:43,794 INFO [finetune.py:976] (2/7) Epoch 22, batch 1800, loss[loss=0.1924, simple_loss=0.2715, pruned_loss=0.05661, over 4884.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2457, pruned_loss=0.05018, over 959917.85 frames. ], batch size: 32, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:50:18,107 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122111.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:50:27,680 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122117.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:50:48,298 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.03 vs. limit=2.0 +2023-04-27 16:50:49,867 INFO [finetune.py:976] (2/7) Epoch 22, batch 1850, loss[loss=0.248, simple_loss=0.3075, pruned_loss=0.09418, over 4905.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2477, pruned_loss=0.05059, over 959189.58 frames. ], batch size: 37, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:51:14,573 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 16:51:22,419 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3990, 1.6318, 1.5007, 1.8421, 1.7607, 2.0432, 1.4441, 3.8855], + device='cuda:2'), covar=tensor([0.0595, 0.0857, 0.0851, 0.1215, 0.0660, 0.0503, 0.0758, 0.0153], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 16:51:27,102 INFO [optim.py:369] (2/7) 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 16:51:48,061 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122178.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 16:51:56,333 INFO [finetune.py:976] (2/7) Epoch 22, batch 1900, loss[loss=0.1461, simple_loss=0.2151, pruned_loss=0.03856, over 4769.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2479, pruned_loss=0.05069, over 958207.46 frames. ], batch size: 26, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:52:10,113 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122195.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:52:20,603 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1243, 2.6432, 2.2225, 2.6316, 1.9227, 2.3550, 2.5647, 1.8209], + device='cuda:2'), covar=tensor([0.2069, 0.1166, 0.0799, 0.1081, 0.3108, 0.1063, 0.1747, 0.2528], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0306, 0.0220, 0.0281, 0.0319, 0.0260, 0.0253, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1593e-04, 1.2117e-04, 8.6970e-05, 1.1113e-04, 1.2943e-04, 1.0274e-04, + 1.0187e-04, 1.0564e-04], device='cuda:2') +2023-04-27 16:52:27,898 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-27 16:52:40,533 INFO [finetune.py:976] (2/7) Epoch 22, batch 1950, loss[loss=0.161, simple_loss=0.2414, pruned_loss=0.04026, over 4773.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2468, pruned_loss=0.04962, over 958250.42 frames. ], batch size: 28, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:52:49,322 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 16:52:55,258 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122256.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:52:58,862 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5729, 3.5446, 0.9915, 1.7496, 1.9565, 2.4549, 1.9709, 1.0692], + device='cuda:2'), covar=tensor([0.1403, 0.1251, 0.2087, 0.1424, 0.1023, 0.1130, 0.1523, 0.1803], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0239, 0.0137, 0.0119, 0.0132, 0.0152, 0.0115, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 16:52:59,360 INFO [optim.py:369] (2/7) 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 16:53:13,159 INFO [finetune.py:976] (2/7) Epoch 22, batch 2000, loss[loss=0.1649, simple_loss=0.2403, pruned_loss=0.04479, over 4795.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2445, pruned_loss=0.04905, over 957280.36 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:53:29,025 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122306.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:53:37,027 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:53:46,289 INFO [finetune.py:976] (2/7) Epoch 22, batch 2050, loss[loss=0.1362, simple_loss=0.2116, pruned_loss=0.03042, over 4912.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2417, pruned_loss=0.04866, over 959147.07 frames. ], batch size: 36, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:54:00,748 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7688, 1.3257, 1.8961, 2.2432, 1.8960, 1.7299, 1.8342, 1.7703], + device='cuda:2'), covar=tensor([0.4273, 0.6903, 0.6076, 0.5082, 0.5476, 0.7845, 0.7509, 0.8627], + device='cuda:2'), in_proj_covar=tensor([0.0432, 0.0414, 0.0509, 0.0507, 0.0460, 0.0491, 0.0496, 0.0505], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:54:12,097 INFO [optim.py:369] (2/7) 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 16:54:35,072 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:54:36,183 INFO [finetune.py:976] (2/7) Epoch 22, batch 2100, loss[loss=0.1294, simple_loss=0.2039, pruned_loss=0.02744, over 4870.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2407, pruned_loss=0.04854, over 957702.48 frames. ], batch size: 34, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:55:40,952 INFO [finetune.py:976] (2/7) Epoch 22, batch 2150, loss[loss=0.1987, simple_loss=0.2624, pruned_loss=0.0675, over 4862.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2435, pruned_loss=0.05006, over 956963.01 frames. ], batch size: 31, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:56:18,178 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5825, 3.6219, 2.6569, 4.1712, 3.6335, 3.5638, 1.5088, 3.5726], + device='cuda:2'), covar=tensor([0.1906, 0.1313, 0.3642, 0.1934, 0.3754, 0.1848, 0.6190, 0.2641], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0215, 0.0250, 0.0304, 0.0293, 0.0245, 0.0274, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:56:19,327 INFO [optim.py:369] (2/7) 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 16:56:22,562 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122468.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:56:30,956 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122473.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 16:56:42,041 INFO [finetune.py:976] (2/7) Epoch 22, batch 2200, loss[loss=0.1989, simple_loss=0.2658, pruned_loss=0.06603, over 4862.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2436, pruned_loss=0.04989, over 953477.38 frames. ], batch size: 31, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:57:15,874 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2665, 1.7214, 2.1745, 2.4631, 2.1224, 1.7082, 1.3804, 1.8246], + device='cuda:2'), covar=tensor([0.3082, 0.3064, 0.1617, 0.2208, 0.2584, 0.2622, 0.4054, 0.2075], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0247, 0.0229, 0.0316, 0.0222, 0.0235, 0.0230, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 16:57:37,002 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8672, 2.1324, 2.0757, 2.2063, 1.8814, 2.1985, 2.2217, 2.0276], + device='cuda:2'), covar=tensor([0.4146, 0.6221, 0.5384, 0.4585, 0.6031, 0.7335, 0.6351, 0.6225], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0340, 0.0349, 0.0396, 0.0357, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:57:38,804 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7634, 2.3310, 1.9720, 1.7593, 1.3886, 1.3460, 2.0557, 1.3986], + device='cuda:2'), covar=tensor([0.1651, 0.1364, 0.1315, 0.1667, 0.2207, 0.1889, 0.0903, 0.1929], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0211, 0.0170, 0.0204, 0.0201, 0.0186, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 16:57:48,020 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 2250, loss[loss=0.1778, simple_loss=0.2458, pruned_loss=0.05493, over 4059.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2444, pruned_loss=0.05005, over 952837.03 frames. ], batch size: 65, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:58:19,798 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122551.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:58:34,012 INFO [optim.py:369] (2/7) 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 16:58:52,866 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0431, 3.9719, 2.8172, 4.6441, 4.1027, 3.9652, 1.6586, 4.0220], + device='cuda:2'), covar=tensor([0.1774, 0.1349, 0.3216, 0.1568, 0.4005, 0.1796, 0.5904, 0.2365], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0217, 0.0253, 0.0308, 0.0296, 0.0247, 0.0277, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 16:58:54,147 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8793, 1.4199, 1.9599, 2.3646, 1.9570, 1.7864, 1.8979, 1.8496], + device='cuda:2'), covar=tensor([0.4659, 0.6888, 0.6579, 0.5648, 0.5924, 0.8024, 0.8016, 0.9295], + device='cuda:2'), in_proj_covar=tensor([0.0433, 0.0415, 0.0510, 0.0508, 0.0461, 0.0492, 0.0498, 0.0506], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 16:58:56,472 INFO [finetune.py:976] (2/7) Epoch 22, batch 2300, loss[loss=0.1607, simple_loss=0.2299, pruned_loss=0.04579, over 4922.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2442, pruned_loss=0.04952, over 954988.36 frames. ], batch size: 42, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 16:59:36,000 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122606.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 16:59:37,297 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5666, 1.1362, 1.3459, 1.2046, 1.6813, 1.3921, 1.0779, 1.2707], + device='cuda:2'), covar=tensor([0.1504, 0.1471, 0.1980, 0.1502, 0.0932, 0.1422, 0.2230, 0.2609], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0311, 0.0350, 0.0289, 0.0325, 0.0308, 0.0301, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.4136e-05, 6.4362e-05, 7.3669e-05, 5.8254e-05, 6.7018e-05, 6.4631e-05, + 6.2831e-05, 7.8818e-05], device='cuda:2') +2023-04-27 17:00:08,779 INFO [finetune.py:976] (2/7) Epoch 22, batch 2350, loss[loss=0.1797, simple_loss=0.2355, pruned_loss=0.06191, over 4785.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2434, pruned_loss=0.04959, over 952196.34 frames. ], batch size: 51, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 17:00:40,760 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 17:01:04,136 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 2400, loss[loss=0.1776, simple_loss=0.2444, pruned_loss=0.0554, over 4930.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2407, pruned_loss=0.04861, over 953285.90 frames. ], batch size: 38, lr: 3.15e-03, grad_scale: 32.0 +2023-04-27 17:01:16,172 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4753, 1.7272, 1.8609, 1.9773, 1.8707, 1.9430, 1.9303, 1.9429], + device='cuda:2'), covar=tensor([0.3818, 0.5180, 0.4337, 0.4346, 0.5370, 0.6758, 0.4797, 0.4657], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0372, 0.0325, 0.0339, 0.0347, 0.0395, 0.0356, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:02:21,393 INFO [finetune.py:976] (2/7) Epoch 22, batch 2450, loss[loss=0.2309, simple_loss=0.2901, pruned_loss=0.08586, over 4807.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2383, pruned_loss=0.04751, over 953142.06 frames. ], batch size: 45, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:02:37,637 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.51 vs. limit=2.0 +2023-04-27 17:02:42,827 INFO [optim.py:369] (2/7) 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 17:02:48,958 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122773.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 17:02:53,910 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-04-27 17:02:54,348 INFO [finetune.py:976] (2/7) Epoch 22, batch 2500, loss[loss=0.2007, simple_loss=0.2773, pruned_loss=0.06206, over 4931.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2385, pruned_loss=0.04773, over 953234.93 frames. ], batch size: 42, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:03:08,269 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2339, 1.5096, 1.3180, 1.7388, 1.5790, 1.9607, 1.3581, 3.6046], + device='cuda:2'), covar=tensor([0.0589, 0.0813, 0.0834, 0.1236, 0.0641, 0.0547, 0.0785, 0.0137], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 17:03:21,722 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122821.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:03:23,546 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 2550, loss[loss=0.166, simple_loss=0.2457, pruned_loss=0.04315, over 4744.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2407, pruned_loss=0.04797, over 954941.13 frames. ], batch size: 59, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:03:40,085 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 17:04:01,375 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 2600, loss[loss=0.1699, simple_loss=0.2552, pruned_loss=0.04234, over 4821.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.243, pruned_loss=0.04884, over 952297.32 frames. ], batch size: 38, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:04:33,759 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0814, 1.4547, 1.3601, 1.6662, 1.5309, 1.6960, 1.3270, 3.0482], + device='cuda:2'), covar=tensor([0.0663, 0.0822, 0.0788, 0.1184, 0.0642, 0.0564, 0.0750, 0.0166], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 17:04:34,929 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122899.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:05:19,783 INFO [finetune.py:976] (2/7) Epoch 22, batch 2650, loss[loss=0.1769, simple_loss=0.2611, pruned_loss=0.04633, over 4745.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2447, pruned_loss=0.04909, over 953562.41 frames. ], batch size: 27, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:05:20,539 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 17:06:15,904 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 2700, loss[loss=0.203, simple_loss=0.2584, pruned_loss=0.07379, over 4915.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2439, pruned_loss=0.04892, over 952566.55 frames. ], batch size: 38, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:06:56,567 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 2750, loss[loss=0.1499, simple_loss=0.2241, pruned_loss=0.0379, over 4825.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2412, pruned_loss=0.04847, over 952924.16 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:07:28,109 INFO [optim.py:369] (2/7) 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 17:07:53,158 INFO [finetune.py:976] (2/7) Epoch 22, batch 2800, loss[loss=0.1575, simple_loss=0.2276, pruned_loss=0.04368, over 4859.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2383, pruned_loss=0.04755, over 953427.96 frames. ], batch size: 44, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:08:48,178 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123124.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:08:53,536 INFO [finetune.py:976] (2/7) Epoch 22, batch 2850, loss[loss=0.1839, simple_loss=0.2513, pruned_loss=0.05822, over 4758.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2376, pruned_loss=0.04768, over 953446.95 frames. ], batch size: 54, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:09:01,426 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0384, 1.5355, 1.4734, 1.6864, 1.6910, 1.9640, 1.3412, 3.4605], + device='cuda:2'), covar=tensor([0.0606, 0.0789, 0.0755, 0.1224, 0.0615, 0.0565, 0.0780, 0.0139], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 17:09:06,926 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3172, 3.0226, 0.8887, 1.6823, 1.8247, 2.1851, 1.7930, 0.9308], + device='cuda:2'), covar=tensor([0.1499, 0.1248, 0.1959, 0.1358, 0.1118, 0.1083, 0.1702, 0.2000], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0241, 0.0138, 0.0120, 0.0133, 0.0153, 0.0118, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:09:12,897 INFO [optim.py:369] (2/7) 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 17:09:19,397 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123172.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:09:27,500 INFO [finetune.py:976] (2/7) Epoch 22, batch 2900, loss[loss=0.2052, simple_loss=0.2769, pruned_loss=0.0667, over 4837.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2401, pruned_loss=0.04863, over 954428.65 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:09:42,982 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3167, 3.4963, 1.0369, 1.9078, 1.9614, 2.5533, 1.9894, 1.0230], + device='cuda:2'), covar=tensor([0.1460, 0.1079, 0.1906, 0.1187, 0.1023, 0.0959, 0.1531, 0.1978], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0240, 0.0137, 0.0119, 0.0132, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:09:48,566 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9752, 2.3030, 0.9769, 1.2329, 1.9973, 1.1970, 3.0263, 1.6222], + device='cuda:2'), covar=tensor([0.0713, 0.0573, 0.0747, 0.1366, 0.0464, 0.1061, 0.0320, 0.0654], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0048, 0.0046, 0.0049, 0.0052, 0.0074, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 17:10:09,345 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 2950, loss[loss=0.1731, simple_loss=0.2567, pruned_loss=0.0447, over 4935.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2441, pruned_loss=0.04957, over 955074.16 frames. ], batch size: 38, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:10:20,211 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 17:10:52,151 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5577, 3.1978, 1.0711, 2.0481, 1.9858, 2.5601, 2.0262, 1.2846], + device='cuda:2'), covar=tensor([0.1135, 0.0875, 0.1694, 0.0985, 0.0932, 0.0803, 0.1234, 0.1901], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0138, 0.0120, 0.0133, 0.0153, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:10:53,369 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6097, 3.7331, 0.9270, 1.9559, 2.1563, 2.5945, 2.0256, 1.0317], + device='cuda:2'), covar=tensor([0.1355, 0.0828, 0.2015, 0.1205, 0.0976, 0.1040, 0.1554, 0.1975], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0138, 0.0120, 0.0133, 0.0153, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:10:55,130 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 22, batch 3000, loss[loss=0.1554, simple_loss=0.2314, pruned_loss=0.03973, over 4908.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2446, pruned_loss=0.04974, over 955243.97 frames. ], batch size: 42, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:11:18,371 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 17:11:24,615 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8216, 2.1910, 1.8351, 1.6007, 1.4062, 1.4045, 1.8198, 1.3829], + device='cuda:2'), covar=tensor([0.1587, 0.1372, 0.1378, 0.1696, 0.2361, 0.1935, 0.1000, 0.2007], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0211, 0.0169, 0.0204, 0.0200, 0.0186, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 17:11:29,116 INFO [finetune.py:1010] (2/7) Epoch 22, validation: loss=0.1537, simple_loss=0.2227, pruned_loss=0.04237, over 2265189.00 frames. +2023-04-27 17:11:29,116 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 17:12:03,187 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1833, 1.5729, 1.4854, 1.7824, 1.7461, 1.9070, 1.4156, 3.6172], + device='cuda:2'), covar=tensor([0.0586, 0.0767, 0.0758, 0.1155, 0.0589, 0.0490, 0.0725, 0.0138], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 17:12:11,382 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123314.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:12:11,557 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-27 17:12:33,514 INFO [finetune.py:976] (2/7) Epoch 22, batch 3050, loss[loss=0.2723, simple_loss=0.3266, pruned_loss=0.1091, over 4207.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2459, pruned_loss=0.05031, over 954075.44 frames. ], batch size: 66, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:12:42,861 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 17:12:59,240 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3591, 1.8517, 1.6581, 2.2332, 2.3897, 1.9431, 1.9217, 1.7353], + device='cuda:2'), covar=tensor([0.1556, 0.1790, 0.1781, 0.1364, 0.1168, 0.2031, 0.2474, 0.2380], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0315, 0.0355, 0.0291, 0.0329, 0.0313, 0.0304, 0.0377], + device='cuda:2'), out_proj_covar=tensor([6.4911e-05, 6.5066e-05, 7.4855e-05, 5.8730e-05, 6.7857e-05, 6.5524e-05, + 6.3455e-05, 8.0009e-05], device='cuda:2') +2023-04-27 17:13:01,591 INFO [optim.py:369] (2/7) 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 17:13:14,308 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123375.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:13:24,033 INFO [finetune.py:976] (2/7) Epoch 22, batch 3100, loss[loss=0.1661, simple_loss=0.2389, pruned_loss=0.04667, over 4853.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2441, pruned_loss=0.04959, over 955701.41 frames. ], batch size: 31, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:13:43,359 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-27 17:14:05,839 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 17:14:29,441 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123428.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:14:31,788 INFO [finetune.py:976] (2/7) Epoch 22, batch 3150, loss[loss=0.1683, simple_loss=0.2334, pruned_loss=0.05157, over 4758.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2425, pruned_loss=0.04969, over 958346.56 frames. ], batch size: 28, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:15:16,033 INFO [optim.py:369] (2/7) 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 17:15:24,646 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5739, 1.4223, 4.4548, 4.2013, 3.9248, 4.2339, 4.0667, 3.9253], + device='cuda:2'), covar=tensor([0.6880, 0.5957, 0.0966, 0.1489, 0.1061, 0.1889, 0.1357, 0.1511], + device='cuda:2'), in_proj_covar=tensor([0.0306, 0.0303, 0.0403, 0.0403, 0.0344, 0.0405, 0.0308, 0.0362], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:15:38,629 INFO [finetune.py:976] (2/7) Epoch 22, batch 3200, loss[loss=0.1591, simple_loss=0.2299, pruned_loss=0.04419, over 4823.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2394, pruned_loss=0.04857, over 955414.85 frames. ], batch size: 41, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:15:48,875 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123489.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:15:56,431 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6600, 1.7208, 1.5337, 1.1149, 1.2780, 1.2488, 1.5229, 1.2356], + device='cuda:2'), covar=tensor([0.1800, 0.1398, 0.1600, 0.1868, 0.2355, 0.2076, 0.1123, 0.2106], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0212, 0.0169, 0.0204, 0.0200, 0.0186, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 17:16:43,112 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123528.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:16:51,257 INFO [finetune.py:976] (2/7) Epoch 22, batch 3250, loss[loss=0.1477, simple_loss=0.229, pruned_loss=0.03323, over 4789.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2398, pruned_loss=0.0486, over 957345.68 frames. ], batch size: 29, lr: 3.14e-03, grad_scale: 64.0 +2023-04-27 17:17:02,484 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-27 17:17:26,393 INFO [optim.py:369] (2/7) 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 17:17:40,415 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123576.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:17:44,044 INFO [finetune.py:976] (2/7) Epoch 22, batch 3300, loss[loss=0.1557, simple_loss=0.2391, pruned_loss=0.03617, over 4765.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2428, pruned_loss=0.04902, over 957516.75 frames. ], batch size: 28, lr: 3.14e-03, grad_scale: 64.0 +2023-04-27 17:17:45,968 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3994, 0.9107, 0.4393, 1.1025, 1.1169, 1.2893, 1.1960, 1.1861], + device='cuda:2'), covar=tensor([0.0498, 0.0418, 0.0408, 0.0559, 0.0305, 0.0529, 0.0506, 0.0562], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 17:18:17,764 INFO [finetune.py:976] (2/7) Epoch 22, batch 3350, loss[loss=0.2006, simple_loss=0.2662, pruned_loss=0.06748, over 4819.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2454, pruned_loss=0.04973, over 958120.26 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 64.0 +2023-04-27 17:18:40,219 INFO [optim.py:369] (2/7) 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 17:18:43,917 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123670.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:18:48,068 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2012, 1.4502, 1.4074, 1.6909, 1.5939, 1.8428, 1.3367, 3.2859], + device='cuda:2'), covar=tensor([0.0599, 0.0808, 0.0761, 0.1198, 0.0612, 0.0516, 0.0742, 0.0150], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 17:18:51,586 INFO [finetune.py:976] (2/7) Epoch 22, batch 3400, loss[loss=0.1757, simple_loss=0.2571, pruned_loss=0.04709, over 4846.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2472, pruned_loss=0.05067, over 956487.34 frames. ], batch size: 44, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:18:59,631 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8120, 1.8009, 0.8284, 1.5062, 2.0016, 1.6602, 1.5910, 1.7086], + device='cuda:2'), covar=tensor([0.0483, 0.0370, 0.0336, 0.0553, 0.0250, 0.0503, 0.0485, 0.0542], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 17:19:25,130 INFO [finetune.py:976] (2/7) Epoch 22, batch 3450, loss[loss=0.1435, simple_loss=0.2112, pruned_loss=0.03793, over 4020.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2466, pruned_loss=0.05021, over 955490.19 frames. ], batch size: 17, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:19:57,043 INFO [optim.py:369] (2/7) 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 17:20:21,185 INFO [finetune.py:976] (2/7) Epoch 22, batch 3500, loss[loss=0.178, simple_loss=0.2412, pruned_loss=0.05739, over 4911.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2453, pruned_loss=0.05053, over 955129.54 frames. ], batch size: 46, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:20:28,351 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:21:22,864 INFO [finetune.py:976] (2/7) Epoch 22, batch 3550, loss[loss=0.1399, simple_loss=0.2061, pruned_loss=0.03687, over 4817.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2424, pruned_loss=0.04951, over 956212.76 frames. ], batch size: 25, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:21:23,631 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5005, 1.8022, 1.9292, 2.0701, 1.8900, 1.9368, 1.9919, 1.9762], + device='cuda:2'), covar=tensor([0.3707, 0.4879, 0.4216, 0.4348, 0.5042, 0.6828, 0.4961, 0.4369], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0374, 0.0326, 0.0340, 0.0348, 0.0396, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:21:43,412 INFO [optim.py:369] (2/7) 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 17:21:49,376 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 3600, loss[loss=0.1823, simple_loss=0.244, pruned_loss=0.06028, over 4872.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2403, pruned_loss=0.04905, over 957375.35 frames. ], batch size: 34, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:21:57,657 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-27 17:22:27,080 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3253, 2.7060, 2.1923, 2.6671, 1.8192, 2.4960, 2.5426, 1.8832], + device='cuda:2'), covar=tensor([0.1731, 0.1007, 0.0780, 0.1192, 0.3325, 0.0997, 0.1677, 0.2338], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0302, 0.0218, 0.0278, 0.0316, 0.0257, 0.0249, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1420e-04, 1.1935e-04, 8.6249e-05, 1.0987e-04, 1.2776e-04, 1.0154e-04, + 1.0067e-04, 1.0418e-04], device='cuda:2') +2023-04-27 17:22:30,655 INFO [finetune.py:976] (2/7) Epoch 22, batch 3650, loss[loss=0.1457, simple_loss=0.2285, pruned_loss=0.03142, over 4908.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2409, pruned_loss=0.04891, over 957770.61 frames. ], batch size: 37, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:22:31,398 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:23:14,431 INFO [optim.py:369] (2/7) 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 17:23:23,384 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123970.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:23:37,124 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3505, 3.5268, 0.9011, 1.7150, 1.8670, 2.4517, 1.9916, 1.0423], + device='cuda:2'), covar=tensor([0.1545, 0.0956, 0.2056, 0.1373, 0.1086, 0.1078, 0.1504, 0.2039], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0241, 0.0139, 0.0120, 0.0133, 0.0153, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:23:37,657 INFO [finetune.py:976] (2/7) Epoch 22, batch 3700, loss[loss=0.1738, simple_loss=0.259, pruned_loss=0.04431, over 4805.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2445, pruned_loss=0.04975, over 956168.15 frames. ], batch size: 40, lr: 3.14e-03, grad_scale: 32.0 +2023-04-27 17:24:10,267 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 17:24:22,553 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:24:44,186 INFO [finetune.py:976] (2/7) Epoch 22, batch 3750, loss[loss=0.1752, simple_loss=0.2513, pruned_loss=0.04956, over 4822.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2456, pruned_loss=0.04994, over 956613.20 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:24:52,266 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:24:55,513 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 17:25:08,732 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124062.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:25:09,361 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4137, 2.8926, 1.0478, 1.6462, 2.4162, 1.4445, 4.0336, 2.1929], + device='cuda:2'), covar=tensor([0.0655, 0.0733, 0.0886, 0.1247, 0.0484, 0.0999, 0.0207, 0.0561], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 17:25:09,866 INFO [optim.py:369] (2/7) 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 17:25:13,603 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1320, 2.6156, 1.0877, 1.4538, 2.3098, 1.3488, 3.5057, 1.9528], + device='cuda:2'), covar=tensor([0.0645, 0.0540, 0.0786, 0.1295, 0.0454, 0.1003, 0.0263, 0.0584], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 17:25:18,822 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7992, 1.7493, 1.6716, 1.3556, 1.7958, 1.6052, 2.2737, 1.4777], + device='cuda:2'), covar=tensor([0.3078, 0.1623, 0.3954, 0.2411, 0.1538, 0.1961, 0.1254, 0.4373], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0344, 0.0420, 0.0349, 0.0375, 0.0370, 0.0365, 0.0413], + device='cuda:2'), out_proj_covar=tensor([9.9175e-05, 1.0290e-04, 1.2756e-04, 1.0517e-04, 1.1170e-04, 1.1016e-04, + 1.0724e-04, 1.2447e-04], device='cuda:2') +2023-04-27 17:25:22,231 INFO [finetune.py:976] (2/7) Epoch 22, batch 3800, loss[loss=0.1696, simple_loss=0.2409, pruned_loss=0.04911, over 4803.00 frames. ], tot_loss[loss=0.174, simple_loss=0.247, pruned_loss=0.05051, over 956929.49 frames. ], batch size: 25, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:25:23,552 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124084.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:25:32,935 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:25:43,904 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1419, 1.8350, 2.3289, 2.6217, 2.1918, 2.0873, 2.2148, 2.1628], + device='cuda:2'), covar=tensor([0.4419, 0.6949, 0.7112, 0.5172, 0.5621, 0.8108, 0.7972, 0.9356], + device='cuda:2'), in_proj_covar=tensor([0.0430, 0.0414, 0.0509, 0.0505, 0.0461, 0.0490, 0.0496, 0.0505], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:25:49,437 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 3850, loss[loss=0.1928, simple_loss=0.2665, pruned_loss=0.05958, over 4819.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2451, pruned_loss=0.04926, over 955247.40 frames. ], batch size: 40, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:25:56,354 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124132.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:26:17,348 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-04-27 17:26:17,750 INFO [optim.py:369] (2/7) 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 17:26:35,209 INFO [finetune.py:976] (2/7) Epoch 22, batch 3900, loss[loss=0.1818, simple_loss=0.2551, pruned_loss=0.05429, over 4909.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2422, pruned_loss=0.04872, over 954748.76 frames. ], batch size: 36, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:27:19,065 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0517, 1.0123, 1.5692, 1.6880, 1.6314, 1.6886, 1.5978, 1.5945], + device='cuda:2'), covar=tensor([0.3728, 0.5121, 0.4392, 0.4422, 0.5326, 0.7236, 0.4650, 0.4533], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0340, 0.0348, 0.0396, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:27:29,074 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4553, 3.3916, 2.4514, 3.9024, 3.3593, 3.4345, 1.4677, 3.3515], + device='cuda:2'), covar=tensor([0.1638, 0.1347, 0.3344, 0.2205, 0.2922, 0.1884, 0.6058, 0.2421], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0217, 0.0252, 0.0305, 0.0295, 0.0246, 0.0275, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:27:38,095 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124228.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:27:40,957 INFO [finetune.py:976] (2/7) Epoch 22, batch 3950, loss[loss=0.1561, simple_loss=0.2356, pruned_loss=0.0383, over 4899.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2395, pruned_loss=0.0477, over 956234.53 frames. ], batch size: 35, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:27:54,340 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-27 17:27:59,546 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 17:28:07,730 INFO [optim.py:369] (2/7) 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 17:28:13,199 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 4000, loss[loss=0.1467, simple_loss=0.2306, pruned_loss=0.03146, over 4820.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2402, pruned_loss=0.04888, over 953280.38 frames. ], batch size: 39, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:28:45,643 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-27 17:29:09,168 INFO [finetune.py:976] (2/7) Epoch 22, batch 4050, loss[loss=0.2007, simple_loss=0.2739, pruned_loss=0.06374, over 4817.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2421, pruned_loss=0.04954, over 952275.26 frames. ], batch size: 40, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:29:09,894 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124333.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:29:32,578 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4306, 3.3281, 1.0398, 1.8689, 1.7458, 2.4721, 1.9537, 1.0454], + device='cuda:2'), covar=tensor([0.1333, 0.0880, 0.1788, 0.1194, 0.1080, 0.0965, 0.1347, 0.2145], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0242, 0.0139, 0.0121, 0.0133, 0.0153, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:29:38,969 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124348.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 17:29:54,262 INFO [optim.py:369] (2/7) 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 17:30:17,167 INFO [finetune.py:976] (2/7) Epoch 22, batch 4100, loss[loss=0.1968, simple_loss=0.2719, pruned_loss=0.06082, over 4842.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2456, pruned_loss=0.05041, over 952032.34 frames. ], batch size: 44, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:30:23,713 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 17:30:34,120 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124392.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:30:36,052 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.96 vs. limit=5.0 +2023-04-27 17:30:58,099 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124409.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 17:31:08,995 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124418.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:31:27,632 INFO [finetune.py:976] (2/7) Epoch 22, batch 4150, loss[loss=0.1744, simple_loss=0.247, pruned_loss=0.05092, over 4860.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2463, pruned_loss=0.05009, over 953053.71 frames. ], batch size: 34, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:31:27,756 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:31:28,991 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0823, 2.4780, 1.2234, 1.4095, 2.0970, 1.2975, 3.0680, 1.6954], + device='cuda:2'), covar=tensor([0.0651, 0.0609, 0.0734, 0.1131, 0.0419, 0.0911, 0.0246, 0.0561], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 17:32:10,275 INFO [optim.py:369] (2/7) 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 17:32:12,306 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-27 17:32:33,359 INFO [finetune.py:976] (2/7) Epoch 22, batch 4200, loss[loss=0.1865, simple_loss=0.2548, pruned_loss=0.05906, over 4907.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2464, pruned_loss=0.04996, over 951025.29 frames. ], batch size: 36, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:32:45,620 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124493.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:33:09,238 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124528.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:33:11,582 INFO [finetune.py:976] (2/7) Epoch 22, batch 4250, loss[loss=0.172, simple_loss=0.2413, pruned_loss=0.05137, over 4715.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2438, pruned_loss=0.04928, over 950413.04 frames. ], batch size: 54, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:33:32,215 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-04-27 17:33:33,183 INFO [optim.py:369] (2/7) 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 17:33:40,672 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124576.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:33:44,718 INFO [finetune.py:976] (2/7) Epoch 22, batch 4300, loss[loss=0.1722, simple_loss=0.2366, pruned_loss=0.05384, over 4719.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2403, pruned_loss=0.04807, over 952022.33 frames. ], batch size: 23, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:33:53,433 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2641, 1.6860, 2.1099, 2.3457, 2.0889, 1.7233, 1.3294, 1.9180], + device='cuda:2'), covar=tensor([0.2777, 0.2989, 0.1511, 0.2256, 0.2368, 0.2319, 0.4002, 0.1849], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0245, 0.0225, 0.0314, 0.0220, 0.0232, 0.0228, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 17:34:14,916 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5912, 1.5088, 1.9300, 1.9492, 1.4827, 1.3090, 1.6309, 1.0196], + device='cuda:2'), covar=tensor([0.0500, 0.0595, 0.0336, 0.0506, 0.0769, 0.1078, 0.0555, 0.0612], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0067, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:34:15,437 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124628.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:34:17,827 INFO [finetune.py:976] (2/7) Epoch 22, batch 4350, loss[loss=0.1471, simple_loss=0.2174, pruned_loss=0.03838, over 4856.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2381, pruned_loss=0.04777, over 951778.86 frames. ], batch size: 47, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:34:22,039 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124638.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:34:38,893 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124663.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:34:39,361 INFO [optim.py:369] (2/7) 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 17:34:51,252 INFO [finetune.py:976] (2/7) Epoch 22, batch 4400, loss[loss=0.1878, simple_loss=0.246, pruned_loss=0.06476, over 4847.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2383, pruned_loss=0.04827, over 952875.05 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:34:58,042 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124692.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:02,332 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124699.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:11,384 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124704.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 17:35:33,077 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124718.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:42,956 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124724.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:35:53,111 INFO [finetune.py:976] (2/7) Epoch 22, batch 4450, loss[loss=0.1525, simple_loss=0.2333, pruned_loss=0.03585, over 4876.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2409, pruned_loss=0.04894, over 951903.92 frames. ], batch size: 32, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:36:03,832 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124766.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:36:40,620 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6680, 2.0410, 1.7532, 1.9720, 1.5454, 1.6994, 1.7134, 1.2536], + device='cuda:2'), covar=tensor([0.1719, 0.1108, 0.0750, 0.1047, 0.3120, 0.1071, 0.1790, 0.2246], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0299, 0.0216, 0.0275, 0.0314, 0.0255, 0.0248, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1343e-04, 1.1832e-04, 8.5249e-05, 1.0832e-04, 1.2715e-04, 1.0107e-04, + 1.0015e-04, 1.0336e-04], device='cuda:2') +2023-04-27 17:37:00,152 INFO [finetune.py:976] (2/7) Epoch 22, batch 4500, loss[loss=0.2146, simple_loss=0.2968, pruned_loss=0.06615, over 4918.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2425, pruned_loss=0.04934, over 951753.66 frames. ], batch size: 42, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:37:03,938 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124788.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:37:26,165 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5976, 1.1560, 4.4132, 4.1070, 3.8419, 4.1913, 4.1329, 3.7953], + device='cuda:2'), covar=tensor([0.7205, 0.6549, 0.1101, 0.1934, 0.1216, 0.1571, 0.1392, 0.1714], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0306, 0.0408, 0.0409, 0.0349, 0.0411, 0.0314, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:37:47,334 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124820.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:37:58,274 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7739, 2.2242, 2.0640, 2.2614, 2.0783, 2.1409, 2.1163, 2.1133], + device='cuda:2'), covar=tensor([0.4457, 0.5906, 0.5457, 0.4105, 0.5879, 0.6740, 0.6289, 0.5458], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0374, 0.0325, 0.0340, 0.0349, 0.0395, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:38:07,812 INFO [finetune.py:976] (2/7) Epoch 22, batch 4550, loss[loss=0.1772, simple_loss=0.2528, pruned_loss=0.05075, over 4781.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2437, pruned_loss=0.04922, over 953207.44 frames. ], batch size: 51, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:38:32,610 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.73 vs. limit=5.0 +2023-04-27 17:38:33,224 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.68 vs. limit=5.0 +2023-04-27 17:38:49,605 INFO [optim.py:369] (2/7) 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 17:38:51,381 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2542, 1.4498, 1.4740, 1.6880, 1.6448, 1.7775, 1.3757, 3.3025], + device='cuda:2'), covar=tensor([0.0612, 0.0842, 0.0755, 0.1247, 0.0653, 0.0547, 0.0754, 0.0159], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0039, 0.0037, 0.0037, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 17:39:13,971 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 4600, loss[loss=0.1294, simple_loss=0.2053, pruned_loss=0.02674, over 3998.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2451, pruned_loss=0.04947, over 954248.92 frames. ], batch size: 17, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:39:36,588 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1046, 2.7573, 1.0797, 1.6490, 2.1928, 1.2942, 3.7459, 1.9387], + device='cuda:2'), covar=tensor([0.0731, 0.0660, 0.0832, 0.1249, 0.0514, 0.1085, 0.0289, 0.0611], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0050, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 17:40:00,421 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124928.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:40:03,269 INFO [finetune.py:976] (2/7) Epoch 22, batch 4650, loss[loss=0.1699, simple_loss=0.2469, pruned_loss=0.04644, over 4844.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2421, pruned_loss=0.04852, over 956608.23 frames. ], batch size: 47, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:40:09,505 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2586, 1.9864, 2.3154, 2.4725, 2.5169, 2.0536, 1.9805, 2.2752], + device='cuda:2'), covar=tensor([0.0775, 0.1064, 0.0617, 0.0623, 0.0554, 0.0822, 0.0675, 0.0577], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0202, 0.0184, 0.0176, 0.0178, 0.0181, 0.0152, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:40:14,479 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8069, 1.3096, 1.9306, 2.2835, 1.9451, 1.8151, 1.8484, 1.8200], + device='cuda:2'), covar=tensor([0.4204, 0.6342, 0.5950, 0.5585, 0.5515, 0.7382, 0.7483, 0.8309], + device='cuda:2'), in_proj_covar=tensor([0.0433, 0.0417, 0.0510, 0.0507, 0.0463, 0.0492, 0.0498, 0.0510], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:40:23,401 INFO [optim.py:369] (2/7) 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 17:40:30,021 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-04-27 17:40:31,673 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124976.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:40:35,875 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0122, 2.2878, 2.2618, 2.3945, 2.0812, 2.2391, 2.2390, 2.1716], + device='cuda:2'), covar=tensor([0.3315, 0.5767, 0.4710, 0.4546, 0.5725, 0.7108, 0.6380, 0.5671], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0374, 0.0324, 0.0340, 0.0348, 0.0394, 0.0358, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:40:36,852 INFO [finetune.py:976] (2/7) Epoch 22, batch 4700, loss[loss=0.1142, simple_loss=0.1913, pruned_loss=0.01851, over 4767.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2388, pruned_loss=0.04736, over 958871.74 frames. ], batch size: 28, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:40:44,818 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124994.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:40:51,153 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125004.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 17:40:56,024 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.82 vs. limit=5.0 +2023-04-27 17:41:00,863 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125019.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:41:10,032 INFO [finetune.py:976] (2/7) Epoch 22, batch 4750, loss[loss=0.1833, simple_loss=0.2461, pruned_loss=0.06026, over 4832.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2381, pruned_loss=0.04773, over 958939.92 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:41:12,366 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9760, 2.8536, 2.3993, 3.3274, 2.9498, 2.9268, 1.2568, 2.8589], + device='cuda:2'), covar=tensor([0.2036, 0.1918, 0.3112, 0.2675, 0.3048, 0.2141, 0.5630, 0.3079], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0215, 0.0248, 0.0303, 0.0293, 0.0244, 0.0272, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:41:23,893 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125052.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 17:41:31,618 INFO [optim.py:369] (2/7) 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 17:41:43,194 INFO [finetune.py:976] (2/7) Epoch 22, batch 4800, loss[loss=0.2028, simple_loss=0.2635, pruned_loss=0.07102, over 4816.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2416, pruned_loss=0.04918, over 958619.84 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:41:48,396 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:41:53,777 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8697, 2.2516, 1.9424, 2.2266, 1.6630, 1.8439, 1.8866, 1.4721], + device='cuda:2'), covar=tensor([0.1789, 0.1197, 0.0797, 0.1116, 0.3272, 0.1212, 0.1925, 0.2485], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0300, 0.0217, 0.0276, 0.0315, 0.0256, 0.0249, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1367e-04, 1.1853e-04, 8.5657e-05, 1.0886e-04, 1.2762e-04, 1.0124e-04, + 1.0039e-04, 1.0362e-04], device='cuda:2') +2023-04-27 17:42:04,686 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6730, 3.4833, 1.1381, 1.8060, 2.1320, 2.6035, 1.9542, 1.1739], + device='cuda:2'), covar=tensor([0.1263, 0.0783, 0.1689, 0.1191, 0.0940, 0.0870, 0.1509, 0.1833], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0120, 0.0132, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:42:14,817 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125128.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:42:17,143 INFO [finetune.py:976] (2/7) Epoch 22, batch 4850, loss[loss=0.1976, simple_loss=0.2771, pruned_loss=0.05907, over 4850.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2441, pruned_loss=0.04972, over 957092.01 frames. ], batch size: 44, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:42:20,149 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:42:23,622 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1650, 1.8090, 2.3391, 2.6337, 2.2210, 2.1077, 2.2315, 2.1659], + device='cuda:2'), covar=tensor([0.4705, 0.7586, 0.7203, 0.5623, 0.6587, 0.8406, 0.9148, 0.9783], + device='cuda:2'), in_proj_covar=tensor([0.0432, 0.0417, 0.0510, 0.0507, 0.0463, 0.0492, 0.0499, 0.0510], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:42:31,985 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1210, 2.5426, 2.3626, 2.5816, 2.2830, 2.5465, 2.4423, 2.3734], + device='cuda:2'), covar=tensor([0.3423, 0.5048, 0.4702, 0.3983, 0.5212, 0.6154, 0.5244, 0.5067], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0373, 0.0324, 0.0339, 0.0347, 0.0393, 0.0357, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:42:44,288 INFO [optim.py:369] (2/7) 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 17:43:03,304 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125176.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:43:06,314 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8076, 2.2682, 2.0927, 2.3463, 2.0413, 2.1882, 2.1229, 2.1223], + device='cuda:2'), covar=tensor([0.4031, 0.5431, 0.5233, 0.3908, 0.5732, 0.6430, 0.6108, 0.5057], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0372, 0.0323, 0.0338, 0.0346, 0.0392, 0.0356, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:43:13,296 INFO [finetune.py:976] (2/7) Epoch 22, batch 4900, loss[loss=0.178, simple_loss=0.2463, pruned_loss=0.05486, over 4924.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2459, pruned_loss=0.05034, over 956196.96 frames. ], batch size: 33, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:43:18,234 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125189.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:43:35,063 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 17:44:14,016 INFO [finetune.py:976] (2/7) Epoch 22, batch 4950, loss[loss=0.1588, simple_loss=0.2317, pruned_loss=0.04295, over 4734.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2462, pruned_loss=0.05024, over 956436.42 frames. ], batch size: 27, lr: 3.13e-03, grad_scale: 32.0 +2023-04-27 17:44:47,946 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4121, 2.2958, 2.1619, 2.0078, 2.4537, 1.9901, 3.0490, 1.9236], + device='cuda:2'), covar=tensor([0.3294, 0.1747, 0.3869, 0.2620, 0.1531, 0.2490, 0.1223, 0.3791], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0349, 0.0426, 0.0354, 0.0381, 0.0374, 0.0369, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:44:50,839 INFO [optim.py:369] (2/7) 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 17:45:13,457 INFO [finetune.py:976] (2/7) Epoch 22, batch 5000, loss[loss=0.1677, simple_loss=0.2242, pruned_loss=0.05556, over 4763.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2456, pruned_loss=0.05018, over 956254.63 frames. ], batch size: 59, lr: 3.13e-03, grad_scale: 16.0 +2023-04-27 17:45:23,252 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3755, 1.2992, 1.6435, 1.6118, 1.3005, 1.2363, 1.2632, 0.6782], + device='cuda:2'), covar=tensor([0.0573, 0.0630, 0.0417, 0.0571, 0.0894, 0.1173, 0.0516, 0.0657], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0069, 0.0068, 0.0069, 0.0076, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:45:32,193 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9128, 2.1751, 2.4717, 2.5975, 2.4105, 2.4579, 2.2783, 4.9815], + device='cuda:2'), covar=tensor([0.0479, 0.0709, 0.0661, 0.1002, 0.0579, 0.0442, 0.0630, 0.0113], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0039, 0.0037, 0.0037, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 17:45:32,810 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125294.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:05,806 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125319.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:18,221 INFO [finetune.py:976] (2/7) Epoch 22, batch 5050, loss[loss=0.1517, simple_loss=0.2252, pruned_loss=0.0391, over 4906.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2428, pruned_loss=0.0494, over 957650.51 frames. ], batch size: 43, lr: 3.13e-03, grad_scale: 16.0 +2023-04-27 17:46:32,850 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125342.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:43,462 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:48,789 INFO [optim.py:369] (2/7) 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 17:46:50,093 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125367.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:51,998 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8947, 2.1896, 2.0770, 2.2818, 2.0292, 2.2151, 2.0652, 2.0759], + device='cuda:2'), covar=tensor([0.4074, 0.6237, 0.4912, 0.4616, 0.5727, 0.6945, 0.6855, 0.6314], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0374, 0.0325, 0.0341, 0.0348, 0.0394, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:46:55,690 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125376.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:46:59,779 INFO [finetune.py:976] (2/7) Epoch 22, batch 5100, loss[loss=0.1492, simple_loss=0.2239, pruned_loss=0.03721, over 4909.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2406, pruned_loss=0.04883, over 959030.72 frames. ], batch size: 32, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:47:24,382 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125417.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 17:47:26,787 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125421.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:47:33,405 INFO [finetune.py:976] (2/7) Epoch 22, batch 5150, loss[loss=0.1153, simple_loss=0.1838, pruned_loss=0.02335, over 3910.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2397, pruned_loss=0.04857, over 955137.70 frames. ], batch size: 17, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:47:37,080 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125437.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:47:43,521 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4956, 1.3338, 1.7772, 1.7265, 1.3573, 1.2412, 1.4272, 0.8419], + device='cuda:2'), covar=tensor([0.0521, 0.0668, 0.0378, 0.0492, 0.0763, 0.1147, 0.0649, 0.0674], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:48:02,909 INFO [optim.py:369] (2/7) 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 17:48:09,680 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125476.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:13,157 INFO [finetune.py:976] (2/7) Epoch 22, batch 5200, loss[loss=0.1932, simple_loss=0.2675, pruned_loss=0.05941, over 4892.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2417, pruned_loss=0.04898, over 953336.39 frames. ], batch size: 36, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:48:13,282 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125482.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:14,967 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125484.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:25,863 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3393, 1.5940, 1.7737, 1.9436, 1.7108, 1.8325, 1.8448, 1.8302], + device='cuda:2'), covar=tensor([0.4673, 0.5503, 0.4376, 0.4250, 0.5494, 0.6849, 0.5112, 0.4528], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0372, 0.0324, 0.0339, 0.0347, 0.0392, 0.0356, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:48:40,621 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1787, 2.6925, 2.2919, 2.6127, 1.8818, 2.3633, 2.5557, 1.8149], + device='cuda:2'), covar=tensor([0.2072, 0.1147, 0.0829, 0.1267, 0.3315, 0.1163, 0.1960, 0.2781], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0301, 0.0218, 0.0278, 0.0316, 0.0257, 0.0250, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1445e-04, 1.1931e-04, 8.6120e-05, 1.0957e-04, 1.2799e-04, 1.0171e-04, + 1.0109e-04, 1.0412e-04], device='cuda:2') +2023-04-27 17:48:42,433 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125524.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:48:47,214 INFO [finetune.py:976] (2/7) Epoch 22, batch 5250, loss[loss=0.1902, simple_loss=0.26, pruned_loss=0.06023, over 4786.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.245, pruned_loss=0.0497, over 954974.21 frames. ], batch size: 29, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:49:09,789 INFO [optim.py:369] (2/7) 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 17:49:20,594 INFO [finetune.py:976] (2/7) Epoch 22, batch 5300, loss[loss=0.136, simple_loss=0.2014, pruned_loss=0.03528, over 4344.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2455, pruned_loss=0.04955, over 954005.64 frames. ], batch size: 19, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:49:34,599 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125603.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:49:43,989 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0460, 2.6575, 2.1234, 2.1837, 1.5332, 1.5354, 2.3121, 1.4835], + device='cuda:2'), covar=tensor([0.1614, 0.1413, 0.1354, 0.1694, 0.2167, 0.1884, 0.0875, 0.1947], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0203, 0.0199, 0.0184, 0.0155, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 17:49:50,540 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9230, 2.3751, 1.9911, 1.7718, 1.4241, 1.4445, 2.0681, 1.4136], + device='cuda:2'), covar=tensor([0.1541, 0.1280, 0.1271, 0.1656, 0.2158, 0.1904, 0.0864, 0.1973], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0203, 0.0199, 0.0184, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 17:49:54,072 INFO [finetune.py:976] (2/7) Epoch 22, batch 5350, loss[loss=0.1449, simple_loss=0.2254, pruned_loss=0.0322, over 4764.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2455, pruned_loss=0.04918, over 953738.69 frames. ], batch size: 28, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:49:55,494 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 17:50:15,155 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125664.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:50:16,103 INFO [optim.py:369] (2/7) 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 17:50:26,744 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5372, 1.2522, 4.2629, 4.0342, 3.7482, 3.9433, 3.9240, 3.7190], + device='cuda:2'), covar=tensor([0.7453, 0.5917, 0.1075, 0.1718, 0.1163, 0.1425, 0.1938, 0.1688], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0305, 0.0405, 0.0406, 0.0347, 0.0409, 0.0313, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:50:38,573 INFO [finetune.py:976] (2/7) Epoch 22, batch 5400, loss[loss=0.1877, simple_loss=0.2449, pruned_loss=0.06524, over 4812.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2434, pruned_loss=0.04913, over 955705.04 frames. ], batch size: 51, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:50:39,452 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 17:51:18,751 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6733, 1.7842, 1.5385, 1.1425, 1.2838, 1.2302, 1.5241, 1.2379], + device='cuda:2'), covar=tensor([0.1660, 0.1226, 0.1435, 0.1758, 0.2192, 0.1900, 0.1027, 0.1969], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0210, 0.0168, 0.0203, 0.0199, 0.0185, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 17:51:19,931 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125712.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 17:51:40,597 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 22, batch 5450, loss[loss=0.1671, simple_loss=0.2326, pruned_loss=0.05082, over 4902.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.24, pruned_loss=0.04794, over 956689.09 frames. ], batch size: 32, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:51:45,327 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125732.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:52:15,236 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.84 vs. limit=5.0 +2023-04-27 17:52:28,292 INFO [optim.py:369] (2/7) 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 17:52:47,951 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125777.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:52:51,409 INFO [finetune.py:976] (2/7) Epoch 22, batch 5500, loss[loss=0.1946, simple_loss=0.2599, pruned_loss=0.06462, over 4845.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.238, pruned_loss=0.04744, over 957761.85 frames. ], batch size: 49, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:52:58,404 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:52:59,633 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125786.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:53:19,176 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.8980, 4.7346, 3.1573, 5.5759, 4.8912, 4.8560, 2.1216, 4.7647], + device='cuda:2'), covar=tensor([0.1426, 0.0902, 0.3189, 0.0826, 0.3316, 0.1678, 0.5625, 0.2185], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0216, 0.0251, 0.0304, 0.0296, 0.0246, 0.0274, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 17:53:57,320 INFO [finetune.py:976] (2/7) Epoch 22, batch 5550, loss[loss=0.1641, simple_loss=0.2525, pruned_loss=0.03787, over 4778.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2395, pruned_loss=0.04784, over 958681.31 frames. ], batch size: 29, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:54:03,350 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125832.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:54:41,113 INFO [optim.py:369] (2/7) 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 17:55:02,079 INFO [finetune.py:976] (2/7) Epoch 22, batch 5600, loss[loss=0.164, simple_loss=0.2496, pruned_loss=0.03923, over 4910.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2425, pruned_loss=0.04852, over 958204.42 frames. ], batch size: 37, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:56:05,133 INFO [finetune.py:976] (2/7) Epoch 22, batch 5650, loss[loss=0.1909, simple_loss=0.2587, pruned_loss=0.06151, over 4916.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2464, pruned_loss=0.04979, over 956352.59 frames. ], batch size: 36, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:56:28,696 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1224, 2.2188, 1.4292, 1.8704, 2.2100, 1.9478, 1.9222, 2.0313], + device='cuda:2'), covar=tensor([0.0415, 0.0283, 0.0286, 0.0452, 0.0224, 0.0441, 0.0425, 0.0446], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0028, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 17:56:29,360 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 17:56:31,893 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125959.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:56:35,408 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 22, batch 5700, loss[loss=0.1763, simple_loss=0.2281, pruned_loss=0.0622, over 3936.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2414, pruned_loss=0.04854, over 937572.30 frames. ], batch size: 17, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:56:54,918 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125998.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:57:07,142 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0344, 3.7736, 1.3278, 1.9869, 2.1915, 2.6778, 2.1233, 1.2720], + device='cuda:2'), covar=tensor([0.1248, 0.0873, 0.1887, 0.1309, 0.1058, 0.1023, 0.1717, 0.1834], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0138, 0.0120, 0.0132, 0.0151, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:57:25,280 INFO [finetune.py:976] (2/7) Epoch 23, batch 0, loss[loss=0.1307, simple_loss=0.2124, pruned_loss=0.02455, over 4857.00 frames. ], tot_loss[loss=0.1307, simple_loss=0.2124, pruned_loss=0.02455, over 4857.00 frames. ], batch size: 44, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:57:25,280 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 17:57:31,393 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5761, 3.0170, 0.9977, 1.8966, 2.0038, 2.1997, 1.9999, 1.1026], + device='cuda:2'), covar=tensor([0.1144, 0.1185, 0.1710, 0.1045, 0.0863, 0.0879, 0.1370, 0.1542], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0138, 0.0120, 0.0132, 0.0151, 0.0118, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 17:57:32,408 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8545, 1.6548, 1.8381, 2.1896, 2.1489, 1.7500, 1.5440, 1.9842], + device='cuda:2'), covar=tensor([0.0825, 0.1181, 0.0765, 0.0579, 0.0633, 0.0863, 0.0765, 0.0565], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0202, 0.0184, 0.0175, 0.0177, 0.0181, 0.0152, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 17:57:41,119 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 17:57:48,905 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126012.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 17:58:05,547 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:23,638 INFO [finetune.py:976] (2/7) Epoch 23, batch 50, loss[loss=0.1537, simple_loss=0.2334, pruned_loss=0.03696, over 4816.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2416, pruned_loss=0.04808, over 216322.42 frames. ], batch size: 33, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:58:24,243 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126059.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:24,774 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 17:58:41,161 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126077.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:42,972 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:58:48,843 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126081.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:59:21,046 INFO [finetune.py:976] (2/7) Epoch 23, batch 100, loss[loss=0.1836, simple_loss=0.2494, pruned_loss=0.05892, over 4691.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2396, pruned_loss=0.04859, over 381432.02 frames. ], batch size: 23, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 17:59:34,265 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1628, 2.5108, 1.0379, 1.3935, 1.9291, 1.3274, 3.0221, 1.6222], + device='cuda:2'), covar=tensor([0.0592, 0.0478, 0.0726, 0.1218, 0.0441, 0.0922, 0.0303, 0.0630], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 17:59:42,929 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126125.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 17:59:53,201 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2274, 2.1720, 1.8561, 1.7945, 2.2643, 1.7429, 2.6775, 1.6847], + device='cuda:2'), covar=tensor([0.3332, 0.1896, 0.4707, 0.2941, 0.1613, 0.2432, 0.1301, 0.4036], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0350, 0.0426, 0.0353, 0.0380, 0.0374, 0.0369, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 18:00:28,538 INFO [finetune.py:976] (2/7) Epoch 23, batch 150, loss[loss=0.1583, simple_loss=0.2299, pruned_loss=0.04338, over 4937.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2358, pruned_loss=0.04807, over 508855.61 frames. ], batch size: 38, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 18:00:38,440 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 23, batch 200, loss[loss=0.1479, simple_loss=0.2207, pruned_loss=0.0375, over 4759.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2327, pruned_loss=0.04644, over 607474.42 frames. ], batch size: 28, lr: 3.12e-03, grad_scale: 16.0 +2023-04-27 18:02:08,289 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 18:02:19,986 INFO [finetune.py:976] (2/7) Epoch 23, batch 250, loss[loss=0.1681, simple_loss=0.2445, pruned_loss=0.04581, over 4819.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2375, pruned_loss=0.04816, over 685649.36 frames. ], batch size: 39, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:02:20,584 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 18:02:33,766 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2366, 4.2541, 2.9360, 4.9070, 4.3155, 4.1580, 1.6990, 4.1533], + device='cuda:2'), covar=tensor([0.1327, 0.1039, 0.3627, 0.0922, 0.2452, 0.1550, 0.5769, 0.2246], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0253, 0.0305, 0.0297, 0.0246, 0.0275, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 18:02:51,717 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126307.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:02:53,383 INFO [finetune.py:976] (2/7) Epoch 23, batch 300, loss[loss=0.1685, simple_loss=0.2459, pruned_loss=0.04561, over 4901.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2407, pruned_loss=0.04875, over 745742.70 frames. ], batch size: 37, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:03:09,673 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 18:03:20,000 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7791, 1.6331, 2.0644, 2.1557, 1.5826, 1.4017, 1.7575, 1.0060], + device='cuda:2'), covar=tensor([0.0660, 0.0595, 0.0421, 0.0694, 0.0662, 0.0968, 0.0599, 0.0674], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:03:22,884 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126354.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:03:26,337 INFO [finetune.py:976] (2/7) Epoch 23, batch 350, loss[loss=0.1458, simple_loss=0.2301, pruned_loss=0.03072, over 4766.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.242, pruned_loss=0.04816, over 791646.22 frames. ], batch size: 28, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:03:30,388 INFO [optim.py:369] (2/7) 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 18:03:41,558 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:03:42,117 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126381.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:03:47,064 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2728, 1.7755, 2.1588, 2.6584, 2.1923, 1.6914, 1.5787, 2.0400], + device='cuda:2'), covar=tensor([0.3055, 0.2928, 0.1494, 0.2175, 0.2362, 0.2436, 0.3671, 0.1797], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0245, 0.0228, 0.0315, 0.0220, 0.0234, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 18:03:54,031 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-04-27 18:03:59,739 INFO [finetune.py:976] (2/7) Epoch 23, batch 400, loss[loss=0.1466, simple_loss=0.2327, pruned_loss=0.03021, over 4849.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2426, pruned_loss=0.04791, over 827857.44 frames. ], batch size: 44, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:04:30,796 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126429.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:04:38,558 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5782, 1.1085, 1.3277, 1.3071, 1.6617, 1.3318, 1.1426, 1.2881], + device='cuda:2'), covar=tensor([0.1779, 0.1562, 0.1944, 0.1355, 0.0916, 0.1812, 0.2111, 0.2499], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0311, 0.0352, 0.0286, 0.0325, 0.0307, 0.0298, 0.0371], + device='cuda:2'), out_proj_covar=tensor([6.4264e-05, 6.4308e-05, 7.4199e-05, 5.7521e-05, 6.6940e-05, 6.4426e-05, + 6.2163e-05, 7.8663e-05], device='cuda:2') +2023-04-27 18:04:42,850 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126441.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:04:46,494 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.5576, 4.5505, 2.9630, 5.2086, 4.6573, 4.4482, 1.9516, 4.4035], + device='cuda:2'), covar=tensor([0.1534, 0.0788, 0.3325, 0.0833, 0.2431, 0.1565, 0.5772, 0.2015], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0217, 0.0253, 0.0305, 0.0297, 0.0246, 0.0275, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 18:04:48,386 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126450.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:04:54,267 INFO [finetune.py:976] (2/7) Epoch 23, batch 450, loss[loss=0.1998, simple_loss=0.2599, pruned_loss=0.06991, over 4834.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2422, pruned_loss=0.04804, over 857111.40 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:04:58,394 INFO [optim.py:369] (2/7) 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 18:05:27,628 INFO [finetune.py:976] (2/7) Epoch 23, batch 500, loss[loss=0.151, simple_loss=0.2167, pruned_loss=0.04265, over 4826.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2389, pruned_loss=0.047, over 879725.86 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:05:29,434 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126511.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:06:06,864 INFO [finetune.py:976] (2/7) Epoch 23, batch 550, loss[loss=0.1999, simple_loss=0.2624, pruned_loss=0.06868, over 4840.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2383, pruned_loss=0.04717, over 898657.56 frames. ], batch size: 47, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:06:16,305 INFO [optim.py:369] (2/7) 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 18:06:59,481 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 18:07:12,070 INFO [finetune.py:976] (2/7) Epoch 23, batch 600, loss[loss=0.1932, simple_loss=0.2742, pruned_loss=0.05609, over 4808.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2393, pruned_loss=0.04787, over 912817.36 frames. ], batch size: 51, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:07:33,586 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 18:07:55,176 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.49 vs. limit=5.0 +2023-04-27 18:08:14,766 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126654.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:08:16,080 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1692, 1.6538, 2.0795, 2.4931, 2.0656, 1.6231, 1.4799, 1.9363], + device='cuda:2'), covar=tensor([0.3229, 0.3183, 0.1655, 0.2299, 0.2537, 0.2512, 0.3982, 0.1983], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0244, 0.0227, 0.0314, 0.0219, 0.0233, 0.0227, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 18:08:17,790 INFO [finetune.py:976] (2/7) Epoch 23, batch 650, loss[loss=0.1825, simple_loss=0.2672, pruned_loss=0.0489, over 4819.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2442, pruned_loss=0.04938, over 923810.35 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:08:24,434 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 18:08:26,271 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 18:08:26,640 INFO [optim.py:369] (2/7) 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 18:09:20,361 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126702.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:09:24,632 INFO [finetune.py:976] (2/7) Epoch 23, batch 700, loss[loss=0.176, simple_loss=0.2539, pruned_loss=0.04908, over 4909.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2446, pruned_loss=0.04919, over 930967.69 frames. ], batch size: 36, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:10:03,670 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126736.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:10:19,571 INFO [finetune.py:976] (2/7) Epoch 23, batch 750, loss[loss=0.1483, simple_loss=0.2325, pruned_loss=0.03209, over 4775.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2462, pruned_loss=0.05014, over 937122.55 frames. ], batch size: 29, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:10:23,175 INFO [optim.py:369] (2/7) 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 18:10:34,011 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 18:10:35,047 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0621, 4.2173, 0.6590, 2.2785, 2.3250, 2.6877, 2.3921, 0.9065], + device='cuda:2'), covar=tensor([0.1174, 0.0805, 0.2194, 0.1170, 0.0998, 0.1096, 0.1416, 0.2233], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0237, 0.0136, 0.0118, 0.0131, 0.0150, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:10:47,355 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0544, 4.2259, 0.8095, 2.2127, 2.3834, 2.6111, 2.4370, 0.9091], + device='cuda:2'), covar=tensor([0.1198, 0.0797, 0.2080, 0.1148, 0.1001, 0.1148, 0.1404, 0.2153], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0238, 0.0136, 0.0119, 0.0131, 0.0150, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:10:51,755 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 800, loss[loss=0.1592, simple_loss=0.2234, pruned_loss=0.04745, over 3775.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2466, pruned_loss=0.05088, over 937999.13 frames. ], batch size: 16, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:10:57,927 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126816.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:11:27,334 INFO [finetune.py:976] (2/7) Epoch 23, batch 850, loss[loss=0.1463, simple_loss=0.2054, pruned_loss=0.04364, over 4552.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2442, pruned_loss=0.0496, over 941036.55 frames. ], batch size: 20, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:11:30,954 INFO [optim.py:369] (2/7) 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 18:11:44,258 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126877.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:12:28,955 INFO [finetune.py:976] (2/7) Epoch 23, batch 900, loss[loss=0.1498, simple_loss=0.2193, pruned_loss=0.04017, over 4746.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2402, pruned_loss=0.04823, over 944827.33 frames. ], batch size: 27, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:13:36,983 INFO [finetune.py:976] (2/7) Epoch 23, batch 950, loss[loss=0.186, simple_loss=0.2384, pruned_loss=0.06674, over 4890.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2395, pruned_loss=0.04862, over 947507.47 frames. ], batch size: 43, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:13:40,660 INFO [optim.py:369] (2/7) 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 18:14:11,166 INFO [finetune.py:976] (2/7) Epoch 23, batch 1000, loss[loss=0.1517, simple_loss=0.2149, pruned_loss=0.04427, over 4789.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.242, pruned_loss=0.04944, over 949735.28 frames. ], batch size: 26, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:14:28,629 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127036.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:14:35,852 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8583, 2.2548, 1.8793, 1.6715, 1.4047, 1.4209, 1.8612, 1.3224], + device='cuda:2'), covar=tensor([0.1689, 0.1347, 0.1460, 0.1752, 0.2321, 0.1921, 0.1005, 0.2038], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0210, 0.0169, 0.0203, 0.0199, 0.0185, 0.0155, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 18:14:45,075 INFO [finetune.py:976] (2/7) Epoch 23, batch 1050, loss[loss=0.1769, simple_loss=0.2558, pruned_loss=0.049, over 4905.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.244, pruned_loss=0.04965, over 950190.64 frames. ], batch size: 36, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:14:48,720 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127084.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:15:16,228 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127106.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:15:18,463 INFO [finetune.py:976] (2/7) Epoch 23, batch 1100, loss[loss=0.1787, simple_loss=0.2465, pruned_loss=0.05538, over 4898.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.245, pruned_loss=0.05013, over 950062.62 frames. ], batch size: 36, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:15:42,016 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-04-27 18:15:47,874 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 1150, loss[loss=0.1535, simple_loss=0.2284, pruned_loss=0.03926, over 4856.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2461, pruned_loss=0.05044, over 949939.77 frames. ], batch size: 31, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:15:56,439 INFO [optim.py:369] (2/7) 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 18:15:58,448 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.08 vs. limit=5.0 +2023-04-27 18:16:00,784 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127172.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:16:25,336 INFO [finetune.py:976] (2/7) Epoch 23, batch 1200, loss[loss=0.1543, simple_loss=0.219, pruned_loss=0.04484, over 4264.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2446, pruned_loss=0.04988, over 951011.07 frames. ], batch size: 65, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:16:31,793 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6197, 1.2941, 1.3627, 1.4730, 1.8305, 1.5184, 1.2044, 1.2641], + device='cuda:2'), covar=tensor([0.1572, 0.1388, 0.1698, 0.1201, 0.0806, 0.1373, 0.2014, 0.2254], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0311, 0.0352, 0.0288, 0.0327, 0.0309, 0.0299, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.4577e-05, 6.4366e-05, 7.4265e-05, 5.8052e-05, 6.7436e-05, 6.4867e-05, + 6.2444e-05, 7.8954e-05], device='cuda:2') +2023-04-27 18:16:52,069 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127249.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:17:01,969 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1463, 2.6253, 1.0259, 1.3871, 2.1480, 1.1695, 3.5143, 1.6356], + device='cuda:2'), covar=tensor([0.0684, 0.0687, 0.0838, 0.1373, 0.0531, 0.1145, 0.0289, 0.0744], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0048, 0.0046, 0.0050, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 18:17:03,695 INFO [finetune.py:976] (2/7) Epoch 23, batch 1250, loss[loss=0.1429, simple_loss=0.2244, pruned_loss=0.0307, over 4779.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2427, pruned_loss=0.04937, over 953619.31 frames. ], batch size: 29, lr: 3.11e-03, grad_scale: 16.0 +2023-04-27 18:17:14,002 INFO [optim.py:369] (2/7) 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 18:18:08,734 INFO [finetune.py:976] (2/7) Epoch 23, batch 1300, loss[loss=0.1189, simple_loss=0.1961, pruned_loss=0.02083, over 4807.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2398, pruned_loss=0.04819, over 955479.47 frames. ], batch size: 25, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:18:09,520 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127310.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:18:10,749 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8231, 2.0239, 1.0560, 1.5564, 2.3550, 1.6593, 1.6886, 1.8065], + device='cuda:2'), covar=tensor([0.0480, 0.0347, 0.0296, 0.0538, 0.0219, 0.0485, 0.0486, 0.0527], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 18:18:49,596 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0305, 2.3784, 1.2029, 1.4343, 2.0027, 1.3109, 2.7746, 1.5325], + device='cuda:2'), covar=tensor([0.0599, 0.0660, 0.0682, 0.1008, 0.0367, 0.0845, 0.0239, 0.0602], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0065, 0.0048, 0.0046, 0.0050, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 18:18:49,605 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3088, 1.5747, 1.4988, 1.7537, 1.7368, 1.7112, 1.5059, 2.8170], + device='cuda:2'), covar=tensor([0.0584, 0.0678, 0.0680, 0.0999, 0.0518, 0.0652, 0.0646, 0.0232], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0037, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 18:19:13,548 INFO [finetune.py:976] (2/7) Epoch 23, batch 1350, loss[loss=0.1906, simple_loss=0.2507, pruned_loss=0.06523, over 4816.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2386, pruned_loss=0.04813, over 952860.72 frames. ], batch size: 39, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:19:23,391 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 23, batch 1400, loss[loss=0.1684, simple_loss=0.2309, pruned_loss=0.05298, over 4766.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2432, pruned_loss=0.04978, over 952984.34 frames. ], batch size: 26, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:20:43,129 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127427.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 18:21:02,761 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4400, 1.7301, 1.6905, 1.8811, 1.9277, 1.9947, 1.6052, 3.7672], + device='cuda:2'), covar=tensor([0.0568, 0.0776, 0.0726, 0.1188, 0.0580, 0.0516, 0.0724, 0.0115], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 18:21:15,606 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6648, 2.1863, 1.6659, 1.5569, 1.2321, 1.2369, 1.7092, 1.1937], + device='cuda:2'), covar=tensor([0.1523, 0.1156, 0.1344, 0.1552, 0.2208, 0.1847, 0.0920, 0.1947], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0209, 0.0167, 0.0202, 0.0198, 0.0184, 0.0154, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 18:21:24,257 INFO [finetune.py:976] (2/7) Epoch 23, batch 1450, loss[loss=0.1701, simple_loss=0.2543, pruned_loss=0.04294, over 4856.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2446, pruned_loss=0.04953, over 953114.34 frames. ], batch size: 44, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:21:34,074 INFO [optim.py:369] (2/7) 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 18:21:43,896 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127472.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:21:44,020 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-27 18:22:06,024 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127488.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 18:22:29,896 INFO [finetune.py:976] (2/7) Epoch 23, batch 1500, loss[loss=0.1646, simple_loss=0.2465, pruned_loss=0.04136, over 4877.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2461, pruned_loss=0.05, over 951996.47 frames. ], batch size: 32, lr: 3.11e-03, grad_scale: 32.0 +2023-04-27 18:22:37,633 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127520.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:22:43,187 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 1550, loss[loss=0.1331, simple_loss=0.2141, pruned_loss=0.02605, over 4791.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2458, pruned_loss=0.04958, over 951235.62 frames. ], batch size: 29, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:23:23,839 INFO [optim.py:369] (2/7) 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 18:23:33,690 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2294, 3.9705, 2.8399, 4.8627, 4.0896, 4.1722, 1.5377, 4.2195], + device='cuda:2'), covar=tensor([0.1707, 0.1412, 0.4089, 0.1032, 0.3026, 0.1712, 0.5960, 0.2104], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0252, 0.0304, 0.0297, 0.0245, 0.0273, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 18:23:41,774 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127590.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:23:50,949 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 1600, loss[loss=0.1192, simple_loss=0.2043, pruned_loss=0.01708, over 4722.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.243, pruned_loss=0.04879, over 950483.68 frames. ], batch size: 59, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:24:20,508 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 18:24:26,892 INFO [finetune.py:976] (2/7) Epoch 23, batch 1650, loss[loss=0.1522, simple_loss=0.217, pruned_loss=0.04368, over 4814.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2406, pruned_loss=0.04816, over 953520.65 frames. ], batch size: 41, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:24:28,866 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127662.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:24:31,042 INFO [optim.py:369] (2/7) 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 18:25:00,252 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8959, 2.8197, 2.1891, 3.2834, 2.8863, 2.8804, 1.1488, 2.8838], + device='cuda:2'), covar=tensor([0.2243, 0.1978, 0.3658, 0.3273, 0.4359, 0.2350, 0.5948, 0.2837], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0253, 0.0305, 0.0297, 0.0245, 0.0274, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 18:25:00,767 INFO [finetune.py:976] (2/7) Epoch 23, batch 1700, loss[loss=0.2459, simple_loss=0.2874, pruned_loss=0.1022, over 4929.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2395, pruned_loss=0.04827, over 955292.54 frames. ], batch size: 38, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:25:10,337 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127723.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:25:34,586 INFO [finetune.py:976] (2/7) Epoch 23, batch 1750, loss[loss=0.164, simple_loss=0.2492, pruned_loss=0.03936, over 4812.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2401, pruned_loss=0.04867, over 952985.00 frames. ], batch size: 45, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:25:38,229 INFO [optim.py:369] (2/7) 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 18:25:40,842 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6778, 2.1867, 1.6630, 1.6623, 1.3869, 1.3251, 1.7158, 1.2975], + device='cuda:2'), covar=tensor([0.1379, 0.1159, 0.1294, 0.1542, 0.2006, 0.1668, 0.0848, 0.1850], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0209, 0.0167, 0.0202, 0.0197, 0.0184, 0.0154, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 18:25:46,631 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3052, 1.3878, 3.5131, 3.2437, 3.1602, 3.3057, 3.4061, 3.1101], + device='cuda:2'), covar=tensor([0.6879, 0.4992, 0.1361, 0.2216, 0.1283, 0.2198, 0.1241, 0.1781], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0307, 0.0405, 0.0408, 0.0348, 0.0410, 0.0314, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 18:25:48,412 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127779.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:25:50,843 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127783.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 18:25:56,103 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4672, 1.2834, 1.5938, 1.6299, 1.3595, 1.1902, 1.2432, 0.6949], + device='cuda:2'), covar=tensor([0.0468, 0.0581, 0.0378, 0.0424, 0.0615, 0.1359, 0.0534, 0.0698], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0096, 0.0073, 0.0065], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:26:14,168 INFO [finetune.py:976] (2/7) Epoch 23, batch 1800, loss[loss=0.2291, simple_loss=0.287, pruned_loss=0.08561, over 4864.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2443, pruned_loss=0.04998, over 952080.07 frames. ], batch size: 31, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:26:57,481 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127840.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:27:09,546 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 1850, loss[loss=0.198, simple_loss=0.2759, pruned_loss=0.06007, over 4728.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.245, pruned_loss=0.04985, over 949588.69 frames. ], batch size: 59, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:27:31,375 INFO [optim.py:369] (2/7) 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 18:27:39,376 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6983, 1.0442, 1.7204, 2.1239, 1.7313, 1.6004, 1.6942, 1.6478], + device='cuda:2'), covar=tensor([0.4408, 0.6225, 0.5819, 0.5294, 0.5636, 0.7449, 0.7072, 0.8728], + device='cuda:2'), in_proj_covar=tensor([0.0432, 0.0416, 0.0507, 0.0504, 0.0462, 0.0492, 0.0496, 0.0507], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 18:27:40,526 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127872.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:27:49,852 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127885.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:27:54,234 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.54 vs. limit=5.0 +2023-04-27 18:28:03,999 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127905.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:28:06,378 INFO [finetune.py:976] (2/7) Epoch 23, batch 1900, loss[loss=0.1925, simple_loss=0.2572, pruned_loss=0.06392, over 4802.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2458, pruned_loss=0.04946, over 952359.48 frames. ], batch size: 51, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:28:07,697 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127911.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:28:32,845 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127933.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:28:42,784 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6818, 4.2101, 0.9585, 2.2224, 2.2992, 2.9316, 2.4079, 1.0419], + device='cuda:2'), covar=tensor([0.1487, 0.1327, 0.2241, 0.1310, 0.1087, 0.1086, 0.1552, 0.2020], + device='cuda:2'), in_proj_covar=tensor([0.0115, 0.0235, 0.0135, 0.0118, 0.0130, 0.0149, 0.0115, 0.0116], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:28:57,590 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 1950, loss[loss=0.1659, simple_loss=0.2336, pruned_loss=0.04908, over 4882.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2444, pruned_loss=0.0488, over 955817.69 frames. ], batch size: 35, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:29:07,713 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2465, 1.5691, 1.4766, 1.7636, 1.7537, 1.9204, 1.4157, 3.6338], + device='cuda:2'), covar=tensor([0.0585, 0.0758, 0.0803, 0.1165, 0.0617, 0.0496, 0.0717, 0.0136], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 18:29:16,665 INFO [optim.py:369] (2/7) 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 18:30:02,324 INFO [finetune.py:976] (2/7) Epoch 23, batch 2000, loss[loss=0.1585, simple_loss=0.236, pruned_loss=0.04053, over 4750.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2411, pruned_loss=0.04769, over 955903.75 frames. ], batch size: 27, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:30:07,967 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128018.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:30:23,743 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1458, 2.6160, 2.1614, 2.0893, 1.6579, 1.6991, 2.2551, 1.6460], + device='cuda:2'), covar=tensor([0.1645, 0.1547, 0.1500, 0.1716, 0.2356, 0.2014, 0.0957, 0.1989], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0211, 0.0169, 0.0204, 0.0200, 0.0186, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 18:30:35,019 INFO [finetune.py:976] (2/7) Epoch 23, batch 2050, loss[loss=0.1587, simple_loss=0.2201, pruned_loss=0.04866, over 4732.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2378, pruned_loss=0.04705, over 955333.92 frames. ], batch size: 59, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:30:37,337 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2621, 1.5331, 1.3274, 1.4776, 1.3419, 1.2807, 1.2974, 1.0119], + device='cuda:2'), covar=tensor([0.1701, 0.1337, 0.0916, 0.1278, 0.3537, 0.1242, 0.1635, 0.2195], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0298, 0.0212, 0.0274, 0.0311, 0.0254, 0.0246, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1302e-04, 1.1791e-04, 8.3710e-05, 1.0827e-04, 1.2560e-04, 1.0055e-04, + 9.9335e-05, 1.0285e-04], device='cuda:2') +2023-04-27 18:30:39,640 INFO [optim.py:369] (2/7) 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 18:30:50,718 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128083.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 18:31:08,824 INFO [finetune.py:976] (2/7) Epoch 23, batch 2100, loss[loss=0.2781, simple_loss=0.331, pruned_loss=0.1126, over 4213.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2379, pruned_loss=0.04768, over 953131.28 frames. ], batch size: 66, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:31:22,861 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128131.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 18:31:25,337 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128135.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:31:42,235 INFO [finetune.py:976] (2/7) Epoch 23, batch 2150, loss[loss=0.1551, simple_loss=0.2257, pruned_loss=0.04221, over 4422.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2404, pruned_loss=0.04844, over 952769.74 frames. ], batch size: 19, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:31:46,836 INFO [optim.py:369] (2/7) 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 18:31:59,140 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:32:15,341 INFO [finetune.py:976] (2/7) Epoch 23, batch 2200, loss[loss=0.1971, simple_loss=0.2804, pruned_loss=0.05693, over 4823.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2432, pruned_loss=0.04923, over 951228.31 frames. ], batch size: 38, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:32:39,402 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128228.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:32:40,691 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8083, 2.2098, 2.1273, 2.2843, 2.0642, 2.1128, 2.1313, 2.0659], + device='cuda:2'), covar=tensor([0.4269, 0.5671, 0.4775, 0.4454, 0.5615, 0.6983, 0.5940, 0.5601], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0371, 0.0323, 0.0336, 0.0345, 0.0392, 0.0356, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 18:32:42,391 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128233.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:33:12,837 INFO [finetune.py:976] (2/7) Epoch 23, batch 2250, loss[loss=0.1805, simple_loss=0.2461, pruned_loss=0.05744, over 4832.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2428, pruned_loss=0.04908, over 949527.71 frames. ], batch size: 30, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:33:22,489 INFO [optim.py:369] (2/7) 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 18:33:57,142 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0850, 2.2517, 1.3202, 1.8236, 2.5419, 1.9358, 1.8685, 2.0669], + device='cuda:2'), covar=tensor([0.0463, 0.0323, 0.0251, 0.0499, 0.0198, 0.0456, 0.0435, 0.0498], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 18:34:26,340 INFO [finetune.py:976] (2/7) Epoch 23, batch 2300, loss[loss=0.1842, simple_loss=0.2663, pruned_loss=0.05105, over 4897.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.243, pruned_loss=0.04877, over 952250.19 frames. ], batch size: 36, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:34:37,303 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128318.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:34:37,949 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8935, 1.7162, 2.1546, 2.3296, 1.6917, 1.5063, 1.8431, 0.9391], + device='cuda:2'), covar=tensor([0.0545, 0.0676, 0.0465, 0.0646, 0.0877, 0.1080, 0.0611, 0.0761], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0067, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:34:39,030 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:35:34,264 INFO [finetune.py:976] (2/7) Epoch 23, batch 2350, loss[loss=0.1593, simple_loss=0.2351, pruned_loss=0.04171, over 4895.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2407, pruned_loss=0.04759, over 951463.91 frames. ], batch size: 35, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:35:37,975 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128366.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:36:01,145 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 2400, loss[loss=0.1636, simple_loss=0.2386, pruned_loss=0.04433, over 4807.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2382, pruned_loss=0.0467, over 953956.13 frames. ], batch size: 25, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:36:34,508 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128413.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:36:58,489 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1990, 3.2031, 0.9040, 1.5327, 1.7422, 2.2593, 1.7517, 1.0439], + device='cuda:2'), covar=tensor([0.1968, 0.1823, 0.2597, 0.1901, 0.1363, 0.1431, 0.1944, 0.2148], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0238, 0.0137, 0.0119, 0.0131, 0.0150, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:37:01,391 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128435.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:37:29,230 INFO [finetune.py:976] (2/7) Epoch 23, batch 2450, loss[loss=0.1416, simple_loss=0.2167, pruned_loss=0.03326, over 4915.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2359, pruned_loss=0.04647, over 956207.20 frames. ], batch size: 37, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:37:35,176 INFO [optim.py:369] (2/7) 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 18:37:46,739 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128474.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:37:56,936 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128483.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:38:27,888 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128506.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:38:29,612 INFO [finetune.py:976] (2/7) Epoch 23, batch 2500, loss[loss=0.2359, simple_loss=0.3016, pruned_loss=0.08512, over 4278.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.238, pruned_loss=0.04747, over 955909.66 frames. ], batch size: 65, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:38:53,791 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128528.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:39:29,767 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128554.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:39:32,693 INFO [finetune.py:976] (2/7) Epoch 23, batch 2550, loss[loss=0.172, simple_loss=0.236, pruned_loss=0.05402, over 4753.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2395, pruned_loss=0.04769, over 954770.09 frames. ], batch size: 23, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:39:42,963 INFO [optim.py:369] (2/7) 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 18:39:55,796 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128576.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:40:40,367 INFO [finetune.py:976] (2/7) Epoch 23, batch 2600, loss[loss=0.2226, simple_loss=0.288, pruned_loss=0.07859, over 4899.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2415, pruned_loss=0.0479, over 954055.74 frames. ], batch size: 36, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:41:01,967 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 18:41:16,254 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3923, 1.9177, 2.2027, 2.8819, 2.2879, 1.7890, 1.9852, 2.1317], + device='cuda:2'), covar=tensor([0.3082, 0.3088, 0.1639, 0.2316, 0.2625, 0.2647, 0.3271, 0.2163], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0245, 0.0227, 0.0314, 0.0219, 0.0234, 0.0227, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 18:41:44,293 INFO [finetune.py:976] (2/7) Epoch 23, batch 2650, loss[loss=0.1909, simple_loss=0.2647, pruned_loss=0.05854, over 4810.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2436, pruned_loss=0.04839, over 955363.69 frames. ], batch size: 40, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:41:53,485 INFO [optim.py:369] (2/7) 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 18:41:54,212 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6293, 3.1649, 2.6471, 3.0911, 2.3498, 2.7278, 2.7936, 2.2907], + device='cuda:2'), covar=tensor([0.1862, 0.1201, 0.0678, 0.1068, 0.2864, 0.1096, 0.1656, 0.2353], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0299, 0.0213, 0.0274, 0.0313, 0.0254, 0.0247, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1309e-04, 1.1826e-04, 8.4090e-05, 1.0823e-04, 1.2639e-04, 1.0057e-04, + 9.9559e-05, 1.0314e-04], device='cuda:2') +2023-04-27 18:42:11,811 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128676.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:42:55,229 INFO [finetune.py:976] (2/7) Epoch 23, batch 2700, loss[loss=0.1908, simple_loss=0.2622, pruned_loss=0.05969, over 4729.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2435, pruned_loss=0.04801, over 952437.93 frames. ], batch size: 54, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:42:55,355 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6981, 1.7514, 1.8604, 1.4944, 1.8731, 1.6019, 2.3771, 1.5895], + device='cuda:2'), covar=tensor([0.3440, 0.1759, 0.4921, 0.2416, 0.1335, 0.2109, 0.1326, 0.4316], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0348, 0.0427, 0.0352, 0.0378, 0.0376, 0.0369, 0.0419], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 18:43:30,532 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-27 18:43:30,954 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1508, 2.5936, 1.0648, 1.4721, 1.9656, 1.2913, 3.5404, 1.8410], + device='cuda:2'), covar=tensor([0.0694, 0.0672, 0.0834, 0.1303, 0.0564, 0.1057, 0.0304, 0.0670], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0050, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 18:44:00,967 INFO [finetune.py:976] (2/7) Epoch 23, batch 2750, loss[loss=0.1397, simple_loss=0.2145, pruned_loss=0.03244, over 4800.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.241, pruned_loss=0.04743, over 952417.08 frames. ], batch size: 29, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:44:04,652 INFO [optim.py:369] (2/7) 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 18:44:12,834 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128769.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:44:24,733 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8274, 3.4260, 0.9590, 1.9491, 2.0601, 2.4515, 1.9447, 0.9586], + device='cuda:2'), covar=tensor([0.1156, 0.0816, 0.2052, 0.1172, 0.0953, 0.0994, 0.1504, 0.1939], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0239, 0.0137, 0.0119, 0.0131, 0.0151, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 18:44:32,560 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 2800, loss[loss=0.1589, simple_loss=0.2428, pruned_loss=0.03747, over 4745.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2392, pruned_loss=0.0468, over 955466.56 frames. ], batch size: 23, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:45:19,273 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0752, 1.2366, 4.9978, 4.3483, 4.3441, 4.6962, 4.3146, 4.1457], + device='cuda:2'), covar=tensor([0.9065, 0.9066, 0.1222, 0.2956, 0.1789, 0.2821, 0.2484, 0.3029], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0308, 0.0408, 0.0412, 0.0352, 0.0414, 0.0317, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 18:45:55,122 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 2850, loss[loss=0.1301, simple_loss=0.1947, pruned_loss=0.03274, over 4357.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2365, pruned_loss=0.04594, over 952651.64 frames. ], batch size: 19, lr: 3.10e-03, grad_scale: 32.0 +2023-04-27 18:46:18,652 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-27 18:46:22,646 INFO [optim.py:369] (2/7) 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 18:46:50,196 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128890.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 18:47:11,862 INFO [finetune.py:976] (2/7) Epoch 23, batch 2900, loss[loss=0.1845, simple_loss=0.2645, pruned_loss=0.05222, over 4904.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2393, pruned_loss=0.04703, over 952694.94 frames. ], batch size: 35, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:47:28,316 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5181, 1.8494, 1.9056, 2.0321, 1.8727, 1.8970, 1.9631, 1.9244], + device='cuda:2'), covar=tensor([0.3927, 0.5310, 0.4427, 0.4155, 0.5175, 0.6941, 0.5222, 0.5016], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0372, 0.0324, 0.0336, 0.0346, 0.0392, 0.0355, 0.0328], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 18:47:56,522 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128951.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 18:48:04,747 INFO [finetune.py:976] (2/7) Epoch 23, batch 2950, loss[loss=0.2076, simple_loss=0.2746, pruned_loss=0.07033, over 4818.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2416, pruned_loss=0.0475, over 952668.94 frames. ], batch size: 40, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:48:10,254 INFO [optim.py:369] (2/7) 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 18:48:22,811 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128976.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:49:01,104 INFO [finetune.py:976] (2/7) Epoch 23, batch 3000, loss[loss=0.139, simple_loss=0.2163, pruned_loss=0.03088, over 4826.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2441, pruned_loss=0.04876, over 953282.04 frames. ], batch size: 47, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:49:01,104 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 18:49:17,638 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 18:49:31,984 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129024.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:50:14,480 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9322, 1.7045, 4.0440, 3.7950, 3.5046, 3.7699, 3.6701, 3.5787], + device='cuda:2'), covar=tensor([0.5925, 0.4636, 0.0987, 0.1524, 0.0998, 0.1560, 0.2734, 0.1393], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0306, 0.0406, 0.0410, 0.0349, 0.0411, 0.0315, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 18:50:16,855 INFO [finetune.py:976] (2/7) Epoch 23, batch 3050, loss[loss=0.1725, simple_loss=0.2398, pruned_loss=0.05264, over 4824.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2451, pruned_loss=0.04909, over 954328.87 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 32.0 +2023-04-27 18:50:25,141 INFO [optim.py:369] (2/7) 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 18:50:25,902 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7822, 1.3516, 1.8525, 2.2854, 1.8428, 1.7589, 1.8481, 1.7024], + device='cuda:2'), covar=tensor([0.4467, 0.6751, 0.6366, 0.5515, 0.5750, 0.7782, 0.7640, 0.8774], + device='cuda:2'), in_proj_covar=tensor([0.0435, 0.0419, 0.0512, 0.0508, 0.0465, 0.0496, 0.0499, 0.0511], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 18:50:34,466 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129069.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:51:09,071 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1237, 2.6392, 2.1230, 2.6289, 1.9365, 2.2402, 2.5306, 1.9200], + device='cuda:2'), covar=tensor([0.2545, 0.1663, 0.1227, 0.1613, 0.3376, 0.1584, 0.1998, 0.2700], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0298, 0.0212, 0.0273, 0.0311, 0.0254, 0.0246, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1251e-04, 1.1790e-04, 8.3585e-05, 1.0757e-04, 1.2583e-04, 1.0023e-04, + 9.9210e-05, 1.0289e-04], device='cuda:2') +2023-04-27 18:51:30,728 INFO [finetune.py:976] (2/7) Epoch 23, batch 3100, loss[loss=0.177, simple_loss=0.2474, pruned_loss=0.05329, over 4747.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2429, pruned_loss=0.04818, over 955172.42 frames. ], batch size: 54, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:51:41,685 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129117.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:52:07,894 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129137.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 18:52:39,819 INFO [finetune.py:976] (2/7) Epoch 23, batch 3150, loss[loss=0.1599, simple_loss=0.2439, pruned_loss=0.03798, over 4917.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.24, pruned_loss=0.04768, over 955532.34 frames. ], batch size: 37, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:52:50,085 INFO [optim.py:369] (2/7) 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 18:53:34,417 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2706, 2.6381, 1.0228, 1.5387, 2.0126, 1.2901, 3.6274, 1.9579], + device='cuda:2'), covar=tensor([0.0649, 0.0658, 0.0837, 0.1228, 0.0543, 0.1055, 0.0223, 0.0606], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 18:53:46,920 INFO [finetune.py:976] (2/7) Epoch 23, batch 3200, loss[loss=0.1464, simple_loss=0.2249, pruned_loss=0.03392, over 4828.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2364, pruned_loss=0.04624, over 955486.69 frames. ], batch size: 33, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:54:30,623 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 18:54:35,414 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2796, 1.4769, 1.2958, 1.4295, 1.3023, 1.2448, 1.2091, 1.0695], + device='cuda:2'), covar=tensor([0.1169, 0.0990, 0.0738, 0.0939, 0.2529, 0.0918, 0.1300, 0.1521], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0299, 0.0213, 0.0274, 0.0312, 0.0255, 0.0247, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1297e-04, 1.1842e-04, 8.4026e-05, 1.0813e-04, 1.2593e-04, 1.0073e-04, + 9.9574e-05, 1.0333e-04], device='cuda:2') +2023-04-27 18:54:42,099 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129246.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 18:54:55,918 INFO [finetune.py:976] (2/7) Epoch 23, batch 3250, loss[loss=0.2291, simple_loss=0.2962, pruned_loss=0.08107, over 4810.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2362, pruned_loss=0.04616, over 954016.48 frames. ], batch size: 41, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:55:06,949 INFO [optim.py:369] (2/7) 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 18:55:51,327 INFO [finetune.py:976] (2/7) Epoch 23, batch 3300, loss[loss=0.1881, simple_loss=0.2553, pruned_loss=0.06046, over 4714.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.241, pruned_loss=0.04825, over 956771.61 frames. ], batch size: 59, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:56:48,060 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0604, 2.1213, 1.7633, 1.7301, 2.1811, 1.7686, 2.7087, 1.5819], + device='cuda:2'), covar=tensor([0.3628, 0.1892, 0.4892, 0.2820, 0.1577, 0.2443, 0.1307, 0.4161], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0346, 0.0424, 0.0347, 0.0374, 0.0371, 0.0364, 0.0415], + device='cuda:2'), out_proj_covar=tensor([9.9352e-05, 1.0344e-04, 1.2843e-04, 1.0443e-04, 1.1110e-04, 1.1061e-04, + 1.0692e-04, 1.2496e-04], device='cuda:2') +2023-04-27 18:56:51,618 INFO [finetune.py:976] (2/7) Epoch 23, batch 3350, loss[loss=0.1566, simple_loss=0.2402, pruned_loss=0.03654, over 4820.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2434, pruned_loss=0.04894, over 957637.72 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:57:01,578 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1807, 1.4112, 1.3517, 1.7794, 1.4864, 1.6295, 1.3040, 3.0427], + device='cuda:2'), covar=tensor([0.0706, 0.1059, 0.1016, 0.1300, 0.0830, 0.0612, 0.1024, 0.0261], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 18:57:02,542 INFO [optim.py:369] (2/7) 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 18:57:34,702 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9638, 3.9626, 2.7652, 4.5599, 4.0080, 3.9217, 1.7314, 3.9295], + device='cuda:2'), covar=tensor([0.1576, 0.1121, 0.2930, 0.1271, 0.2969, 0.1616, 0.5455, 0.2295], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0217, 0.0252, 0.0304, 0.0295, 0.0245, 0.0272, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 18:57:58,598 INFO [finetune.py:976] (2/7) Epoch 23, batch 3400, loss[loss=0.2429, simple_loss=0.3197, pruned_loss=0.0831, over 4173.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2463, pruned_loss=0.05033, over 957940.64 frames. ], batch size: 65, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:58:06,877 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129413.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 18:58:08,701 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 18:58:40,741 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 3450, loss[loss=0.2174, simple_loss=0.2809, pruned_loss=0.07689, over 4829.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2474, pruned_loss=0.05098, over 957249.41 frames. ], batch size: 47, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 18:59:15,876 INFO [optim.py:369] (2/7) 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 18:59:26,975 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129474.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 18:59:39,675 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129485.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:00:05,986 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 19:00:07,510 INFO [finetune.py:976] (2/7) Epoch 23, batch 3500, loss[loss=0.1614, simple_loss=0.2357, pruned_loss=0.04358, over 4873.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2443, pruned_loss=0.05022, over 957103.08 frames. ], batch size: 34, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:00:40,766 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3142, 1.6851, 1.8087, 1.9487, 1.8057, 1.9413, 1.8617, 1.8775], + device='cuda:2'), covar=tensor([0.3476, 0.4255, 0.3869, 0.3590, 0.4959, 0.6195, 0.4313, 0.4025], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0377, 0.0329, 0.0340, 0.0349, 0.0397, 0.0359, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 19:00:46,476 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129546.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:00:47,222 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 19:01:00,754 INFO [finetune.py:976] (2/7) Epoch 23, batch 3550, loss[loss=0.1576, simple_loss=0.2322, pruned_loss=0.04155, over 4916.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2412, pruned_loss=0.04956, over 955489.43 frames. ], batch size: 46, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:01:05,537 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.78 vs. limit=5.0 +2023-04-27 19:01:08,390 INFO [optim.py:369] (2/7) 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 19:01:30,931 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 19:01:38,378 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129594.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:01:57,767 INFO [finetune.py:976] (2/7) Epoch 23, batch 3600, loss[loss=0.1514, simple_loss=0.2219, pruned_loss=0.04044, over 4872.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2388, pruned_loss=0.04838, over 955512.51 frames. ], batch size: 31, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:01:59,686 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129612.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:02:47,466 INFO [finetune.py:976] (2/7) Epoch 23, batch 3650, loss[loss=0.1857, simple_loss=0.2586, pruned_loss=0.05635, over 4825.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2427, pruned_loss=0.04987, over 951690.68 frames. ], batch size: 47, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:02:51,890 INFO [optim.py:369] (2/7) 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 19:02:52,039 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7890, 3.7258, 2.8613, 4.4117, 3.9091, 3.8065, 1.6445, 3.7407], + device='cuda:2'), covar=tensor([0.1752, 0.1355, 0.3095, 0.1803, 0.4128, 0.1776, 0.6271, 0.2947], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0218, 0.0253, 0.0306, 0.0296, 0.0246, 0.0274, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 19:03:00,803 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9327, 2.3739, 2.1438, 2.3192, 2.1100, 2.2432, 2.3044, 2.2097], + device='cuda:2'), covar=tensor([0.3969, 0.5679, 0.5423, 0.4804, 0.5602, 0.7260, 0.6253, 0.5323], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0374, 0.0327, 0.0339, 0.0347, 0.0394, 0.0357, 0.0330], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 19:03:02,055 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129673.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:03:26,404 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6920, 1.5849, 1.8938, 2.0928, 1.6003, 1.4026, 1.7236, 1.0240], + device='cuda:2'), covar=tensor([0.0608, 0.0672, 0.0484, 0.0700, 0.0705, 0.0926, 0.0683, 0.0695], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0066, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 19:03:26,451 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.20 vs. limit=5.0 +2023-04-27 19:03:55,977 INFO [finetune.py:976] (2/7) Epoch 23, batch 3700, loss[loss=0.1686, simple_loss=0.2422, pruned_loss=0.04752, over 4764.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2466, pruned_loss=0.0514, over 953000.67 frames. ], batch size: 26, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:04:41,812 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6864, 1.1948, 4.4459, 4.1817, 3.9303, 4.2363, 4.1520, 3.9219], + device='cuda:2'), covar=tensor([0.6895, 0.6474, 0.0996, 0.1674, 0.1123, 0.1911, 0.1498, 0.1640], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0307, 0.0407, 0.0406, 0.0347, 0.0409, 0.0315, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:05:03,363 INFO [finetune.py:976] (2/7) Epoch 23, batch 3750, loss[loss=0.1689, simple_loss=0.2489, pruned_loss=0.04445, over 4920.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.247, pruned_loss=0.05114, over 954162.23 frames. ], batch size: 42, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:05:12,866 INFO [optim.py:369] (2/7) 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 19:05:15,279 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129769.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:05:23,582 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8531, 1.4334, 2.0293, 2.3566, 1.9325, 1.8167, 1.9585, 1.8613], + device='cuda:2'), covar=tensor([0.4795, 0.7127, 0.6577, 0.5460, 0.6301, 0.8315, 0.7887, 0.8611], + device='cuda:2'), in_proj_covar=tensor([0.0433, 0.0417, 0.0511, 0.0507, 0.0462, 0.0495, 0.0499, 0.0509], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:05:45,714 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129791.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:05:46,404 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6681, 1.0332, 1.6979, 2.1354, 1.7279, 1.5771, 1.6807, 1.6343], + device='cuda:2'), covar=tensor([0.3825, 0.5941, 0.4989, 0.4626, 0.5244, 0.6416, 0.6249, 0.7317], + device='cuda:2'), in_proj_covar=tensor([0.0432, 0.0416, 0.0510, 0.0506, 0.0461, 0.0494, 0.0498, 0.0509], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:06:08,152 INFO [finetune.py:976] (2/7) Epoch 23, batch 3800, loss[loss=0.181, simple_loss=0.2592, pruned_loss=0.05143, over 4810.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2472, pruned_loss=0.05088, over 953272.76 frames. ], batch size: 39, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:07:04,472 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129852.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:07:13,929 INFO [finetune.py:976] (2/7) Epoch 23, batch 3850, loss[loss=0.1633, simple_loss=0.2283, pruned_loss=0.04913, over 4785.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2454, pruned_loss=0.05022, over 954682.87 frames. ], batch size: 51, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:07:23,739 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8799, 1.4087, 1.4568, 1.6777, 2.0567, 1.6825, 1.3955, 1.4147], + device='cuda:2'), covar=tensor([0.1397, 0.1481, 0.1768, 0.1207, 0.0826, 0.1730, 0.2089, 0.2215], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0312, 0.0353, 0.0287, 0.0326, 0.0310, 0.0301, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.4779e-05, 6.4355e-05, 7.4393e-05, 5.7706e-05, 6.7175e-05, 6.5126e-05, + 6.2839e-05, 7.9335e-05], device='cuda:2') +2023-04-27 19:07:24,864 INFO [optim.py:369] (2/7) 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 19:07:59,419 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-04-27 19:07:59,901 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2042, 1.6933, 2.0866, 2.1538, 2.0289, 1.6126, 1.0367, 1.7004], + device='cuda:2'), covar=tensor([0.3156, 0.3142, 0.1780, 0.2082, 0.2687, 0.2731, 0.4280, 0.2004], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0246, 0.0229, 0.0315, 0.0221, 0.0236, 0.0228, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 19:08:11,816 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1979, 3.4835, 1.5099, 2.3545, 2.9224, 2.3112, 4.6830, 2.9238], + device='cuda:2'), covar=tensor([0.0485, 0.0599, 0.0712, 0.1004, 0.0440, 0.0790, 0.0261, 0.0463], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 19:08:15,872 INFO [finetune.py:976] (2/7) Epoch 23, batch 3900, loss[loss=0.1557, simple_loss=0.2241, pruned_loss=0.04361, over 4848.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2422, pruned_loss=0.04903, over 953132.91 frames. ], batch size: 49, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:09:17,305 INFO [finetune.py:976] (2/7) Epoch 23, batch 3950, loss[loss=0.1757, simple_loss=0.239, pruned_loss=0.05618, over 4902.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2392, pruned_loss=0.04804, over 954587.65 frames. ], batch size: 32, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:09:27,939 INFO [optim.py:369] (2/7) 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 19:09:35,335 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129968.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:09:35,388 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5797, 1.4227, 1.7807, 1.8144, 1.4364, 1.3773, 1.4781, 0.8989], + device='cuda:2'), covar=tensor([0.0455, 0.0549, 0.0394, 0.0493, 0.0678, 0.0935, 0.0516, 0.0588], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0068, 0.0066, 0.0068, 0.0074, 0.0095, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 19:10:16,063 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130000.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:10:22,143 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 4000, loss[loss=0.1485, simple_loss=0.229, pruned_loss=0.03402, over 4881.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.239, pruned_loss=0.04834, over 955485.05 frames. ], batch size: 32, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:10:36,158 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5937, 1.7355, 0.8066, 1.2953, 1.8576, 1.4404, 1.3459, 1.4531], + device='cuda:2'), covar=tensor([0.0485, 0.0353, 0.0354, 0.0537, 0.0271, 0.0498, 0.0479, 0.0556], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 19:10:46,631 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1395, 1.7592, 2.2503, 2.5539, 2.1219, 2.0174, 2.1667, 2.1138], + device='cuda:2'), covar=tensor([0.4560, 0.7008, 0.7301, 0.5339, 0.6184, 0.8568, 0.8945, 1.0062], + device='cuda:2'), in_proj_covar=tensor([0.0433, 0.0416, 0.0511, 0.0506, 0.0461, 0.0495, 0.0499, 0.0510], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:11:43,267 INFO [finetune.py:976] (2/7) Epoch 23, batch 4050, loss[loss=0.1869, simple_loss=0.2636, pruned_loss=0.05506, over 4755.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2423, pruned_loss=0.04964, over 953952.89 frames. ], batch size: 59, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:11:44,608 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130061.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:11:45,888 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130063.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:11:46,468 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2485, 1.3584, 1.2267, 1.5565, 1.4116, 1.6281, 1.2256, 3.0314], + device='cuda:2'), covar=tensor([0.0661, 0.1082, 0.1059, 0.1403, 0.0852, 0.0636, 0.1032, 0.0244], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 19:11:53,610 INFO [optim.py:369] (2/7) 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 19:11:54,369 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8728, 2.0219, 1.9463, 1.6664, 2.1637, 1.6486, 2.6558, 1.7073], + device='cuda:2'), covar=tensor([0.3553, 0.1633, 0.4186, 0.2690, 0.1443, 0.2416, 0.1218, 0.4004], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0345, 0.0422, 0.0347, 0.0373, 0.0371, 0.0364, 0.0414], + device='cuda:2'), out_proj_covar=tensor([9.9278e-05, 1.0316e-04, 1.2793e-04, 1.0442e-04, 1.1086e-04, 1.1069e-04, + 1.0693e-04, 1.2459e-04], device='cuda:2') +2023-04-27 19:11:56,082 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130069.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:12:52,034 INFO [finetune.py:976] (2/7) Epoch 23, batch 4100, loss[loss=0.2719, simple_loss=0.3255, pruned_loss=0.1091, over 4063.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2444, pruned_loss=0.04967, over 955273.46 frames. ], batch size: 65, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:13:01,825 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130117.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:13:32,113 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130147.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:13:51,930 INFO [finetune.py:976] (2/7) Epoch 23, batch 4150, loss[loss=0.1618, simple_loss=0.2364, pruned_loss=0.04357, over 4257.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2462, pruned_loss=0.0506, over 956405.85 frames. ], batch size: 66, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:14:02,699 INFO [optim.py:369] (2/7) 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 19:14:11,462 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4646, 1.7650, 1.6659, 1.9231, 1.9758, 2.1861, 1.6211, 3.6883], + device='cuda:2'), covar=tensor([0.0493, 0.0677, 0.0675, 0.0987, 0.0496, 0.0538, 0.0657, 0.0170], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 19:14:48,093 INFO [finetune.py:976] (2/7) Epoch 23, batch 4200, loss[loss=0.1751, simple_loss=0.253, pruned_loss=0.04859, over 4902.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.246, pruned_loss=0.05037, over 954529.00 frames. ], batch size: 36, lr: 3.09e-03, grad_scale: 16.0 +2023-04-27 19:15:51,651 INFO [finetune.py:976] (2/7) Epoch 23, batch 4250, loss[loss=0.1517, simple_loss=0.2303, pruned_loss=0.03657, over 4900.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2425, pruned_loss=0.04887, over 954472.69 frames. ], batch size: 36, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:16:01,965 INFO [optim.py:369] (2/7) 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 19:16:03,337 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130268.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:16:53,665 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 19:16:55,316 INFO [finetune.py:976] (2/7) Epoch 23, batch 4300, loss[loss=0.1509, simple_loss=0.2185, pruned_loss=0.04163, over 4936.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2397, pruned_loss=0.04787, over 955186.24 frames. ], batch size: 38, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:17:05,108 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130316.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:17:44,201 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8884, 2.6419, 1.9740, 2.0005, 1.4362, 1.4274, 2.0388, 1.3919], + device='cuda:2'), covar=tensor([0.1615, 0.1211, 0.1320, 0.1525, 0.2200, 0.1925, 0.0916, 0.1938], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0212, 0.0169, 0.0206, 0.0201, 0.0186, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 19:17:58,665 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:18:05,097 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130358.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 19:18:05,662 INFO [finetune.py:976] (2/7) Epoch 23, batch 4350, loss[loss=0.1997, simple_loss=0.2672, pruned_loss=0.06604, over 4823.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2376, pruned_loss=0.04723, over 957277.93 frames. ], batch size: 40, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:18:09,001 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1513, 2.5014, 1.3707, 1.9177, 2.5617, 1.9607, 1.9047, 1.9875], + device='cuda:2'), covar=tensor([0.0430, 0.0309, 0.0265, 0.0494, 0.0216, 0.0471, 0.0482, 0.0484], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 19:18:10,119 INFO [optim.py:369] (2/7) 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 19:18:11,075 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 19:18:29,070 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 4400, loss[loss=0.1865, simple_loss=0.2631, pruned_loss=0.05494, over 4836.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.239, pruned_loss=0.04765, over 957200.75 frames. ], batch size: 44, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:19:34,808 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:19:45,197 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130438.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:19:55,329 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130445.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:19:56,565 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130447.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:20:16,250 INFO [finetune.py:976] (2/7) Epoch 23, batch 4450, loss[loss=0.1861, simple_loss=0.2615, pruned_loss=0.05533, over 4811.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2416, pruned_loss=0.04782, over 957067.05 frames. ], batch size: 45, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:20:25,829 INFO [optim.py:369] (2/7) 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 19:20:28,568 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5891, 3.8559, 0.8316, 1.9987, 2.1765, 2.6499, 2.2571, 0.9993], + device='cuda:2'), covar=tensor([0.1376, 0.0913, 0.2115, 0.1196, 0.0966, 0.0999, 0.1460, 0.2068], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0138, 0.0121, 0.0133, 0.0152, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 19:20:36,470 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1650, 2.0100, 2.1923, 2.5638, 2.5321, 2.0861, 1.6923, 2.3569], + device='cuda:2'), covar=tensor([0.0770, 0.1043, 0.0686, 0.0511, 0.0543, 0.0800, 0.0738, 0.0474], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0200, 0.0181, 0.0171, 0.0174, 0.0178, 0.0147, 0.0176], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:21:01,278 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130493.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:21:02,959 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130495.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:21:11,830 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130499.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:21:23,653 INFO [finetune.py:976] (2/7) Epoch 23, batch 4500, loss[loss=0.1494, simple_loss=0.2283, pruned_loss=0.03529, over 4763.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2413, pruned_loss=0.04699, over 956260.48 frames. ], batch size: 28, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:22:09,637 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3743, 1.5566, 1.4358, 1.7213, 1.7348, 1.9591, 1.3771, 3.4670], + device='cuda:2'), covar=tensor([0.0579, 0.0805, 0.0808, 0.1189, 0.0594, 0.0451, 0.0734, 0.0162], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 19:22:17,974 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 4550, loss[loss=0.1502, simple_loss=0.2268, pruned_loss=0.03685, over 4779.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2427, pruned_loss=0.04768, over 955210.18 frames. ], batch size: 29, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:22:31,556 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7290, 3.7029, 2.7365, 4.3087, 3.7484, 3.7101, 1.4808, 3.7322], + device='cuda:2'), covar=tensor([0.1908, 0.1292, 0.3482, 0.1901, 0.2274, 0.1804, 0.6058, 0.2526], + device='cuda:2'), in_proj_covar=tensor([0.0248, 0.0221, 0.0256, 0.0309, 0.0300, 0.0250, 0.0278, 0.0277], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 19:22:42,286 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 19:23:36,952 INFO [finetune.py:976] (2/7) Epoch 23, batch 4600, loss[loss=0.1827, simple_loss=0.2497, pruned_loss=0.0579, over 4786.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2424, pruned_loss=0.04743, over 955303.87 frames. ], batch size: 29, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:23:37,077 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130609.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:23:56,075 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130626.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:24:36,992 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130656.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:24:38,736 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130658.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:24:39,265 INFO [finetune.py:976] (2/7) Epoch 23, batch 4650, loss[loss=0.1698, simple_loss=0.2286, pruned_loss=0.05549, over 4729.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2399, pruned_loss=0.04746, over 955639.53 frames. ], batch size: 59, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:24:48,331 INFO [optim.py:369] (2/7) 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 19:25:39,303 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:25:40,550 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130706.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:25:47,205 INFO [finetune.py:976] (2/7) Epoch 23, batch 4700, loss[loss=0.1607, simple_loss=0.2305, pruned_loss=0.04548, over 4822.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2376, pruned_loss=0.04709, over 956404.50 frames. ], batch size: 39, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:25:49,719 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4631, 1.4317, 1.4811, 1.1010, 1.4304, 1.2405, 1.7137, 1.4350], + device='cuda:2'), covar=tensor([0.3610, 0.1821, 0.5039, 0.2628, 0.1566, 0.2169, 0.1541, 0.4648], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0348, 0.0425, 0.0350, 0.0377, 0.0374, 0.0367, 0.0418], + device='cuda:2'), out_proj_covar=tensor([9.9931e-05, 1.0394e-04, 1.2876e-04, 1.0514e-04, 1.1208e-04, 1.1143e-04, + 1.0785e-04, 1.2585e-04], device='cuda:2') +2023-04-27 19:26:22,718 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130740.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:26:50,253 INFO [finetune.py:976] (2/7) Epoch 23, batch 4750, loss[loss=0.142, simple_loss=0.2238, pruned_loss=0.03012, over 4790.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2362, pruned_loss=0.04683, over 955161.25 frames. ], batch size: 29, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:27:00,091 INFO [optim.py:369] (2/7) 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 19:27:24,059 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130788.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:27:33,679 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 4800, loss[loss=0.2084, simple_loss=0.2855, pruned_loss=0.06566, over 4848.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2391, pruned_loss=0.04797, over 953480.05 frames. ], batch size: 47, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:28:05,971 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9536, 1.7540, 2.0899, 2.4126, 2.0331, 1.9325, 2.0117, 1.9572], + device='cuda:2'), covar=tensor([0.4681, 0.7075, 0.7468, 0.5361, 0.6226, 0.9049, 0.8675, 0.9868], + device='cuda:2'), in_proj_covar=tensor([0.0436, 0.0418, 0.0512, 0.0506, 0.0464, 0.0497, 0.0501, 0.0512], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:28:13,829 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-27 19:28:54,603 INFO [finetune.py:976] (2/7) Epoch 23, batch 4850, loss[loss=0.1863, simple_loss=0.2671, pruned_loss=0.05277, over 4908.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2419, pruned_loss=0.04905, over 953529.69 frames. ], batch size: 36, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:28:54,769 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0074, 2.6034, 2.0206, 2.0116, 1.5010, 1.5186, 2.0668, 1.4360], + device='cuda:2'), covar=tensor([0.1751, 0.1429, 0.1458, 0.1750, 0.2392, 0.2042, 0.0994, 0.2096], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0211, 0.0168, 0.0204, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 19:29:05,299 INFO [optim.py:369] (2/7) 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 19:29:05,512 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7995, 1.3808, 1.8525, 2.2853, 1.8815, 1.8176, 1.8753, 1.7900], + device='cuda:2'), covar=tensor([0.4613, 0.6815, 0.6699, 0.5547, 0.5895, 0.7607, 0.7951, 0.9578], + device='cuda:2'), in_proj_covar=tensor([0.0435, 0.0418, 0.0511, 0.0507, 0.0464, 0.0497, 0.0500, 0.0513], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:29:26,590 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 19:30:00,410 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130904.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:30:03,974 INFO [finetune.py:976] (2/7) Epoch 23, batch 4900, loss[loss=0.161, simple_loss=0.2338, pruned_loss=0.04413, over 4923.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2436, pruned_loss=0.0491, over 954826.56 frames. ], batch size: 33, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:30:23,706 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130921.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:30:44,789 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9792, 1.8741, 1.7147, 1.6220, 2.0705, 1.6672, 2.4235, 1.5284], + device='cuda:2'), covar=tensor([0.3471, 0.1741, 0.4668, 0.2731, 0.1597, 0.2257, 0.1389, 0.4352], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0348, 0.0424, 0.0350, 0.0376, 0.0373, 0.0368, 0.0418], + device='cuda:2'), out_proj_covar=tensor([9.9746e-05, 1.0374e-04, 1.2857e-04, 1.0518e-04, 1.1165e-04, 1.1123e-04, + 1.0808e-04, 1.2576e-04], device='cuda:2') +2023-04-27 19:31:11,883 INFO [finetune.py:976] (2/7) Epoch 23, batch 4950, loss[loss=0.1436, simple_loss=0.221, pruned_loss=0.03307, over 4920.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2446, pruned_loss=0.04959, over 953950.86 frames. ], batch size: 37, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:31:21,969 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 23, batch 5000, loss[loss=0.182, simple_loss=0.2513, pruned_loss=0.05636, over 4281.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.243, pruned_loss=0.04893, over 955143.91 frames. ], batch size: 65, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:33:10,248 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131040.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:33:35,487 INFO [finetune.py:976] (2/7) Epoch 23, batch 5050, loss[loss=0.1295, simple_loss=0.2093, pruned_loss=0.02486, over 4751.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2398, pruned_loss=0.04785, over 952517.88 frames. ], batch size: 27, lr: 3.08e-03, grad_scale: 16.0 +2023-04-27 19:33:41,159 INFO [optim.py:369] (2/7) 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 19:33:53,122 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-04-27 19:34:14,398 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:34:14,425 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:34:18,269 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131094.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:34:38,867 INFO [finetune.py:976] (2/7) Epoch 23, batch 5100, loss[loss=0.1452, simple_loss=0.2165, pruned_loss=0.03692, over 4904.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2374, pruned_loss=0.04691, over 954219.33 frames. ], batch size: 37, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:34:59,053 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-04-27 19:35:11,742 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131136.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:35:19,507 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7137, 2.0285, 0.9839, 1.4862, 2.0668, 1.5932, 1.5220, 1.5808], + device='cuda:2'), covar=tensor([0.0475, 0.0339, 0.0319, 0.0534, 0.0259, 0.0500, 0.0502, 0.0544], + device='cuda:2'), in_proj_covar=tensor([0.0027, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0027], + device='cuda:2'), out_proj_covar=tensor([0.0050, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0049, 0.0051], + device='cuda:2') +2023-04-27 19:35:21,309 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131142.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:35:41,501 INFO [finetune.py:976] (2/7) Epoch 23, batch 5150, loss[loss=0.1304, simple_loss=0.2037, pruned_loss=0.02849, over 4733.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2368, pruned_loss=0.04696, over 953445.57 frames. ], batch size: 23, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:35:51,236 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131174.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:36:42,389 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131204.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:36:45,362 INFO [finetune.py:976] (2/7) Epoch 23, batch 5200, loss[loss=0.1955, simple_loss=0.2804, pruned_loss=0.05527, over 4909.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2407, pruned_loss=0.04804, over 952756.62 frames. ], batch size: 36, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:37:04,406 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131235.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 19:37:43,653 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131252.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:37:52,702 INFO [finetune.py:976] (2/7) Epoch 23, batch 5250, loss[loss=0.1533, simple_loss=0.226, pruned_loss=0.04029, over 4787.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2436, pruned_loss=0.04855, over 953736.97 frames. ], batch size: 51, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:37:57,554 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131269.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:38:14,726 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-04-27 19:38:50,182 INFO [finetune.py:976] (2/7) Epoch 23, batch 5300, loss[loss=0.173, simple_loss=0.2472, pruned_loss=0.04935, over 4770.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.244, pruned_loss=0.0488, over 954291.44 frames. ], batch size: 28, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:39:15,167 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-27 19:39:24,513 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8227, 2.4735, 1.8560, 1.8420, 1.3044, 1.3798, 1.9562, 1.2980], + device='cuda:2'), covar=tensor([0.1731, 0.1309, 0.1374, 0.1642, 0.2321, 0.1953, 0.0947, 0.1988], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0208, 0.0167, 0.0203, 0.0197, 0.0183, 0.0154, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 19:39:56,069 INFO [finetune.py:976] (2/7) Epoch 23, batch 5350, loss[loss=0.125, simple_loss=0.1881, pruned_loss=0.03095, over 4749.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2434, pruned_loss=0.04836, over 953360.43 frames. ], batch size: 23, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:39:56,862 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-27 19:40:07,007 INFO [optim.py:369] (2/7) 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 19:40:18,407 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 19:41:03,583 INFO [finetune.py:976] (2/7) Epoch 23, batch 5400, loss[loss=0.2039, simple_loss=0.2722, pruned_loss=0.06775, over 4897.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.241, pruned_loss=0.0474, over 955057.47 frames. ], batch size: 43, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:41:03,677 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131409.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:41:04,287 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6192, 1.3069, 4.4450, 4.1509, 3.8420, 4.1273, 4.0564, 3.8722], + device='cuda:2'), covar=tensor([0.7574, 0.5972, 0.1064, 0.1849, 0.1133, 0.1754, 0.1894, 0.1661], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0308, 0.0408, 0.0408, 0.0348, 0.0411, 0.0318, 0.0368], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:41:12,777 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131416.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:42:12,941 INFO [finetune.py:976] (2/7) Epoch 23, batch 5450, loss[loss=0.1828, simple_loss=0.2477, pruned_loss=0.05894, over 4838.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2392, pruned_loss=0.04719, over 955448.62 frames. ], batch size: 47, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:42:22,592 INFO [optim.py:369] (2/7) 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 19:42:31,938 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131470.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:42:43,237 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131492.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:43:07,761 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-27 19:43:22,051 INFO [finetune.py:976] (2/7) Epoch 23, batch 5500, loss[loss=0.1552, simple_loss=0.2059, pruned_loss=0.05225, over 4458.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2368, pruned_loss=0.04637, over 956560.89 frames. ], batch size: 19, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:43:52,911 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131530.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 19:43:59,269 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8999, 1.5426, 1.5121, 1.7220, 2.1200, 1.6430, 1.4710, 1.4255], + device='cuda:2'), covar=tensor([0.1633, 0.1591, 0.1820, 0.1265, 0.0883, 0.1787, 0.1982, 0.2232], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0308, 0.0351, 0.0284, 0.0326, 0.0305, 0.0298, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.4054e-05, 6.3645e-05, 7.3858e-05, 5.6953e-05, 6.7175e-05, 6.3967e-05, + 6.2177e-05, 7.8892e-05], device='cuda:2') +2023-04-27 19:44:10,970 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5213, 1.4762, 0.7529, 1.2411, 1.5294, 1.3239, 1.2892, 1.2967], + device='cuda:2'), covar=tensor([0.0602, 0.0349, 0.0372, 0.0616, 0.0300, 0.0633, 0.0625, 0.0651], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0049, 0.0051], + device='cuda:2') +2023-04-27 19:44:16,311 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 23, batch 5550, loss[loss=0.186, simple_loss=0.2678, pruned_loss=0.05207, over 4899.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2384, pruned_loss=0.04734, over 956335.51 frames. ], batch size: 32, lr: 3.08e-03, grad_scale: 32.0 +2023-04-27 19:44:32,380 INFO [optim.py:369] (2/7) 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 19:45:32,464 INFO [finetune.py:976] (2/7) Epoch 23, batch 5600, loss[loss=0.1774, simple_loss=0.2525, pruned_loss=0.05113, over 4912.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2414, pruned_loss=0.04822, over 955883.61 frames. ], batch size: 43, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:46:28,599 INFO [finetune.py:976] (2/7) Epoch 23, batch 5650, loss[loss=0.1569, simple_loss=0.2356, pruned_loss=0.03906, over 4875.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2442, pruned_loss=0.04888, over 952403.47 frames. ], batch size: 34, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:46:38,006 INFO [optim.py:369] (2/7) 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 19:47:29,413 INFO [finetune.py:976] (2/7) Epoch 23, batch 5700, loss[loss=0.1473, simple_loss=0.2147, pruned_loss=0.03991, over 3750.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2413, pruned_loss=0.04876, over 935961.45 frames. ], batch size: 16, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:48:21,855 INFO [finetune.py:976] (2/7) Epoch 24, batch 0, loss[loss=0.1763, simple_loss=0.2541, pruned_loss=0.04927, over 4919.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2541, pruned_loss=0.04927, over 4919.00 frames. ], batch size: 42, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:48:21,856 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 19:48:31,593 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7624, 1.6708, 0.9675, 1.4972, 1.6027, 1.5786, 1.5362, 1.5860], + device='cuda:2'), covar=tensor([0.0394, 0.0325, 0.0312, 0.0465, 0.0272, 0.0456, 0.0399, 0.0463], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 19:48:37,426 INFO [finetune.py:1010] (2/7) Epoch 24, validation: loss=0.1552, simple_loss=0.2243, pruned_loss=0.04308, over 2265189.00 frames. +2023-04-27 19:48:37,426 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 19:48:43,777 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4064, 2.8395, 1.3209, 1.6776, 2.1234, 1.4530, 3.5429, 1.7572], + device='cuda:2'), covar=tensor([0.0531, 0.0868, 0.0737, 0.1002, 0.0433, 0.0823, 0.0166, 0.0538], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0050, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 19:48:46,849 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3434, 1.6820, 1.5468, 1.9436, 1.9456, 2.0320, 1.4942, 3.7679], + device='cuda:2'), covar=tensor([0.0595, 0.0801, 0.0771, 0.1183, 0.0573, 0.0497, 0.0752, 0.0144], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 19:49:05,132 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 19:49:12,805 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131765.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:49:13,311 INFO [optim.py:369] (2/7) 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 19:49:21,985 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131772.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:49:23,423 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-04-27 19:49:36,760 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9713, 1.8169, 2.2903, 2.4955, 1.8550, 1.6112, 1.8532, 0.9256], + device='cuda:2'), covar=tensor([0.0595, 0.0674, 0.0372, 0.0622, 0.0675, 0.1069, 0.0688, 0.0733], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0068, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 19:49:37,727 INFO [finetune.py:976] (2/7) Epoch 24, batch 50, loss[loss=0.1526, simple_loss=0.225, pruned_loss=0.04009, over 4813.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2435, pruned_loss=0.04774, over 214968.07 frames. ], batch size: 38, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:50:16,898 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0289, 3.9864, 2.8598, 4.6719, 4.0801, 4.0350, 1.8022, 3.9789], + device='cuda:2'), covar=tensor([0.1723, 0.1138, 0.3334, 0.1330, 0.2950, 0.1787, 0.5705, 0.2329], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0217, 0.0251, 0.0302, 0.0292, 0.0244, 0.0271, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 19:50:39,505 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131830.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:50:48,614 INFO [finetune.py:976] (2/7) Epoch 24, batch 100, loss[loss=0.1719, simple_loss=0.2465, pruned_loss=0.04869, over 4907.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2384, pruned_loss=0.04604, over 380694.57 frames. ], batch size: 36, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:50:56,376 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131848.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:51:01,837 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131857.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:51:07,129 INFO [optim.py:369] (2/7) 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 19:51:23,455 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131878.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:51:35,586 INFO [finetune.py:976] (2/7) Epoch 24, batch 150, loss[loss=0.1834, simple_loss=0.2576, pruned_loss=0.05464, over 4900.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.235, pruned_loss=0.04567, over 509128.45 frames. ], batch size: 37, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:51:45,079 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1122, 1.4462, 1.2893, 1.7303, 1.6340, 1.7427, 1.3208, 3.0687], + device='cuda:2'), covar=tensor([0.0687, 0.0860, 0.0861, 0.1260, 0.0662, 0.0510, 0.0795, 0.0183], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 19:52:18,775 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 200, loss[loss=0.1811, simple_loss=0.2583, pruned_loss=0.05192, over 4803.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2349, pruned_loss=0.04658, over 608078.10 frames. ], batch size: 51, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:53:12,560 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-04-27 19:53:12,911 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 250, loss[loss=0.2257, simple_loss=0.2909, pruned_loss=0.08025, over 4807.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2396, pruned_loss=0.04898, over 685072.81 frames. ], batch size: 51, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:53:30,828 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-04-27 19:53:57,465 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132031.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:54:01,445 INFO [finetune.py:976] (2/7) Epoch 24, batch 300, loss[loss=0.1657, simple_loss=0.2418, pruned_loss=0.04481, over 4827.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2423, pruned_loss=0.04908, over 746780.69 frames. ], batch size: 39, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:54:17,446 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-04-27 19:54:24,636 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132054.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:54:36,453 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 19:54:46,354 INFO [zipformer.py:1188] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.6468, 1.4150, 0.5782, 1.3229, 1.4690, 1.4942, 1.3812, 1.3801], + device='cuda:2'), covar=tensor([0.0470, 0.0377, 0.0385, 0.0566, 0.0291, 0.0538, 0.0502, 0.0565], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0045, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0051], + device='cuda:2') +2023-04-27 19:55:00,439 INFO [finetune.py:976] (2/7) Epoch 24, batch 350, loss[loss=0.1565, simple_loss=0.2468, pruned_loss=0.03306, over 4911.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2446, pruned_loss=0.04899, over 793890.72 frames. ], batch size: 38, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:55:10,379 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132092.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:55:40,630 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132113.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:55:41,934 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132115.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 19:55:44,963 INFO [zipformer.py:1188] (2/7) 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] (2/7) attn_weights_entropy = tensor([2.5785, 2.9653, 1.1001, 1.8207, 2.4077, 1.5622, 4.2743, 1.8818], + device='cuda:2'), covar=tensor([0.0584, 0.0940, 0.0869, 0.1310, 0.0507, 0.0991, 0.0344, 0.0645], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0050, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 19:56:06,850 INFO [finetune.py:976] (2/7) Epoch 24, batch 400, loss[loss=0.1599, simple_loss=0.2358, pruned_loss=0.04198, over 4927.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2465, pruned_loss=0.04956, over 830434.00 frames. ], batch size: 33, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:56:25,725 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132148.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:56:49,858 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 450, loss[loss=0.1378, simple_loss=0.2046, pruned_loss=0.03545, over 4293.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2453, pruned_loss=0.04934, over 858361.88 frames. ], batch size: 18, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:57:12,345 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7819, 1.8369, 1.6448, 1.4130, 1.7772, 1.5228, 2.2795, 1.5024], + device='cuda:2'), covar=tensor([0.3514, 0.1641, 0.4831, 0.2711, 0.1698, 0.2135, 0.1485, 0.4441], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0353, 0.0428, 0.0354, 0.0380, 0.0378, 0.0370, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 19:57:15,196 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132196.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:57:27,639 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132213.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:57:31,901 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6121, 1.3881, 1.3664, 1.5256, 1.8499, 1.5242, 1.3373, 1.2137], + device='cuda:2'), covar=tensor([0.1809, 0.1293, 0.1754, 0.1238, 0.0762, 0.1474, 0.1845, 0.2363], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0306, 0.0348, 0.0283, 0.0325, 0.0304, 0.0297, 0.0370], + device='cuda:2'), out_proj_covar=tensor([6.3617e-05, 6.3214e-05, 7.3273e-05, 5.6743e-05, 6.6991e-05, 6.3633e-05, + 6.1923e-05, 7.8484e-05], device='cuda:2') +2023-04-27 19:57:42,207 INFO [finetune.py:976] (2/7) Epoch 24, batch 500, loss[loss=0.2083, simple_loss=0.2722, pruned_loss=0.07214, over 4907.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2421, pruned_loss=0.0479, over 881268.38 frames. ], batch size: 43, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:58:03,236 INFO [optim.py:369] (2/7) 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 19:58:06,507 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8373, 2.1591, 2.1334, 2.2605, 2.0762, 2.1535, 2.1516, 2.1360], + device='cuda:2'), covar=tensor([0.3886, 0.6042, 0.4746, 0.4295, 0.5758, 0.6940, 0.6254, 0.5145], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0374, 0.0327, 0.0339, 0.0349, 0.0394, 0.0357, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 19:58:16,036 INFO [finetune.py:976] (2/7) Epoch 24, batch 550, loss[loss=0.1512, simple_loss=0.2254, pruned_loss=0.03852, over 4816.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2386, pruned_loss=0.04681, over 899914.47 frames. ], batch size: 39, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:58:47,059 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-04-27 19:58:49,840 INFO [finetune.py:976] (2/7) Epoch 24, batch 600, loss[loss=0.1992, simple_loss=0.2819, pruned_loss=0.05822, over 4807.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2408, pruned_loss=0.04816, over 912960.59 frames. ], batch size: 51, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:59:10,243 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 650, loss[loss=0.1679, simple_loss=0.2433, pruned_loss=0.04621, over 4940.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2428, pruned_loss=0.049, over 922729.04 frames. ], batch size: 38, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 19:59:23,069 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132387.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 19:59:39,120 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132410.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 20:00:05,289 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 20:00:08,097 INFO [finetune.py:976] (2/7) Epoch 24, batch 700, loss[loss=0.209, simple_loss=0.2784, pruned_loss=0.06981, over 4808.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2433, pruned_loss=0.04881, over 929928.50 frames. ], batch size: 45, lr: 3.07e-03, grad_scale: 32.0 +2023-04-27 20:00:50,346 INFO [optim.py:369] (2/7) 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 20:00:52,839 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5718, 1.6633, 1.3959, 1.1212, 1.1794, 1.1647, 1.4085, 1.1389], + device='cuda:2'), covar=tensor([0.1609, 0.1255, 0.1451, 0.1650, 0.2225, 0.1972, 0.1040, 0.1997], + device='cuda:2'), in_proj_covar=tensor([0.0195, 0.0209, 0.0167, 0.0203, 0.0199, 0.0184, 0.0154, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 20:01:15,825 INFO [finetune.py:976] (2/7) Epoch 24, batch 750, loss[loss=0.1953, simple_loss=0.2652, pruned_loss=0.0627, over 4853.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2438, pruned_loss=0.04889, over 933785.12 frames. ], batch size: 31, lr: 3.06e-03, grad_scale: 32.0 +2023-04-27 20:01:22,922 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-04-27 20:01:47,424 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4134, 1.6604, 1.8529, 1.9075, 1.8061, 1.8558, 1.8892, 1.8537], + device='cuda:2'), covar=tensor([0.4175, 0.5844, 0.4260, 0.4606, 0.5529, 0.7364, 0.5559, 0.4838], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0376, 0.0328, 0.0340, 0.0350, 0.0395, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:01:48,566 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 800, loss[loss=0.1612, simple_loss=0.2204, pruned_loss=0.05106, over 4927.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.244, pruned_loss=0.04865, over 938505.26 frames. ], batch size: 33, lr: 3.06e-03, grad_scale: 32.0 +2023-04-27 20:02:34,658 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4300, 1.2979, 1.6814, 1.6183, 1.3054, 1.2339, 1.3484, 0.7739], + device='cuda:2'), covar=tensor([0.0485, 0.0628, 0.0401, 0.0501, 0.0748, 0.1174, 0.0470, 0.0595], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0068, 0.0074, 0.0095, 0.0072, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:02:53,593 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132561.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:03:01,811 INFO [optim.py:369] (2/7) 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 20:03:27,812 INFO [finetune.py:976] (2/7) Epoch 24, batch 850, loss[loss=0.1497, simple_loss=0.2231, pruned_loss=0.03814, over 4726.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2412, pruned_loss=0.0477, over 942559.65 frames. ], batch size: 59, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:04:34,502 INFO [finetune.py:976] (2/7) Epoch 24, batch 900, loss[loss=0.1771, simple_loss=0.2495, pruned_loss=0.05235, over 4807.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2387, pruned_loss=0.0468, over 945433.22 frames. ], batch size: 45, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:05:02,117 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132660.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:05:12,901 INFO [optim.py:369] (2/7) 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 20:05:26,148 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-27 20:05:44,310 INFO [finetune.py:976] (2/7) Epoch 24, batch 950, loss[loss=0.198, simple_loss=0.2748, pruned_loss=0.06057, over 4911.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2379, pruned_loss=0.0469, over 949301.41 frames. ], batch size: 43, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:05:44,398 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132687.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:06:08,254 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132710.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 20:06:27,924 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132721.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:06:43,787 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 1000, loss[loss=0.1767, simple_loss=0.2476, pruned_loss=0.05295, over 4919.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2406, pruned_loss=0.04767, over 951714.98 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:06:55,173 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132746.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:07:02,345 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132758.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:07:02,377 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9899, 4.3180, 1.0446, 2.1636, 2.6336, 3.0041, 2.5736, 1.0538], + device='cuda:2'), covar=tensor([0.1288, 0.1078, 0.2143, 0.1365, 0.0972, 0.0953, 0.1433, 0.2078], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0239, 0.0137, 0.0120, 0.0133, 0.0152, 0.0116, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:07:07,765 INFO [optim.py:369] (2/7) 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 20:07:22,252 INFO [finetune.py:976] (2/7) Epoch 24, batch 1050, loss[loss=0.1497, simple_loss=0.2186, pruned_loss=0.04043, over 4784.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2428, pruned_loss=0.04808, over 952353.12 frames. ], batch size: 26, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:07:35,145 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132807.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:07:49,843 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0730, 1.2522, 4.8013, 4.5042, 4.1758, 4.5611, 4.3397, 4.2889], + device='cuda:2'), covar=tensor([0.6519, 0.6261, 0.1140, 0.1938, 0.1198, 0.1413, 0.1351, 0.1562], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0306, 0.0404, 0.0405, 0.0346, 0.0408, 0.0314, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:07:56,050 INFO [finetune.py:976] (2/7) Epoch 24, batch 1100, loss[loss=0.1758, simple_loss=0.2522, pruned_loss=0.04968, over 4918.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2452, pruned_loss=0.04883, over 954384.20 frames. ], batch size: 33, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:08:14,644 INFO [optim.py:369] (2/7) 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 20:08:28,712 INFO [finetune.py:976] (2/7) Epoch 24, batch 1150, loss[loss=0.1685, simple_loss=0.2474, pruned_loss=0.04478, over 4828.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2472, pruned_loss=0.05006, over 954507.04 frames. ], batch size: 47, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:08:33,019 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 20:09:11,141 INFO [finetune.py:976] (2/7) Epoch 24, batch 1200, loss[loss=0.1222, simple_loss=0.203, pruned_loss=0.02071, over 4870.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2459, pruned_loss=0.04936, over 955586.87 frames. ], batch size: 31, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:09:48,369 INFO [optim.py:369] (2/7) 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 20:10:17,851 INFO [finetune.py:976] (2/7) Epoch 24, batch 1250, loss[loss=0.1804, simple_loss=0.2442, pruned_loss=0.05831, over 4916.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2431, pruned_loss=0.04895, over 953823.16 frames. ], batch size: 37, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:10:41,414 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-27 20:10:53,171 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133016.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:11:01,198 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9168, 4.3303, 0.9644, 2.2076, 2.7461, 2.9606, 2.6430, 0.9419], + device='cuda:2'), covar=tensor([0.1381, 0.0970, 0.2222, 0.1334, 0.0943, 0.1146, 0.1391, 0.2182], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0242, 0.0139, 0.0122, 0.0134, 0.0154, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:11:17,181 INFO [finetune.py:976] (2/7) Epoch 24, batch 1300, loss[loss=0.1425, simple_loss=0.2191, pruned_loss=0.03298, over 4757.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2396, pruned_loss=0.04776, over 952228.18 frames. ], batch size: 27, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:11:37,391 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0724, 1.9767, 2.1483, 2.5164, 2.5018, 2.0201, 1.6704, 2.1289], + device='cuda:2'), covar=tensor([0.0758, 0.0975, 0.0692, 0.0533, 0.0509, 0.0806, 0.0722, 0.0557], + device='cuda:2'), in_proj_covar=tensor([0.0186, 0.0203, 0.0185, 0.0172, 0.0177, 0.0179, 0.0149, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:11:37,887 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 1350, loss[loss=0.1639, simple_loss=0.2401, pruned_loss=0.04389, over 4821.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2401, pruned_loss=0.04865, over 953124.06 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:12:07,606 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133102.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:12:10,561 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3994, 1.5963, 1.7756, 1.8958, 1.7747, 1.8150, 1.8292, 1.8635], + device='cuda:2'), covar=tensor([0.3708, 0.5569, 0.4342, 0.4066, 0.5291, 0.6890, 0.5285, 0.4316], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0378, 0.0330, 0.0341, 0.0350, 0.0396, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:12:52,452 INFO [finetune.py:976] (2/7) Epoch 24, batch 1400, loss[loss=0.1513, simple_loss=0.2405, pruned_loss=0.03108, over 4845.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2421, pruned_loss=0.04899, over 954817.54 frames. ], batch size: 44, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:13:29,242 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 1450, loss[loss=0.1401, simple_loss=0.2255, pruned_loss=0.0273, over 4801.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2441, pruned_loss=0.04958, over 955400.58 frames. ], batch size: 41, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:13:42,797 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2547, 1.6279, 1.5014, 1.9897, 1.8688, 1.8877, 1.4599, 4.1946], + device='cuda:2'), covar=tensor([0.0548, 0.0795, 0.0793, 0.1143, 0.0642, 0.0648, 0.0736, 0.0106], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 20:14:31,502 INFO [finetune.py:976] (2/7) Epoch 24, batch 1500, loss[loss=0.1762, simple_loss=0.2438, pruned_loss=0.05435, over 4866.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2455, pruned_loss=0.05, over 956165.17 frames. ], batch size: 31, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:15:14,559 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 1550, loss[loss=0.155, simple_loss=0.2348, pruned_loss=0.03765, over 4842.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2456, pruned_loss=0.0494, over 957675.61 frames. ], batch size: 44, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:16:10,961 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133316.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:16:34,615 INFO [finetune.py:976] (2/7) Epoch 24, batch 1600, loss[loss=0.1674, simple_loss=0.2303, pruned_loss=0.05225, over 4733.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2427, pruned_loss=0.04884, over 957568.16 frames. ], batch size: 54, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:16:43,861 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0879, 1.8649, 2.0783, 2.4709, 2.5375, 1.8916, 1.7890, 2.1194], + device='cuda:2'), covar=tensor([0.0786, 0.1127, 0.0711, 0.0580, 0.0520, 0.0926, 0.0718, 0.0592], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0205, 0.0187, 0.0174, 0.0179, 0.0181, 0.0151, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:16:54,213 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0551, 4.0194, 2.9051, 4.6669, 4.0840, 4.0843, 1.8954, 3.9958], + device='cuda:2'), covar=tensor([0.1940, 0.1257, 0.3215, 0.1507, 0.2285, 0.1743, 0.5371, 0.2387], + device='cuda:2'), in_proj_covar=tensor([0.0243, 0.0217, 0.0250, 0.0303, 0.0293, 0.0245, 0.0272, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:17:17,458 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133364.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:17:19,227 INFO [optim.py:369] (2/7) 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 20:17:29,309 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133374.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 20:17:30,575 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133376.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:17:42,834 INFO [finetune.py:976] (2/7) Epoch 24, batch 1650, loss[loss=0.1659, simple_loss=0.2349, pruned_loss=0.04841, over 4770.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2403, pruned_loss=0.048, over 957866.60 frames. ], batch size: 28, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:17:51,565 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 20:18:02,179 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133402.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:18:36,633 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0092, 2.6541, 2.8537, 3.6428, 2.7711, 2.3657, 2.5624, 2.9573], + device='cuda:2'), covar=tensor([0.3028, 0.2755, 0.1595, 0.2298, 0.2735, 0.2666, 0.3303, 0.1735], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0245, 0.0228, 0.0314, 0.0221, 0.0235, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 20:18:44,380 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2312, 1.2373, 3.8112, 3.5154, 3.3702, 3.6695, 3.6753, 3.3570], + device='cuda:2'), covar=tensor([0.7299, 0.5894, 0.1242, 0.2188, 0.1416, 0.1795, 0.1504, 0.1700], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0305, 0.0403, 0.0406, 0.0347, 0.0407, 0.0314, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:18:48,096 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 1700, loss[loss=0.1871, simple_loss=0.2513, pruned_loss=0.06145, over 4907.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2379, pruned_loss=0.04696, over 958482.28 frames. ], batch size: 43, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:18:49,265 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133437.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:19:08,018 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133450.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:19:31,793 INFO [optim.py:369] (2/7) 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 20:19:41,299 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133474.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:19:55,039 INFO [finetune.py:976] (2/7) Epoch 24, batch 1750, loss[loss=0.1852, simple_loss=0.2573, pruned_loss=0.0566, over 4894.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2391, pruned_loss=0.04761, over 958286.90 frames. ], batch size: 35, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:21:01,601 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133535.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:21:08,174 INFO [finetune.py:976] (2/7) Epoch 24, batch 1800, loss[loss=0.1561, simple_loss=0.2355, pruned_loss=0.03829, over 4829.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2416, pruned_loss=0.04838, over 958512.07 frames. ], batch size: 33, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:21:42,021 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8539, 2.3589, 1.8855, 1.7495, 1.3470, 1.3842, 1.9128, 1.3030], + device='cuda:2'), covar=tensor([0.1572, 0.1217, 0.1322, 0.1533, 0.2213, 0.1827, 0.0899, 0.1955], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0211, 0.0169, 0.0205, 0.0200, 0.0186, 0.0157, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 20:21:44,189 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 1850, loss[loss=0.1618, simple_loss=0.2405, pruned_loss=0.04158, over 4871.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2432, pruned_loss=0.04886, over 958245.24 frames. ], batch size: 32, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:22:42,122 INFO [finetune.py:976] (2/7) Epoch 24, batch 1900, loss[loss=0.1682, simple_loss=0.2514, pruned_loss=0.04253, over 4804.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2438, pruned_loss=0.04905, over 954700.92 frames. ], batch size: 40, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:22:51,330 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 20:23:26,868 INFO [finetune.py:976] (2/7) Epoch 24, batch 1950, loss[loss=0.1672, simple_loss=0.2287, pruned_loss=0.05278, over 4782.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.242, pruned_loss=0.04787, over 955914.24 frames. ], batch size: 29, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:23:43,323 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133713.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:23:56,165 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133730.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 20:23:57,388 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133732.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:24:00,817 INFO [finetune.py:976] (2/7) Epoch 24, batch 2000, loss[loss=0.1979, simple_loss=0.2579, pruned_loss=0.06893, over 4748.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2407, pruned_loss=0.04786, over 957945.19 frames. ], batch size: 54, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:24:25,086 INFO [optim.py:369] (2/7) 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 20:24:55,308 INFO [finetune.py:976] (2/7) Epoch 24, batch 2050, loss[loss=0.175, simple_loss=0.2428, pruned_loss=0.05359, over 4919.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2377, pruned_loss=0.04707, over 959944.89 frames. ], batch size: 37, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:25:48,583 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133830.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:25:57,778 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7398, 3.5158, 1.0702, 2.1985, 2.0976, 2.7333, 2.3126, 1.3454], + device='cuda:2'), covar=tensor([0.1103, 0.0911, 0.1827, 0.0972, 0.0958, 0.0800, 0.1181, 0.1901], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0240, 0.0137, 0.0121, 0.0133, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:25:58,914 INFO [finetune.py:976] (2/7) Epoch 24, batch 2100, loss[loss=0.1817, simple_loss=0.2595, pruned_loss=0.05191, over 4803.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2372, pruned_loss=0.04687, over 958835.29 frames. ], batch size: 45, lr: 3.06e-03, grad_scale: 16.0 +2023-04-27 20:26:07,183 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8543, 3.7762, 2.7902, 4.4562, 3.8891, 3.8611, 1.6831, 3.7917], + device='cuda:2'), covar=tensor([0.1615, 0.1308, 0.3093, 0.1695, 0.4201, 0.1837, 0.5938, 0.2334], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0220, 0.0253, 0.0306, 0.0296, 0.0247, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:26:22,908 INFO [optim.py:369] (2/7) 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 20:26:37,513 INFO [finetune.py:976] (2/7) Epoch 24, batch 2150, loss[loss=0.1936, simple_loss=0.2649, pruned_loss=0.06116, over 4935.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2403, pruned_loss=0.04778, over 958882.08 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:26:50,711 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.8685, 3.7985, 3.1155, 4.4812, 3.7856, 3.8229, 1.7639, 3.8909], + device='cuda:2'), covar=tensor([0.1811, 0.1304, 0.3777, 0.1116, 0.2891, 0.1666, 0.5002, 0.1877], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0219, 0.0253, 0.0306, 0.0296, 0.0247, 0.0275, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:27:09,764 INFO [finetune.py:976] (2/7) Epoch 24, batch 2200, loss[loss=0.2208, simple_loss=0.2732, pruned_loss=0.08417, over 4058.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.243, pruned_loss=0.04867, over 956415.77 frames. ], batch size: 65, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:27:48,467 INFO [optim.py:369] (2/7) 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 20:28:06,395 INFO [finetune.py:976] (2/7) Epoch 24, batch 2250, loss[loss=0.183, simple_loss=0.2534, pruned_loss=0.0563, over 4779.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2451, pruned_loss=0.04947, over 954878.72 frames. ], batch size: 29, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:28:06,472 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.1367, 4.1063, 3.0551, 4.7781, 4.1645, 4.1154, 1.6343, 4.0726], + device='cuda:2'), covar=tensor([0.1710, 0.1031, 0.3750, 0.0898, 0.2390, 0.1518, 0.5943, 0.2097], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0218, 0.0252, 0.0306, 0.0295, 0.0247, 0.0274, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:28:06,513 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6991, 2.0661, 0.9632, 1.4290, 2.0509, 1.5670, 1.4761, 1.5573], + device='cuda:2'), covar=tensor([0.0493, 0.0322, 0.0297, 0.0510, 0.0244, 0.0477, 0.0453, 0.0543], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 20:28:14,968 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134008.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:28:36,964 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134030.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 20:28:38,199 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134032.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:28:41,583 INFO [finetune.py:976] (2/7) Epoch 24, batch 2300, loss[loss=0.1996, simple_loss=0.2666, pruned_loss=0.06633, over 4919.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2447, pruned_loss=0.0491, over 954889.76 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:28:55,891 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134058.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:29:01,257 INFO [optim.py:369] (2/7) 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 20:29:02,811 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-27 20:29:08,092 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134078.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 20:29:09,324 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:29:14,453 INFO [finetune.py:976] (2/7) Epoch 24, batch 2350, loss[loss=0.1244, simple_loss=0.1984, pruned_loss=0.02523, over 4744.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2421, pruned_loss=0.0483, over 955394.75 frames. ], batch size: 23, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:29:42,023 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 2400, loss[loss=0.1732, simple_loss=0.2429, pruned_loss=0.05176, over 4897.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2403, pruned_loss=0.04843, over 955871.49 frames. ], batch size: 32, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:30:00,890 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3313, 1.8080, 2.2111, 2.5215, 2.1747, 1.6929, 1.5203, 2.0197], + device='cuda:2'), covar=tensor([0.3127, 0.2920, 0.1575, 0.2342, 0.2509, 0.2665, 0.4195, 0.1944], + device='cuda:2'), in_proj_covar=tensor([0.0296, 0.0246, 0.0230, 0.0316, 0.0222, 0.0236, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 20:30:06,804 INFO [optim.py:369] (2/7) 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 20:30:13,576 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134178.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:30:19,542 INFO [finetune.py:976] (2/7) Epoch 24, batch 2450, loss[loss=0.1973, simple_loss=0.2628, pruned_loss=0.06587, over 4859.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2389, pruned_loss=0.04846, over 955253.91 frames. ], batch size: 44, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:30:29,686 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.22 vs. limit=5.0 +2023-04-27 20:31:07,210 INFO [finetune.py:976] (2/7) Epoch 24, batch 2500, loss[loss=0.125, simple_loss=0.2059, pruned_loss=0.02205, over 4758.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2389, pruned_loss=0.04837, over 953650.86 frames. ], batch size: 27, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:31:39,321 INFO [optim.py:369] (2/7) 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 20:32:00,835 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3549, 1.2862, 1.6605, 1.5816, 1.2209, 1.1622, 1.2920, 0.8874], + device='cuda:2'), covar=tensor([0.0508, 0.0601, 0.0359, 0.0675, 0.0800, 0.1061, 0.0668, 0.0517], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0068, 0.0074, 0.0094, 0.0073, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:32:02,531 INFO [finetune.py:976] (2/7) Epoch 24, batch 2550, loss[loss=0.1929, simple_loss=0.2798, pruned_loss=0.05297, over 4804.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2419, pruned_loss=0.04891, over 954422.32 frames. ], batch size: 41, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:32:14,114 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1994, 1.6127, 1.5425, 2.1968, 2.3855, 1.8401, 1.7554, 1.5656], + device='cuda:2'), covar=tensor([0.2323, 0.2064, 0.2177, 0.1844, 0.1274, 0.2361, 0.2598, 0.2816], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0308, 0.0350, 0.0284, 0.0328, 0.0306, 0.0299, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.3859e-05, 6.3459e-05, 7.3647e-05, 5.6975e-05, 6.7586e-05, 6.4097e-05, + 6.2208e-05, 7.8951e-05], device='cuda:2') +2023-04-27 20:32:26,937 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134304.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:32:34,600 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134308.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:32:35,212 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134309.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:32:58,867 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7245, 3.1323, 1.0495, 1.9223, 2.7096, 1.5996, 4.5134, 2.0512], + device='cuda:2'), covar=tensor([0.0580, 0.0813, 0.0805, 0.1211, 0.0478, 0.0955, 0.0310, 0.0623], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 20:33:08,808 INFO [finetune.py:976] (2/7) Epoch 24, batch 2600, loss[loss=0.1801, simple_loss=0.2578, pruned_loss=0.05117, over 4909.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2428, pruned_loss=0.04851, over 955585.00 frames. ], batch size: 37, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:33:27,028 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134353.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:33:30,514 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134356.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:33:42,155 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134365.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:33:43,227 INFO [optim.py:369] (2/7) 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 20:33:45,769 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 2650, loss[loss=0.1917, simple_loss=0.253, pruned_loss=0.0652, over 4770.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2444, pruned_loss=0.04852, over 958533.31 frames. ], batch size: 26, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:34:16,587 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.53 vs. limit=2.0 +2023-04-27 20:35:02,815 INFO [finetune.py:976] (2/7) Epoch 24, batch 2700, loss[loss=0.1432, simple_loss=0.2197, pruned_loss=0.03332, over 4866.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2436, pruned_loss=0.04777, over 959585.89 frames. ], batch size: 31, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:35:23,806 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 2750, loss[loss=0.1839, simple_loss=0.2518, pruned_loss=0.05806, over 4908.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2416, pruned_loss=0.04763, over 959678.56 frames. ], batch size: 37, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:35:54,516 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0802, 0.7652, 0.9250, 0.8148, 1.1937, 0.9807, 0.8760, 0.9405], + device='cuda:2'), covar=tensor([0.1768, 0.1354, 0.1814, 0.1473, 0.0873, 0.1475, 0.1472, 0.2128], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0304, 0.0347, 0.0281, 0.0325, 0.0303, 0.0296, 0.0369], + device='cuda:2'), 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:2') +2023-04-27 20:36:16,405 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 20:36:43,198 INFO [finetune.py:976] (2/7) Epoch 24, batch 2800, loss[loss=0.1369, simple_loss=0.2081, pruned_loss=0.03288, over 4917.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2378, pruned_loss=0.04639, over 957307.39 frames. ], batch size: 46, lr: 3.05e-03, grad_scale: 16.0 +2023-04-27 20:37:21,564 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7980, 1.7720, 1.7857, 1.4212, 1.7825, 1.6511, 2.3068, 1.6065], + device='cuda:2'), covar=tensor([0.3119, 0.1767, 0.4314, 0.2511, 0.1506, 0.2054, 0.1291, 0.3996], + device='cuda:2'), in_proj_covar=tensor([0.0335, 0.0347, 0.0422, 0.0346, 0.0375, 0.0370, 0.0365, 0.0417], + device='cuda:2'), 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:2') +2023-04-27 20:37:23,293 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 2850, loss[loss=0.2017, simple_loss=0.2794, pruned_loss=0.062, over 4911.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2368, pruned_loss=0.04644, over 956203.93 frames. ], batch size: 37, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:37:55,984 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6093, 1.1040, 1.4106, 1.2196, 1.6619, 1.4288, 1.1009, 1.3439], + device='cuda:2'), covar=tensor([0.1448, 0.1413, 0.1738, 0.1352, 0.0941, 0.1392, 0.1820, 0.2168], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0305, 0.0348, 0.0282, 0.0326, 0.0304, 0.0297, 0.0370], + device='cuda:2'), 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:2') +2023-04-27 20:38:59,967 INFO [finetune.py:976] (2/7) Epoch 24, batch 2900, loss[loss=0.1951, simple_loss=0.269, pruned_loss=0.06059, over 4840.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2393, pruned_loss=0.04723, over 957777.47 frames. ], batch size: 49, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:39:10,240 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 20:39:15,060 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134653.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:39:24,562 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134660.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:39:33,041 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 20:40:04,278 INFO [finetune.py:976] (2/7) Epoch 24, batch 2950, loss[loss=0.1955, simple_loss=0.2836, pruned_loss=0.05369, over 4907.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2421, pruned_loss=0.04746, over 956944.97 frames. ], batch size: 36, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:40:18,489 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134701.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:40:48,699 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.2393, 2.5010, 2.4765, 2.5330, 2.2016, 2.4222, 2.5542, 2.4975], + device='cuda:2'), covar=tensor([0.4087, 0.5374, 0.4313, 0.4139, 0.5688, 0.6721, 0.5319, 0.4803], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0375, 0.0328, 0.0340, 0.0348, 0.0395, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:41:09,548 INFO [finetune.py:976] (2/7) Epoch 24, batch 3000, loss[loss=0.1477, simple_loss=0.2222, pruned_loss=0.03655, over 4732.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2436, pruned_loss=0.04818, over 956033.73 frames. ], batch size: 59, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:41:09,548 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 20:41:25,508 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 20:41:27,955 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4806, 3.4618, 0.9837, 1.8582, 2.0027, 2.6376, 1.9719, 1.0882], + device='cuda:2'), covar=tensor([0.1354, 0.0820, 0.1871, 0.1218, 0.1034, 0.0827, 0.1471, 0.1914], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0238, 0.0135, 0.0120, 0.0132, 0.0151, 0.0116, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:41:44,217 INFO [optim.py:369] (2/7) 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 20:41:47,450 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5818, 1.9646, 2.4865, 3.0014, 2.3933, 1.8466, 1.9234, 2.3246], + device='cuda:2'), covar=tensor([0.2906, 0.3119, 0.1506, 0.2239, 0.2541, 0.2625, 0.3645, 0.2032], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0245, 0.0228, 0.0315, 0.0221, 0.0235, 0.0228, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 20:41:57,440 INFO [finetune.py:976] (2/7) Epoch 24, batch 3050, loss[loss=0.1531, simple_loss=0.2323, pruned_loss=0.03694, over 4864.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2447, pruned_loss=0.04891, over 955921.37 frames. ], batch size: 31, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:42:02,390 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-04-27 20:42:30,088 INFO [finetune.py:976] (2/7) Epoch 24, batch 3100, loss[loss=0.1567, simple_loss=0.2265, pruned_loss=0.0434, over 4745.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2442, pruned_loss=0.04884, over 957278.17 frames. ], batch size: 27, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:42:47,422 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 20:43:02,614 INFO [finetune.py:976] (2/7) Epoch 24, batch 3150, loss[loss=0.1508, simple_loss=0.2172, pruned_loss=0.04223, over 4771.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2408, pruned_loss=0.04835, over 957239.66 frames. ], batch size: 28, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:43:21,320 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5599, 1.4564, 1.8870, 1.9111, 1.3799, 1.2859, 1.5108, 0.9016], + device='cuda:2'), covar=tensor([0.0549, 0.0622, 0.0356, 0.0540, 0.0709, 0.1086, 0.0592, 0.0582], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0067, 0.0074, 0.0095, 0.0073, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:43:23,155 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1606, 2.3784, 1.8463, 1.9437, 2.2037, 1.7945, 2.7826, 1.4928], + device='cuda:2'), covar=tensor([0.3718, 0.1877, 0.4482, 0.2748, 0.1922, 0.2695, 0.1349, 0.4751], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0351, 0.0426, 0.0350, 0.0379, 0.0375, 0.0370, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:43:28,085 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134923.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:43:36,613 INFO [finetune.py:976] (2/7) Epoch 24, batch 3200, loss[loss=0.1664, simple_loss=0.2332, pruned_loss=0.04981, over 4818.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2375, pruned_loss=0.04737, over 954290.40 frames. ], batch size: 41, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:43:53,561 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134960.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:43:54,751 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4617, 1.3425, 4.1925, 3.9170, 3.6913, 4.0263, 3.9892, 3.6072], + device='cuda:2'), covar=tensor([0.7199, 0.5841, 0.1070, 0.1878, 0.1157, 0.1826, 0.1161, 0.1708], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0303, 0.0400, 0.0406, 0.0345, 0.0406, 0.0315, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:43:56,581 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134965.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:43:57,658 INFO [optim.py:369] (2/7) 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 20:44:10,010 INFO [finetune.py:976] (2/7) Epoch 24, batch 3250, loss[loss=0.1681, simple_loss=0.247, pruned_loss=0.04458, over 4913.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.237, pruned_loss=0.04714, over 952579.57 frames. ], batch size: 36, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:44:25,316 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135008.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:44:25,975 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9566, 1.5556, 1.8826, 2.2734, 2.2159, 1.8229, 1.6392, 2.0188], + device='cuda:2'), covar=tensor([0.0824, 0.1215, 0.0776, 0.0544, 0.0602, 0.0872, 0.0762, 0.0566], + device='cuda:2'), in_proj_covar=tensor([0.0187, 0.0204, 0.0186, 0.0173, 0.0178, 0.0179, 0.0150, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:44:28,804 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135013.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:44:30,655 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7696, 2.2127, 1.1682, 1.5240, 2.3183, 1.6228, 1.5507, 1.7418], + device='cuda:2'), covar=tensor([0.0486, 0.0329, 0.0297, 0.0554, 0.0238, 0.0488, 0.0499, 0.0540], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0024, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 20:44:42,366 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 20:44:43,392 INFO [finetune.py:976] (2/7) Epoch 24, batch 3300, loss[loss=0.1643, simple_loss=0.2251, pruned_loss=0.05173, over 4389.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2399, pruned_loss=0.0479, over 953741.10 frames. ], batch size: 19, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:45:15,629 INFO [optim.py:369] (2/7) 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 20:45:44,561 INFO [finetune.py:976] (2/7) Epoch 24, batch 3350, loss[loss=0.2032, simple_loss=0.2733, pruned_loss=0.06658, over 4829.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2434, pruned_loss=0.04899, over 955462.86 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:46:49,143 INFO [finetune.py:976] (2/7) Epoch 24, batch 3400, loss[loss=0.2751, simple_loss=0.3191, pruned_loss=0.1155, over 4883.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2466, pruned_loss=0.05064, over 955454.28 frames. ], batch size: 43, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:47:21,652 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4511, 1.7300, 1.8405, 1.9734, 1.8183, 1.8377, 1.9118, 1.9026], + device='cuda:2'), covar=tensor([0.4193, 0.5033, 0.4331, 0.3777, 0.5304, 0.6920, 0.4794, 0.4663], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0339, 0.0349, 0.0394, 0.0357, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:47:25,515 INFO [optim.py:369] (2/7) 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 20:47:33,950 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:47:42,975 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 20:47:54,155 INFO [finetune.py:976] (2/7) Epoch 24, batch 3450, loss[loss=0.1187, simple_loss=0.1938, pruned_loss=0.02174, over 4783.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2454, pruned_loss=0.04975, over 954802.63 frames. ], batch size: 29, lr: 3.05e-03, grad_scale: 32.0 +2023-04-27 20:48:27,446 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135218.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:48:36,545 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135232.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:48:39,439 INFO [finetune.py:976] (2/7) Epoch 24, batch 3500, loss[loss=0.1702, simple_loss=0.2432, pruned_loss=0.04854, over 4821.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2423, pruned_loss=0.04926, over 954550.20 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:48:45,834 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-04-27 20:48:59,253 INFO [optim.py:369] (2/7) 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 20:49:05,145 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1617, 2.4324, 2.2009, 2.0579, 1.6952, 1.7807, 2.2220, 1.6497], + device='cuda:2'), covar=tensor([0.1580, 0.1359, 0.1341, 0.1460, 0.2177, 0.1764, 0.0857, 0.1911], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0209, 0.0168, 0.0204, 0.0199, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 20:49:13,319 INFO [finetune.py:976] (2/7) Epoch 24, batch 3550, loss[loss=0.1396, simple_loss=0.2128, pruned_loss=0.03316, over 4800.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2386, pruned_loss=0.04778, over 954756.65 frames. ], batch size: 51, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:49:26,787 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5494, 3.4212, 0.8699, 1.8325, 2.0586, 2.4648, 1.9446, 1.0011], + device='cuda:2'), covar=tensor([0.1426, 0.0995, 0.2186, 0.1318, 0.1022, 0.1003, 0.1672, 0.1865], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0235, 0.0134, 0.0119, 0.0130, 0.0150, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:49:47,282 INFO [finetune.py:976] (2/7) Epoch 24, batch 3600, loss[loss=0.1401, simple_loss=0.2223, pruned_loss=0.02897, over 4756.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2366, pruned_loss=0.04699, over 955598.30 frames. ], batch size: 54, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:50:05,962 INFO [optim.py:369] (2/7) 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 20:50:20,184 INFO [finetune.py:976] (2/7) Epoch 24, batch 3650, loss[loss=0.173, simple_loss=0.2468, pruned_loss=0.04961, over 4795.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2393, pruned_loss=0.04837, over 955169.51 frames. ], batch size: 45, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:50:53,968 INFO [finetune.py:976] (2/7) Epoch 24, batch 3700, loss[loss=0.1555, simple_loss=0.2221, pruned_loss=0.04448, over 4081.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2431, pruned_loss=0.04943, over 955111.33 frames. ], batch size: 17, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:50:57,614 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9309, 1.0740, 1.5331, 1.6525, 1.6093, 1.6923, 1.6039, 1.5884], + device='cuda:2'), covar=tensor([0.3733, 0.4725, 0.4201, 0.4221, 0.5186, 0.6762, 0.4465, 0.4298], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0374, 0.0327, 0.0339, 0.0349, 0.0394, 0.0357, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:51:12,481 INFO [optim.py:369] (2/7) 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 20:51:21,301 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4208, 3.4037, 2.6802, 4.0095, 3.4746, 3.3970, 1.5560, 3.4368], + device='cuda:2'), covar=tensor([0.1910, 0.1320, 0.3551, 0.2005, 0.2631, 0.1986, 0.5478, 0.2714], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0219, 0.0252, 0.0307, 0.0296, 0.0249, 0.0275, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:51:27,090 INFO [finetune.py:976] (2/7) Epoch 24, batch 3750, loss[loss=0.153, simple_loss=0.2355, pruned_loss=0.03527, over 4814.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2449, pruned_loss=0.05019, over 953914.23 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:51:44,453 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1183, 2.0408, 1.7040, 1.6667, 2.0864, 1.8288, 2.4389, 1.5711], + device='cuda:2'), covar=tensor([0.3394, 0.1646, 0.4653, 0.2485, 0.1412, 0.2033, 0.1338, 0.4089], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0352, 0.0425, 0.0351, 0.0379, 0.0376, 0.0369, 0.0422], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:51:47,497 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5914, 4.6605, 1.4089, 2.6553, 3.1237, 3.5306, 2.8520, 1.3173], + device='cuda:2'), covar=tensor([0.1082, 0.1196, 0.1850, 0.1136, 0.0817, 0.0826, 0.1341, 0.1897], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0236, 0.0135, 0.0120, 0.0131, 0.0151, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:52:05,312 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135518.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:52:16,495 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135527.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:52:17,873 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 20:52:29,077 INFO [finetune.py:976] (2/7) Epoch 24, batch 3800, loss[loss=0.1611, simple_loss=0.2377, pruned_loss=0.04224, over 4745.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2451, pruned_loss=0.04978, over 952525.14 frames. ], batch size: 59, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:53:08,363 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135566.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:53:08,924 INFO [optim.py:369] (2/7) 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 20:53:32,877 INFO [finetune.py:976] (2/7) Epoch 24, batch 3850, loss[loss=0.1274, simple_loss=0.207, pruned_loss=0.02387, over 4748.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2438, pruned_loss=0.04906, over 951505.77 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:53:56,256 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1368, 1.7834, 2.0457, 2.3638, 2.3635, 1.9737, 1.6234, 2.1625], + device='cuda:2'), covar=tensor([0.0735, 0.1088, 0.0679, 0.0571, 0.0563, 0.0755, 0.0799, 0.0581], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0206, 0.0188, 0.0175, 0.0179, 0.0180, 0.0152, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 20:54:28,831 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135630.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:54:33,841 INFO [finetune.py:976] (2/7) Epoch 24, batch 3900, loss[loss=0.1709, simple_loss=0.2393, pruned_loss=0.05128, over 4752.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2419, pruned_loss=0.04881, over 954433.30 frames. ], batch size: 26, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:55:15,856 INFO [optim.py:369] (2/7) 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 20:55:17,857 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2049, 2.5230, 0.7611, 1.5041, 1.6151, 1.8917, 1.7015, 0.8258], + device='cuda:2'), covar=tensor([0.1452, 0.1442, 0.1889, 0.1300, 0.1167, 0.1046, 0.1533, 0.1849], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0120, 0.0132, 0.0151, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:55:44,888 INFO [finetune.py:976] (2/7) Epoch 24, batch 3950, loss[loss=0.1502, simple_loss=0.2205, pruned_loss=0.03997, over 4830.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2388, pruned_loss=0.04788, over 953969.99 frames. ], batch size: 25, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:55:47,453 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135691.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:56:21,154 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1028, 2.8106, 0.8460, 1.4474, 1.5542, 2.1710, 1.6845, 0.9143], + device='cuda:2'), covar=tensor([0.1862, 0.1844, 0.2308, 0.1826, 0.1417, 0.1301, 0.1847, 0.2170], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0237, 0.0135, 0.0120, 0.0131, 0.0151, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 20:56:44,336 INFO [finetune.py:976] (2/7) Epoch 24, batch 4000, loss[loss=0.2328, simple_loss=0.2876, pruned_loss=0.08902, over 4907.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2384, pruned_loss=0.0477, over 954820.19 frames. ], batch size: 36, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:57:22,843 INFO [optim.py:369] (2/7) 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 20:57:52,046 INFO [finetune.py:976] (2/7) Epoch 24, batch 4050, loss[loss=0.1806, simple_loss=0.2512, pruned_loss=0.05501, over 4756.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2416, pruned_loss=0.04876, over 953221.12 frames. ], batch size: 28, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:58:05,438 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4547, 3.4218, 2.5497, 4.0186, 3.4916, 3.4138, 1.6083, 3.4194], + device='cuda:2'), covar=tensor([0.2039, 0.1329, 0.3814, 0.1990, 0.3147, 0.1884, 0.5722, 0.2497], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0219, 0.0253, 0.0307, 0.0295, 0.0248, 0.0274, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:58:36,237 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.3424, 4.2752, 3.0080, 4.9797, 4.2612, 4.3250, 1.7161, 4.2457], + device='cuda:2'), covar=tensor([0.1518, 0.1073, 0.3662, 0.0973, 0.3078, 0.1588, 0.5888, 0.2216], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0220, 0.0253, 0.0307, 0.0296, 0.0249, 0.0275, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 20:58:39,892 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135827.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:58:52,538 INFO [finetune.py:976] (2/7) Epoch 24, batch 4100, loss[loss=0.1547, simple_loss=0.2294, pruned_loss=0.04001, over 4754.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2434, pruned_loss=0.04895, over 953318.06 frames. ], batch size: 28, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 20:59:28,241 INFO [optim.py:369] (2/7) 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 20:59:33,114 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135875.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 20:59:40,401 INFO [finetune.py:976] (2/7) Epoch 24, batch 4150, loss[loss=0.1967, simple_loss=0.273, pruned_loss=0.06024, over 4915.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.245, pruned_loss=0.05016, over 948950.70 frames. ], batch size: 46, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:00:14,127 INFO [finetune.py:976] (2/7) Epoch 24, batch 4200, loss[loss=0.1135, simple_loss=0.1926, pruned_loss=0.01717, over 4749.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2451, pruned_loss=0.0499, over 950220.02 frames. ], batch size: 27, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:00:35,175 INFO [optim.py:369] (2/7) 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 21:00:42,964 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9618, 1.1256, 1.6111, 1.7327, 1.6790, 1.7436, 1.6460, 1.6182], + device='cuda:2'), covar=tensor([0.3522, 0.4811, 0.3802, 0.4021, 0.4926, 0.6547, 0.4010, 0.4077], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0375, 0.0327, 0.0341, 0.0349, 0.0396, 0.0358, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:00:44,768 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135982.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:00:46,622 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8759, 2.4399, 1.9499, 1.9793, 1.5250, 1.4873, 1.9528, 1.4593], + device='cuda:2'), covar=tensor([0.1238, 0.1109, 0.1115, 0.1311, 0.1848, 0.1594, 0.0807, 0.1703], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0210, 0.0169, 0.0205, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 21:00:47,169 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 4250, loss[loss=0.2197, simple_loss=0.2824, pruned_loss=0.07848, over 4852.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2439, pruned_loss=0.04979, over 951091.73 frames. ], batch size: 47, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:00:52,111 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135994.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:01:22,183 INFO [finetune.py:976] (2/7) Epoch 24, batch 4300, loss[loss=0.1376, simple_loss=0.2167, pruned_loss=0.02923, over 4827.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2414, pruned_loss=0.04888, over 948862.30 frames. ], batch size: 41, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:01:23,719 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 21:01:25,995 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136043.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:01:34,853 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 21:01:55,533 INFO [finetune.py:976] (2/7) Epoch 24, batch 4350, loss[loss=0.1771, simple_loss=0.2452, pruned_loss=0.05449, over 4853.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2395, pruned_loss=0.04802, over 950088.37 frames. ], batch size: 44, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:02:33,974 INFO [finetune.py:976] (2/7) Epoch 24, batch 4400, loss[loss=0.1875, simple_loss=0.2616, pruned_loss=0.05667, over 4904.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2403, pruned_loss=0.04807, over 951010.18 frames. ], batch size: 37, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:02:42,581 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136142.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:02:45,226 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-27 21:03:16,195 INFO [optim.py:369] (2/7) 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 21:03:40,891 INFO [finetune.py:976] (2/7) Epoch 24, batch 4450, loss[loss=0.2225, simple_loss=0.2863, pruned_loss=0.07936, over 4820.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2457, pruned_loss=0.05029, over 951503.69 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:04:02,257 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136203.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:04:53,815 INFO [finetune.py:976] (2/7) Epoch 24, batch 4500, loss[loss=0.1939, simple_loss=0.261, pruned_loss=0.06339, over 4748.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2471, pruned_loss=0.0503, over 952902.14 frames. ], batch size: 27, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:05:29,627 INFO [optim.py:369] (2/7) 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 21:05:58,010 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 4550, loss[loss=0.1917, simple_loss=0.2611, pruned_loss=0.06114, over 4816.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2461, pruned_loss=0.04955, over 953748.30 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:06:24,458 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5761, 1.5913, 0.6915, 1.2627, 1.6324, 1.4433, 1.3672, 1.4229], + device='cuda:2'), covar=tensor([0.0454, 0.0337, 0.0350, 0.0520, 0.0277, 0.0453, 0.0453, 0.0502], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 21:06:47,092 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9086, 2.4444, 0.9625, 1.3842, 1.8768, 1.2492, 3.3424, 1.8779], + device='cuda:2'), covar=tensor([0.0876, 0.0794, 0.1008, 0.1753, 0.0675, 0.1337, 0.0295, 0.0785], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0048, 0.0050, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0007, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 21:06:56,737 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136334.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:06:57,426 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 4600, loss[loss=0.1879, simple_loss=0.2553, pruned_loss=0.06026, over 4819.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2433, pruned_loss=0.04839, over 953164.88 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:07:04,992 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136338.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:07:18,413 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136350.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:07:26,773 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 21:07:40,422 INFO [optim.py:369] (2/7) 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 21:08:03,892 INFO [finetune.py:976] (2/7) Epoch 24, batch 4650, loss[loss=0.1797, simple_loss=0.2397, pruned_loss=0.05979, over 4834.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2412, pruned_loss=0.04812, over 953997.88 frames. ], batch size: 30, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:08:20,807 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136396.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:08:26,273 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6782, 1.8069, 0.8761, 1.3799, 1.9477, 1.5162, 1.4161, 1.5930], + device='cuda:2'), covar=tensor([0.0504, 0.0361, 0.0324, 0.0548, 0.0250, 0.0501, 0.0509, 0.0572], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 21:09:13,810 INFO [finetune.py:976] (2/7) Epoch 24, batch 4700, loss[loss=0.1366, simple_loss=0.211, pruned_loss=0.03108, over 4808.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2386, pruned_loss=0.04775, over 953884.71 frames. ], batch size: 51, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:09:36,611 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0178, 1.4910, 1.6323, 1.7356, 2.1618, 1.8411, 1.5292, 1.4770], + device='cuda:2'), covar=tensor([0.1722, 0.1598, 0.2023, 0.1406, 0.0914, 0.1597, 0.2090, 0.2506], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0308, 0.0351, 0.0285, 0.0327, 0.0305, 0.0298, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4114e-05, 6.3416e-05, 7.3925e-05, 5.7272e-05, 6.7194e-05, 6.3951e-05, + 6.1988e-05, 7.9100e-05], device='cuda:2') +2023-04-27 21:09:45,097 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 4750, loss[loss=0.1949, simple_loss=0.2479, pruned_loss=0.07098, over 4200.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2359, pruned_loss=0.04682, over 952878.04 frames. ], batch size: 65, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:10:07,141 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136498.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:10:47,890 INFO [finetune.py:976] (2/7) Epoch 24, batch 4800, loss[loss=0.1596, simple_loss=0.2429, pruned_loss=0.03809, over 4756.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.24, pruned_loss=0.04869, over 952054.30 frames. ], batch size: 26, lr: 3.04e-03, grad_scale: 32.0 +2023-04-27 21:11:29,602 INFO [optim.py:369] (2/7) 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 21:11:44,086 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1686, 1.5060, 1.3510, 1.6680, 1.6113, 1.8064, 1.4075, 3.5036], + device='cuda:2'), covar=tensor([0.0645, 0.0858, 0.0864, 0.1244, 0.0659, 0.0633, 0.0762, 0.0146], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 21:11:53,501 INFO [finetune.py:976] (2/7) Epoch 24, batch 4850, loss[loss=0.2028, simple_loss=0.2776, pruned_loss=0.064, over 4717.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2431, pruned_loss=0.04898, over 953467.06 frames. ], batch size: 59, lr: 3.04e-03, grad_scale: 64.0 +2023-04-27 21:12:26,524 INFO [finetune.py:976] (2/7) Epoch 24, batch 4900, loss[loss=0.2075, simple_loss=0.2771, pruned_loss=0.06892, over 4213.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2446, pruned_loss=0.04962, over 951772.00 frames. ], batch size: 65, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:12:27,225 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136638.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:12:30,277 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:12:35,619 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136650.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:12:38,099 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3138, 1.6398, 1.9117, 1.9572, 1.8023, 1.8463, 1.9247, 1.9418], + device='cuda:2'), covar=tensor([0.4155, 0.5206, 0.3832, 0.4364, 0.5349, 0.6781, 0.4809, 0.4289], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0371, 0.0323, 0.0337, 0.0345, 0.0392, 0.0355, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:12:46,890 INFO [optim.py:369] (2/7) 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 21:12:47,016 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8371, 1.7259, 2.1956, 2.1801, 1.6838, 1.4785, 1.8257, 1.0243], + device='cuda:2'), covar=tensor([0.0584, 0.0692, 0.0409, 0.0817, 0.0722, 0.1041, 0.0613, 0.0657], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0075, 0.0095, 0.0073, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 21:12:58,275 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 4950, loss[loss=0.1806, simple_loss=0.2553, pruned_loss=0.05295, over 4830.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2456, pruned_loss=0.0499, over 952529.21 frames. ], batch size: 30, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:13:02,152 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136691.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:13:07,455 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136698.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:13:09,257 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1877, 1.5447, 1.4475, 1.7736, 1.6232, 1.9443, 1.4132, 3.7119], + device='cuda:2'), covar=tensor([0.0599, 0.0825, 0.0756, 0.1153, 0.0627, 0.0567, 0.0749, 0.0127], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0038, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 21:13:11,646 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:13:25,474 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-27 21:13:32,451 INFO [finetune.py:976] (2/7) Epoch 24, batch 5000, loss[loss=0.1407, simple_loss=0.2187, pruned_loss=0.03138, over 4906.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2438, pruned_loss=0.04924, over 953385.64 frames. ], batch size: 37, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:13:53,647 INFO [optim.py:369] (2/7) 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 21:14:05,878 INFO [finetune.py:976] (2/7) Epoch 24, batch 5050, loss[loss=0.1311, simple_loss=0.2018, pruned_loss=0.03025, over 4736.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2412, pruned_loss=0.04848, over 954702.71 frames. ], batch size: 54, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:14:18,296 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-27 21:14:18,798 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136798.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:14:29,847 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 5100, loss[loss=0.1867, simple_loss=0.2493, pruned_loss=0.06209, over 4905.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2393, pruned_loss=0.04834, over 954603.43 frames. ], batch size: 43, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:15:07,737 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136846.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:15:22,628 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 24, batch 5150, loss[loss=0.2276, simple_loss=0.3103, pruned_loss=0.07248, over 4218.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2388, pruned_loss=0.04817, over 952138.73 frames. ], batch size: 65, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:16:29,340 INFO [finetune.py:976] (2/7) Epoch 24, batch 5200, loss[loss=0.1731, simple_loss=0.2384, pruned_loss=0.05392, over 4896.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2425, pruned_loss=0.04902, over 952416.47 frames. ], batch size: 32, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:16:29,586 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 21:16:29,592 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-04-27 21:16:59,706 INFO [optim.py:369] (2/7) 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 21:17:22,716 INFO [finetune.py:976] (2/7) Epoch 24, batch 5250, loss[loss=0.1537, simple_loss=0.2147, pruned_loss=0.04635, over 4340.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2443, pruned_loss=0.04905, over 953413.52 frames. ], batch size: 19, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:17:25,238 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136991.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:17:36,059 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136999.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:18:10,122 INFO [finetune.py:976] (2/7) Epoch 24, batch 5300, loss[loss=0.1861, simple_loss=0.2643, pruned_loss=0.05391, over 4835.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2453, pruned_loss=0.04869, over 954768.91 frames. ], batch size: 49, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:18:11,866 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137039.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:18:28,031 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9571, 1.2276, 3.1452, 2.9222, 2.8305, 3.0803, 3.0567, 2.8022], + device='cuda:2'), covar=tensor([0.6597, 0.4948, 0.1408, 0.1898, 0.1370, 0.1895, 0.1320, 0.1836], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0305, 0.0403, 0.0404, 0.0347, 0.0406, 0.0318, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:18:29,854 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3207, 2.1988, 1.8971, 1.9318, 2.3931, 1.9622, 2.9057, 1.7934], + device='cuda:2'), covar=tensor([0.3338, 0.1962, 0.4592, 0.2985, 0.1625, 0.2512, 0.1277, 0.4085], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0351, 0.0423, 0.0348, 0.0375, 0.0373, 0.0364, 0.0419], + device='cuda:2'), out_proj_covar=tensor([9.9710e-05, 1.0485e-04, 1.2806e-04, 1.0452e-04, 1.1125e-04, 1.1111e-04, + 1.0698e-04, 1.2614e-04], device='cuda:2') +2023-04-27 21:18:30,931 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 24, batch 5350, loss[loss=0.1804, simple_loss=0.2494, pruned_loss=0.05573, over 4910.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2462, pruned_loss=0.04875, over 956343.02 frames. ], batch size: 37, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:18:48,947 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2456, 2.8837, 2.4799, 2.7300, 2.0164, 2.3872, 2.4720, 1.8818], + device='cuda:2'), covar=tensor([0.2250, 0.1279, 0.0827, 0.1269, 0.3400, 0.1171, 0.1981, 0.2991], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0297, 0.0215, 0.0275, 0.0311, 0.0254, 0.0249, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1294e-04, 1.1725e-04, 8.4512e-05, 1.0850e-04, 1.2546e-04, 1.0011e-04, + 1.0031e-04, 1.0381e-04], device='cuda:2') +2023-04-27 21:19:16,914 INFO [finetune.py:976] (2/7) Epoch 24, batch 5400, loss[loss=0.1769, simple_loss=0.2496, pruned_loss=0.0521, over 4818.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2436, pruned_loss=0.04834, over 955918.18 frames. ], batch size: 33, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:19:33,100 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=137162.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:19:37,581 INFO [optim.py:369] (2/7) 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 21:19:50,713 INFO [finetune.py:976] (2/7) Epoch 24, batch 5450, loss[loss=0.1501, simple_loss=0.2103, pruned_loss=0.04493, over 4825.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.241, pruned_loss=0.04782, over 955026.98 frames. ], batch size: 30, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:19:56,289 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8630, 2.7351, 2.1564, 3.2785, 2.8741, 2.8530, 1.1998, 2.7453], + device='cuda:2'), covar=tensor([0.2329, 0.1936, 0.3940, 0.3256, 0.3274, 0.2274, 0.6247, 0.3164], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0221, 0.0253, 0.0308, 0.0299, 0.0248, 0.0277, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:20:24,724 INFO [finetune.py:976] (2/7) Epoch 24, batch 5500, loss[loss=0.2201, simple_loss=0.2877, pruned_loss=0.07622, over 4818.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2384, pruned_loss=0.04714, over 953713.32 frames. ], batch size: 38, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:20:44,122 INFO [optim.py:369] (2/7) 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 21:20:52,453 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8676, 2.4030, 1.8535, 1.6724, 1.3796, 1.4085, 1.8866, 1.3249], + device='cuda:2'), covar=tensor([0.1778, 0.1297, 0.1488, 0.1804, 0.2350, 0.2022, 0.1036, 0.2119], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0211, 0.0168, 0.0204, 0.0199, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 21:20:57,634 INFO [finetune.py:976] (2/7) Epoch 24, batch 5550, loss[loss=0.1635, simple_loss=0.2499, pruned_loss=0.03852, over 4790.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2408, pruned_loss=0.04807, over 955078.67 frames. ], batch size: 51, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:21:05,046 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 24, batch 5600, loss[loss=0.154, simple_loss=0.2313, pruned_loss=0.0384, over 4846.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2433, pruned_loss=0.04869, over 955771.10 frames. ], batch size: 47, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:21:46,034 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137347.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:22:14,767 INFO [optim.py:369] (2/7) 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 21:22:37,496 INFO [finetune.py:976] (2/7) Epoch 24, batch 5650, loss[loss=0.1712, simple_loss=0.2528, pruned_loss=0.04481, over 4793.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2454, pruned_loss=0.049, over 953286.56 frames. ], batch size: 51, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:23:13,097 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.2675, 3.7781, 3.3854, 3.6303, 2.9630, 3.2370, 3.4164, 2.7808], + device='cuda:2'), covar=tensor([0.1241, 0.0740, 0.0506, 0.0646, 0.2288, 0.0796, 0.1336, 0.1976], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0297, 0.0214, 0.0275, 0.0311, 0.0253, 0.0248, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1308e-04, 1.1747e-04, 8.4147e-05, 1.0827e-04, 1.2540e-04, 9.9774e-05, + 1.0017e-04, 1.0379e-04], device='cuda:2') +2023-04-27 21:23:23,669 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-27 21:23:28,790 INFO [finetune.py:976] (2/7) Epoch 24, batch 5700, loss[loss=0.1399, simple_loss=0.2044, pruned_loss=0.03776, over 4217.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2421, pruned_loss=0.04812, over 939674.72 frames. ], batch size: 18, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:23:36,087 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-27 21:23:37,743 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8168, 1.1177, 2.8411, 2.6747, 2.5681, 2.8020, 2.7871, 2.4570], + device='cuda:2'), covar=tensor([0.6332, 0.4398, 0.1314, 0.1787, 0.1221, 0.1572, 0.1249, 0.1543], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0304, 0.0404, 0.0406, 0.0347, 0.0407, 0.0317, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:23:42,557 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.02 vs. limit=5.0 +2023-04-27 21:23:43,839 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=137462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:23:57,746 INFO [finetune.py:976] (2/7) Epoch 25, batch 0, loss[loss=0.1413, simple_loss=0.2189, pruned_loss=0.03184, over 4771.00 frames. ], tot_loss[loss=0.1413, simple_loss=0.2189, pruned_loss=0.03184, over 4771.00 frames. ], batch size: 26, lr: 3.03e-03, grad_scale: 32.0 +2023-04-27 21:23:57,746 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 21:24:03,951 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4761, 3.5343, 2.5518, 3.9415, 3.6195, 3.5463, 1.5771, 3.5424], + device='cuda:2'), covar=tensor([0.1745, 0.1315, 0.3086, 0.2009, 0.2366, 0.1732, 0.5182, 0.2256], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0220, 0.0253, 0.0308, 0.0299, 0.0248, 0.0277, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:24:08,111 INFO [finetune.py:1010] (2/7) Epoch 25, validation: loss=0.155, simple_loss=0.224, pruned_loss=0.04295, over 2265189.00 frames. +2023-04-27 21:24:08,111 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 21:24:09,923 INFO [optim.py:369] (2/7) 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 21:24:37,392 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137510.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:24:40,447 INFO [finetune.py:976] (2/7) Epoch 25, batch 50, loss[loss=0.1739, simple_loss=0.2533, pruned_loss=0.04728, over 4812.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2523, pruned_loss=0.05427, over 217232.05 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:24:42,242 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7847, 2.0345, 1.2004, 1.5425, 2.2489, 1.6263, 1.5882, 1.7560], + device='cuda:2'), covar=tensor([0.0440, 0.0320, 0.0264, 0.0515, 0.0226, 0.0471, 0.0451, 0.0498], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 21:24:44,133 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-04-27 21:24:55,461 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0718, 2.5387, 1.0732, 1.3409, 1.9446, 1.1852, 3.2168, 1.5693], + device='cuda:2'), covar=tensor([0.0639, 0.0495, 0.0709, 0.1216, 0.0460, 0.1017, 0.0209, 0.0623], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0048, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0009, 0.0007, 0.0007, 0.0007, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 21:24:57,772 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7855, 4.3131, 0.8637, 2.4988, 2.5079, 2.9564, 2.5492, 1.1019], + device='cuda:2'), covar=tensor([0.1550, 0.1303, 0.2443, 0.1235, 0.1061, 0.1171, 0.1488, 0.2104], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0236, 0.0135, 0.0119, 0.0131, 0.0151, 0.0117, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 21:24:59,601 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4321, 1.8871, 1.6657, 2.3043, 2.4426, 2.0700, 2.0460, 1.6286], + device='cuda:2'), covar=tensor([0.1708, 0.1568, 0.1977, 0.1557, 0.1252, 0.1569, 0.1914, 0.2500], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0304, 0.0348, 0.0282, 0.0324, 0.0302, 0.0296, 0.0369], + device='cuda:2'), out_proj_covar=tensor([6.3359e-05, 6.2663e-05, 7.3322e-05, 5.6520e-05, 6.6681e-05, 6.3163e-05, + 6.1477e-05, 7.8147e-05], device='cuda:2') +2023-04-27 21:25:13,440 INFO [finetune.py:976] (2/7) Epoch 25, batch 100, loss[loss=0.2007, simple_loss=0.2556, pruned_loss=0.07292, over 4822.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2414, pruned_loss=0.04979, over 383049.85 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:25:15,237 INFO [optim.py:369] (2/7) 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 21:25:46,402 INFO [finetune.py:976] (2/7) Epoch 25, batch 150, loss[loss=0.1871, simple_loss=0.2586, pruned_loss=0.05782, over 4830.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2372, pruned_loss=0.04822, over 510079.98 frames. ], batch size: 39, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:25:55,298 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 21:26:20,276 INFO [finetune.py:976] (2/7) Epoch 25, batch 200, loss[loss=0.172, simple_loss=0.2367, pruned_loss=0.05364, over 4888.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2359, pruned_loss=0.04723, over 609693.79 frames. ], batch size: 35, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:26:22,074 INFO [optim.py:369] (2/7) 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 21:26:33,043 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 21:27:09,383 INFO [finetune.py:976] (2/7) Epoch 25, batch 250, loss[loss=0.1878, simple_loss=0.2554, pruned_loss=0.06014, over 4934.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2398, pruned_loss=0.04873, over 687151.61 frames. ], batch size: 33, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:27:19,172 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9888, 2.2297, 2.1785, 2.3274, 2.0858, 2.2774, 2.2245, 2.1479], + device='cuda:2'), covar=tensor([0.4134, 0.5665, 0.4449, 0.4160, 0.5365, 0.6523, 0.5942, 0.5385], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0372, 0.0324, 0.0338, 0.0348, 0.0392, 0.0356, 0.0329], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:28:04,796 INFO [finetune.py:976] (2/7) Epoch 25, batch 300, loss[loss=0.1531, simple_loss=0.2273, pruned_loss=0.03945, over 4193.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2412, pruned_loss=0.04858, over 746180.45 frames. ], batch size: 18, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:28:05,572 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7787, 0.7841, 1.6167, 2.1065, 1.8449, 1.6572, 1.6824, 1.6507], + device='cuda:2'), covar=tensor([0.4183, 0.6480, 0.5589, 0.5638, 0.5614, 0.7119, 0.7817, 0.7176], + device='cuda:2'), in_proj_covar=tensor([0.0436, 0.0420, 0.0512, 0.0508, 0.0466, 0.0500, 0.0503, 0.0516], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:28:06,619 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 25, batch 350, loss[loss=0.1729, simple_loss=0.2503, pruned_loss=0.0478, over 4892.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2438, pruned_loss=0.04952, over 794395.35 frames. ], batch size: 43, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:29:09,277 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-04-27 21:29:21,924 INFO [finetune.py:976] (2/7) Epoch 25, batch 400, loss[loss=0.1457, simple_loss=0.2262, pruned_loss=0.03258, over 4806.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2442, pruned_loss=0.0488, over 829674.13 frames. ], batch size: 45, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:29:29,038 INFO [optim.py:369] (2/7) 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 21:30:01,120 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137892.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:30:02,926 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137895.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:30:27,175 INFO [finetune.py:976] (2/7) Epoch 25, batch 450, loss[loss=0.1567, simple_loss=0.2429, pruned_loss=0.03522, over 4747.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2437, pruned_loss=0.04897, over 858215.69 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:30:28,477 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0398, 2.3172, 1.0619, 1.4162, 1.8515, 1.2336, 2.9690, 1.5568], + device='cuda:2'), covar=tensor([0.0640, 0.0527, 0.0671, 0.1162, 0.0423, 0.0940, 0.0226, 0.0606], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0048, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0007, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 21:31:27,310 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137953.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:31:33,956 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137956.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:31:44,826 INFO [finetune.py:976] (2/7) Epoch 25, batch 500, loss[loss=0.2097, simple_loss=0.2694, pruned_loss=0.07503, over 4904.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2413, pruned_loss=0.04775, over 881171.21 frames. ], batch size: 36, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:31:46,606 INFO [optim.py:369] (2/7) 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 21:32:25,642 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-27 21:32:27,211 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3535, 1.2760, 1.6139, 1.6213, 1.2557, 1.1903, 1.3799, 0.8870], + device='cuda:2'), covar=tensor([0.0523, 0.0579, 0.0349, 0.0484, 0.0793, 0.1099, 0.0436, 0.0516], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0068, 0.0075, 0.0096, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 21:32:30,093 INFO [finetune.py:976] (2/7) Epoch 25, batch 550, loss[loss=0.1673, simple_loss=0.2314, pruned_loss=0.05154, over 4842.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2387, pruned_loss=0.04716, over 899170.92 frames. ], batch size: 49, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:33:20,384 INFO [finetune.py:976] (2/7) Epoch 25, batch 600, loss[loss=0.2016, simple_loss=0.2804, pruned_loss=0.06144, over 4856.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2389, pruned_loss=0.04793, over 910251.40 frames. ], batch size: 47, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:33:22,208 INFO [optim.py:369] (2/7) 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 21:33:47,242 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2340, 1.8345, 2.1082, 2.6866, 2.2758, 1.7591, 1.7915, 2.0497], + device='cuda:2'), covar=tensor([0.2663, 0.2804, 0.1473, 0.2155, 0.2325, 0.2403, 0.3615, 0.2036], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0246, 0.0228, 0.0314, 0.0221, 0.0234, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 21:33:53,155 INFO [finetune.py:976] (2/7) Epoch 25, batch 650, loss[loss=0.151, simple_loss=0.2336, pruned_loss=0.03418, over 4838.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2422, pruned_loss=0.04908, over 917988.18 frames. ], batch size: 47, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:34:02,643 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6421, 1.3905, 4.3508, 4.0841, 3.7699, 4.1742, 4.0097, 3.7739], + device='cuda:2'), covar=tensor([0.7238, 0.6170, 0.1149, 0.1777, 0.1242, 0.2127, 0.1792, 0.1612], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0307, 0.0407, 0.0409, 0.0348, 0.0411, 0.0319, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:34:26,502 INFO [finetune.py:976] (2/7) Epoch 25, batch 700, loss[loss=0.2015, simple_loss=0.2709, pruned_loss=0.06607, over 4817.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2438, pruned_loss=0.04906, over 927304.94 frames. ], batch size: 51, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:34:28,314 INFO [optim.py:369] (2/7) 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 21:34:55,457 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.68 vs. limit=5.0 +2023-04-27 21:35:26,623 INFO [finetune.py:976] (2/7) Epoch 25, batch 750, loss[loss=0.1769, simple_loss=0.2512, pruned_loss=0.0513, over 4827.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2432, pruned_loss=0.04814, over 931301.43 frames. ], batch size: 47, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:35:58,716 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4019, 3.3876, 0.8215, 1.7871, 1.7947, 2.3713, 1.9442, 1.0221], + device='cuda:2'), covar=tensor([0.1404, 0.1012, 0.2132, 0.1267, 0.1100, 0.1063, 0.1397, 0.1984], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0120, 0.0132, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 21:36:09,233 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138248.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:36:16,833 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138251.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:36:20,981 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4056, 2.0626, 2.2794, 2.9315, 2.4520, 1.9727, 2.0330, 2.2798], + device='cuda:2'), covar=tensor([0.2469, 0.2447, 0.1383, 0.1789, 0.2016, 0.2135, 0.3239, 0.1713], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0247, 0.0229, 0.0315, 0.0222, 0.0235, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 21:36:24,038 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2267, 1.4804, 1.7742, 1.9001, 1.8379, 1.9157, 1.7995, 1.8144], + device='cuda:2'), covar=tensor([0.4232, 0.5051, 0.4588, 0.4714, 0.5411, 0.6583, 0.5203, 0.4613], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0375, 0.0327, 0.0341, 0.0350, 0.0395, 0.0359, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:36:25,762 INFO [finetune.py:976] (2/7) Epoch 25, batch 800, loss[loss=0.1729, simple_loss=0.2562, pruned_loss=0.04474, over 4726.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2439, pruned_loss=0.04822, over 937009.98 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:36:27,566 INFO [optim.py:369] (2/7) 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 21:36:28,888 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138270.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:36:47,452 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 21:37:07,632 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0549, 2.4624, 2.4966, 2.7872, 2.6716, 2.6148, 2.3940, 4.9071], + device='cuda:2'), covar=tensor([0.0446, 0.0624, 0.0654, 0.0961, 0.0494, 0.0385, 0.0556, 0.0095], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0038, 0.0037, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0012, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 21:37:09,965 INFO [finetune.py:976] (2/7) Epoch 25, batch 850, loss[loss=0.1887, simple_loss=0.2464, pruned_loss=0.06545, over 4776.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2434, pruned_loss=0.04877, over 941267.95 frames. ], batch size: 51, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:37:14,325 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138322.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:37:16,895 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-04-27 21:37:19,551 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7557, 1.9213, 0.9744, 1.5112, 2.2452, 1.6432, 1.5760, 1.6491], + device='cuda:2'), covar=tensor([0.0470, 0.0326, 0.0286, 0.0503, 0.0215, 0.0464, 0.0444, 0.0523], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 21:37:24,954 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138331.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:37:50,342 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-27 21:38:00,618 INFO [finetune.py:976] (2/7) Epoch 25, batch 900, loss[loss=0.2159, simple_loss=0.2781, pruned_loss=0.07685, over 4739.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2409, pruned_loss=0.04818, over 945545.13 frames. ], batch size: 26, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:38:02,478 INFO [optim.py:369] (2/7) 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 21:38:21,991 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138383.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:38:35,742 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 21:39:01,961 INFO [finetune.py:976] (2/7) Epoch 25, batch 950, loss[loss=0.219, simple_loss=0.2883, pruned_loss=0.07489, over 4901.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2394, pruned_loss=0.04817, over 948038.85 frames. ], batch size: 43, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:39:25,185 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2004, 1.4760, 1.4068, 1.6565, 1.5714, 1.8010, 1.3550, 3.1839], + device='cuda:2'), covar=tensor([0.0611, 0.0765, 0.0780, 0.1204, 0.0623, 0.0554, 0.0744, 0.0162], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 21:39:32,171 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 21:39:34,922 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7541, 1.3227, 1.4132, 1.5523, 1.9386, 1.6203, 1.4017, 1.3481], + device='cuda:2'), covar=tensor([0.1664, 0.1675, 0.1981, 0.1442, 0.0855, 0.1659, 0.1997, 0.2362], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0305, 0.0348, 0.0281, 0.0323, 0.0302, 0.0296, 0.0368], + device='cuda:2'), out_proj_covar=tensor([6.3368e-05, 6.2738e-05, 7.3247e-05, 5.6388e-05, 6.6212e-05, 6.3239e-05, + 6.1524e-05, 7.8067e-05], device='cuda:2') +2023-04-27 21:40:06,147 INFO [finetune.py:976] (2/7) Epoch 25, batch 1000, loss[loss=0.1696, simple_loss=0.2389, pruned_loss=0.05013, over 4837.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2397, pruned_loss=0.04763, over 948314.27 frames. ], batch size: 30, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:40:07,976 INFO [optim.py:369] (2/7) 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 21:40:16,465 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3678, 1.8543, 2.2886, 2.5881, 2.2664, 1.7778, 1.4285, 2.0476], + device='cuda:2'), covar=tensor([0.3309, 0.3057, 0.1770, 0.2312, 0.2714, 0.2744, 0.4039, 0.1789], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0246, 0.0229, 0.0314, 0.0223, 0.0235, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 21:41:09,819 INFO [finetune.py:976] (2/7) Epoch 25, batch 1050, loss[loss=0.1833, simple_loss=0.2523, pruned_loss=0.05711, over 4898.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2396, pruned_loss=0.04677, over 948004.54 frames. ], batch size: 35, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:41:52,261 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138548.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:41:54,996 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138551.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:42:14,666 INFO [finetune.py:976] (2/7) Epoch 25, batch 1100, loss[loss=0.1748, simple_loss=0.2504, pruned_loss=0.04963, over 4718.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2413, pruned_loss=0.04779, over 948555.87 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:42:16,467 INFO [optim.py:369] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138599.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:43:18,598 INFO [finetune.py:976] (2/7) Epoch 25, batch 1150, loss[loss=0.1877, simple_loss=0.2637, pruned_loss=0.05583, over 4819.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2423, pruned_loss=0.04808, over 950104.29 frames. ], batch size: 39, lr: 3.02e-03, grad_scale: 32.0 +2023-04-27 21:43:38,297 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138626.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:44:23,367 INFO [finetune.py:976] (2/7) Epoch 25, batch 1200, loss[loss=0.1506, simple_loss=0.2213, pruned_loss=0.03991, over 4837.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2411, pruned_loss=0.04752, over 949913.06 frames. ], batch size: 47, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:44:26,076 INFO [optim.py:369] (2/7) 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 21:44:35,089 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4699, 2.0652, 2.3957, 2.8733, 2.8007, 2.2612, 2.0179, 2.5546], + device='cuda:2'), covar=tensor([0.0910, 0.1231, 0.0788, 0.0637, 0.0641, 0.1011, 0.0825, 0.0601], + device='cuda:2'), in_proj_covar=tensor([0.0186, 0.0202, 0.0186, 0.0172, 0.0178, 0.0177, 0.0151, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:44:43,179 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138678.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:44:44,998 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138681.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:44:53,114 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138686.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:45:03,881 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1672, 1.5462, 1.4018, 1.8168, 1.6251, 1.8208, 1.3922, 3.1219], + device='cuda:2'), covar=tensor([0.0665, 0.0820, 0.0809, 0.1170, 0.0658, 0.0460, 0.0720, 0.0160], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 21:45:05,143 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138697.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:45:29,241 INFO [finetune.py:976] (2/7) Epoch 25, batch 1250, loss[loss=0.1624, simple_loss=0.2248, pruned_loss=0.05004, over 4907.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.24, pruned_loss=0.04745, over 950139.91 frames. ], batch size: 43, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:46:07,995 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138742.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:46:09,210 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5691, 1.4139, 4.3029, 4.0050, 3.7774, 4.1080, 4.0652, 3.7721], + device='cuda:2'), covar=tensor([0.7158, 0.5762, 0.1078, 0.1795, 0.1141, 0.1658, 0.1468, 0.1611], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0307, 0.0407, 0.0411, 0.0348, 0.0412, 0.0319, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:46:09,228 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1978, 2.5920, 1.0543, 1.4474, 1.9917, 1.2445, 3.4664, 1.7656], + device='cuda:2'), covar=tensor([0.0628, 0.0576, 0.0780, 0.1212, 0.0500, 0.1005, 0.0225, 0.0618], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 21:46:11,076 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138747.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:46:19,918 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5486, 1.4309, 4.3524, 4.0667, 3.8282, 4.1481, 4.0307, 3.8206], + device='cuda:2'), covar=tensor([0.7236, 0.5510, 0.0956, 0.1532, 0.1021, 0.1696, 0.1690, 0.1418], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0306, 0.0406, 0.0411, 0.0348, 0.0412, 0.0319, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:46:23,979 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138758.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:46:33,340 INFO [finetune.py:976] (2/7) Epoch 25, batch 1300, loss[loss=0.1831, simple_loss=0.2387, pruned_loss=0.06377, over 4824.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2388, pruned_loss=0.04748, over 952542.14 frames. ], batch size: 40, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:46:40,291 INFO [optim.py:369] (2/7) 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 21:47:44,177 INFO [finetune.py:976] (2/7) Epoch 25, batch 1350, loss[loss=0.1322, simple_loss=0.207, pruned_loss=0.02868, over 4764.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2388, pruned_loss=0.0476, over 952464.55 frames. ], batch size: 27, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:48:48,890 INFO [finetune.py:976] (2/7) Epoch 25, batch 1400, loss[loss=0.1338, simple_loss=0.2063, pruned_loss=0.0307, over 4795.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.242, pruned_loss=0.04792, over 952675.34 frames. ], batch size: 25, lr: 3.02e-03, grad_scale: 64.0 +2023-04-27 21:48:50,720 INFO [optim.py:369] (2/7) 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 21:49:44,979 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138909.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:49:54,329 INFO [finetune.py:976] (2/7) Epoch 25, batch 1450, loss[loss=0.1358, simple_loss=0.213, pruned_loss=0.02933, over 4788.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2436, pruned_loss=0.04813, over 952590.93 frames. ], batch size: 29, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:50:07,864 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138926.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:50:39,516 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1314, 1.8324, 2.0396, 2.3427, 2.4057, 1.9679, 1.5802, 2.1796], + device='cuda:2'), covar=tensor([0.0741, 0.1163, 0.0779, 0.0593, 0.0570, 0.0906, 0.0804, 0.0556], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0200, 0.0185, 0.0172, 0.0178, 0.0176, 0.0150, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:51:01,097 INFO [finetune.py:976] (2/7) Epoch 25, batch 1500, loss[loss=0.2236, simple_loss=0.2906, pruned_loss=0.07829, over 4894.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2462, pruned_loss=0.04958, over 953352.10 frames. ], batch size: 32, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:51:03,880 INFO [optim.py:369] (2/7) 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 21:51:10,082 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138970.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 21:51:12,536 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138974.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:51:19,565 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138978.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:51:34,814 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9636, 2.4920, 2.1459, 2.1092, 1.3365, 1.4367, 2.2932, 1.3818], + device='cuda:2'), covar=tensor([0.1828, 0.1721, 0.1401, 0.1781, 0.2361, 0.2010, 0.0913, 0.2208], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0209, 0.0168, 0.0203, 0.0199, 0.0186, 0.0155, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 21:52:02,854 INFO [finetune.py:976] (2/7) Epoch 25, batch 1550, loss[loss=0.1414, simple_loss=0.2199, pruned_loss=0.03144, over 4755.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2465, pruned_loss=0.04947, over 955533.15 frames. ], batch size: 28, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:52:11,153 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139026.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:18,847 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139037.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:22,336 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139042.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:29,092 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139053.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:52:41,930 INFO [finetune.py:976] (2/7) Epoch 25, batch 1600, loss[loss=0.1529, simple_loss=0.2398, pruned_loss=0.03299, over 4737.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2428, pruned_loss=0.04803, over 954571.00 frames. ], batch size: 26, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:52:44,232 INFO [optim.py:369] (2/7) 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 21:53:47,321 INFO [finetune.py:976] (2/7) Epoch 25, batch 1650, loss[loss=0.1408, simple_loss=0.2139, pruned_loss=0.03383, over 4924.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2406, pruned_loss=0.04763, over 957250.16 frames. ], batch size: 37, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:53:50,088 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 21:54:33,489 INFO [finetune.py:976] (2/7) Epoch 25, batch 1700, loss[loss=0.1363, simple_loss=0.2042, pruned_loss=0.03421, over 4825.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.239, pruned_loss=0.04734, over 958691.70 frames. ], batch size: 25, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:54:35,337 INFO [optim.py:369] (2/7) 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 21:55:07,048 INFO [finetune.py:976] (2/7) Epoch 25, batch 1750, loss[loss=0.1707, simple_loss=0.2432, pruned_loss=0.04909, over 4888.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.241, pruned_loss=0.0484, over 958278.92 frames. ], batch size: 32, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:55:18,548 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8350, 2.9419, 2.4600, 2.7150, 3.0989, 2.7280, 3.9904, 2.2842], + device='cuda:2'), covar=tensor([0.3255, 0.2488, 0.3953, 0.2863, 0.1526, 0.2381, 0.1144, 0.3449], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0354, 0.0427, 0.0351, 0.0381, 0.0377, 0.0368, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 21:55:23,350 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139238.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:55:32,747 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5536, 1.4305, 0.8065, 1.2308, 1.4740, 1.4003, 1.2750, 1.3488], + device='cuda:2'), covar=tensor([0.0505, 0.0392, 0.0361, 0.0581, 0.0307, 0.0514, 0.0535, 0.0589], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 21:55:39,582 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-27 21:55:41,181 INFO [finetune.py:976] (2/7) Epoch 25, batch 1800, loss[loss=0.188, simple_loss=0.2562, pruned_loss=0.05989, over 4786.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2434, pruned_loss=0.04856, over 957095.85 frames. ], batch size: 29, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:55:41,253 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139265.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 21:55:42,989 INFO [optim.py:369] (2/7) 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 21:55:48,017 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139276.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:55:57,464 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139289.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:04,594 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:14,572 INFO [finetune.py:976] (2/7) Epoch 25, batch 1850, loss[loss=0.145, simple_loss=0.2273, pruned_loss=0.03132, over 4765.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2443, pruned_loss=0.04882, over 958023.03 frames. ], batch size: 28, lr: 3.01e-03, grad_scale: 64.0 +2023-04-27 21:56:16,538 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9191, 2.0430, 1.0869, 1.6140, 2.2129, 1.7573, 1.7265, 1.8403], + device='cuda:2'), covar=tensor([0.0483, 0.0343, 0.0278, 0.0520, 0.0229, 0.0474, 0.0460, 0.0541], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 21:56:29,234 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139337.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:29,262 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139337.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 21:56:33,194 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139342.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:43,883 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139350.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:56:51,941 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139353.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:57:05,217 INFO [finetune.py:976] (2/7) Epoch 25, batch 1900, loss[loss=0.2244, simple_loss=0.2855, pruned_loss=0.08162, over 4891.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2445, pruned_loss=0.04872, over 956977.65 frames. ], batch size: 35, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 21:57:07,684 INFO [optim.py:369] (2/7) 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 21:57:28,904 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139385.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:57:36,779 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:57:49,498 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139401.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 21:58:09,750 INFO [finetune.py:976] (2/7) Epoch 25, batch 1950, loss[loss=0.1472, simple_loss=0.2214, pruned_loss=0.03652, over 4788.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2429, pruned_loss=0.04796, over 956037.96 frames. ], batch size: 45, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 21:58:32,553 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8567, 2.0831, 2.1350, 2.2361, 2.0546, 2.0910, 2.2044, 2.1289], + device='cuda:2'), covar=tensor([0.3998, 0.6158, 0.4574, 0.4568, 0.5478, 0.7090, 0.5752, 0.5757], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0375, 0.0327, 0.0340, 0.0350, 0.0393, 0.0358, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:58:33,086 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1644, 3.1989, 2.5710, 3.6973, 3.0676, 3.1616, 1.5065, 3.1770], + device='cuda:2'), covar=tensor([0.1871, 0.1504, 0.4300, 0.2358, 0.3355, 0.1898, 0.5272, 0.2518], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0217, 0.0249, 0.0304, 0.0296, 0.0247, 0.0273, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 21:59:13,359 INFO [finetune.py:976] (2/7) Epoch 25, batch 2000, loss[loss=0.1391, simple_loss=0.214, pruned_loss=0.03213, over 4854.00 frames. ], tot_loss[loss=0.168, simple_loss=0.241, pruned_loss=0.04756, over 953211.59 frames. ], batch size: 47, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 21:59:15,797 INFO [optim.py:369] (2/7) 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 22:00:17,505 INFO [finetune.py:976] (2/7) Epoch 25, batch 2050, loss[loss=0.1587, simple_loss=0.2327, pruned_loss=0.04238, over 4823.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2384, pruned_loss=0.04635, over 955108.61 frames. ], batch size: 30, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:00:37,210 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 22:00:41,952 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3686, 3.0618, 0.8476, 1.5937, 1.6852, 2.0741, 1.8013, 0.9861], + device='cuda:2'), covar=tensor([0.1410, 0.0986, 0.1961, 0.1344, 0.1113, 0.1096, 0.1596, 0.1753], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0237, 0.0136, 0.0120, 0.0131, 0.0151, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 22:01:21,673 INFO [finetune.py:976] (2/7) Epoch 25, batch 2100, loss[loss=0.1921, simple_loss=0.2621, pruned_loss=0.06104, over 4891.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2379, pruned_loss=0.04655, over 954400.42 frames. ], batch size: 32, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:01:21,759 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139565.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 22:01:24,120 INFO [optim.py:369] (2/7) 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 22:01:30,512 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139571.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:01:37,273 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139582.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:01:44,575 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139613.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:01:59,114 INFO [finetune.py:976] (2/7) Epoch 25, batch 2150, loss[loss=0.1645, simple_loss=0.2329, pruned_loss=0.04804, over 4902.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2421, pruned_loss=0.04827, over 951398.87 frames. ], batch size: 32, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:02:10,543 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139632.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 22:02:10,595 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139632.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:02:17,290 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139645.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:02:25,614 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 22:02:31,933 INFO [finetune.py:976] (2/7) Epoch 25, batch 2200, loss[loss=0.1796, simple_loss=0.2415, pruned_loss=0.05881, over 4692.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2449, pruned_loss=0.04904, over 952981.46 frames. ], batch size: 23, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:02:34,822 INFO [optim.py:369] (2/7) 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 22:02:47,826 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-27 22:03:04,782 INFO [finetune.py:976] (2/7) Epoch 25, batch 2250, loss[loss=0.1755, simple_loss=0.2535, pruned_loss=0.04872, over 4924.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2461, pruned_loss=0.0493, over 952756.73 frames. ], batch size: 42, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:03:06,656 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-04-27 22:03:24,008 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4737, 1.2686, 1.2145, 1.3468, 1.7009, 1.3957, 1.2435, 1.1760], + device='cuda:2'), covar=tensor([0.1458, 0.1221, 0.1612, 0.1251, 0.0657, 0.1193, 0.1655, 0.1866], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0305, 0.0347, 0.0283, 0.0325, 0.0304, 0.0295, 0.0369], + device='cuda:2'), out_proj_covar=tensor([6.3171e-05, 6.2755e-05, 7.3048e-05, 5.6694e-05, 6.6704e-05, 6.3621e-05, + 6.1237e-05, 7.8305e-05], device='cuda:2') +2023-04-27 22:03:38,678 INFO [finetune.py:976] (2/7) Epoch 25, batch 2300, loss[loss=0.1682, simple_loss=0.2498, pruned_loss=0.04335, over 4798.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2461, pruned_loss=0.04898, over 951594.60 frames. ], batch size: 41, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:03:41,540 INFO [optim.py:369] (2/7) 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 22:03:41,673 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139769.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:03:45,266 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2971, 3.0197, 2.2391, 2.4768, 1.6912, 1.6339, 2.5154, 1.6179], + device='cuda:2'), covar=tensor([0.1601, 0.1335, 0.1374, 0.1539, 0.2238, 0.1856, 0.0899, 0.2011], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0210, 0.0168, 0.0203, 0.0199, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 22:03:54,758 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139789.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:04:03,988 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139804.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:04:05,254 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139806.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:04:21,936 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9304, 2.0926, 1.8200, 1.6375, 1.9146, 1.4956, 2.5639, 1.4279], + device='cuda:2'), covar=tensor([0.3880, 0.1738, 0.4843, 0.2783, 0.1972, 0.3011, 0.1410, 0.5246], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0354, 0.0428, 0.0353, 0.0382, 0.0376, 0.0370, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:04:23,019 INFO [finetune.py:976] (2/7) Epoch 25, batch 2350, loss[loss=0.1932, simple_loss=0.2617, pruned_loss=0.06237, over 4787.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2442, pruned_loss=0.04874, over 951403.16 frames. ], batch size: 29, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:04:33,536 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-04-27 22:04:44,831 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139830.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:06,879 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:06,888 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:08,872 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-27 22:05:17,197 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4098, 1.6505, 1.3793, 1.5472, 1.3290, 1.2660, 1.4446, 1.0768], + device='cuda:2'), covar=tensor([0.1487, 0.1093, 0.0810, 0.1145, 0.3237, 0.1129, 0.1524, 0.1985], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0300, 0.0213, 0.0277, 0.0311, 0.0255, 0.0248, 0.0263], + device='cuda:2'), 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:2') +2023-04-27 22:05:17,687 INFO [finetune.py:976] (2/7) Epoch 25, batch 2400, loss[loss=0.2077, simple_loss=0.2629, pruned_loss=0.07631, over 4828.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2418, pruned_loss=0.04828, over 952538.22 frames. ], batch size: 39, lr: 3.01e-03, grad_scale: 32.0 +2023-04-27 22:05:17,798 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:19,529 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:05:23,145 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1985, 2.1591, 2.3640, 2.7048, 2.6956, 2.1191, 1.8920, 2.4160], + device='cuda:2'), covar=tensor([0.0837, 0.1018, 0.0663, 0.0540, 0.0636, 0.0991, 0.0783, 0.0576], + device='cuda:2'), in_proj_covar=tensor([0.0188, 0.0206, 0.0189, 0.0175, 0.0181, 0.0180, 0.0153, 0.0181], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:05:37,246 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139911.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:05:55,628 INFO [finetune.py:976] (2/7) Epoch 25, batch 2450, loss[loss=0.1191, simple_loss=0.1944, pruned_loss=0.02186, over 4755.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.239, pruned_loss=0.04719, over 955153.26 frames. ], batch size: 28, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:06:04,991 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-27 22:06:15,192 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139927.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:18,758 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139932.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 22:06:27,876 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139938.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:35,937 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139942.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:37,754 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139945.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:06:46,904 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-04-27 22:06:59,678 INFO [finetune.py:976] (2/7) Epoch 25, batch 2500, loss[loss=0.1909, simple_loss=0.27, pruned_loss=0.05585, over 4846.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2402, pruned_loss=0.04764, over 956276.96 frames. ], batch size: 47, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:07:04,984 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139980.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:07:12,839 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0792, 1.5059, 1.9565, 2.2917, 2.0256, 1.5816, 1.2430, 1.7310], + device='cuda:2'), covar=tensor([0.3190, 0.3241, 0.1715, 0.2236, 0.2639, 0.2796, 0.4326, 0.1947], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0247, 0.0228, 0.0314, 0.0222, 0.0235, 0.0229, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 22:07:21,149 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139993.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:07:23,123 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-27 22:07:35,748 INFO [finetune.py:976] (2/7) Epoch 25, batch 2550, loss[loss=0.232, simple_loss=0.2965, pruned_loss=0.08378, over 4806.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2432, pruned_loss=0.04857, over 956529.64 frames. ], batch size: 41, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:07:36,491 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2451, 1.3818, 1.6995, 1.8020, 1.6843, 1.8401, 1.7696, 1.7205], + device='cuda:2'), covar=tensor([0.4049, 0.5068, 0.4235, 0.4163, 0.5404, 0.6648, 0.4812, 0.4557], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0339, 0.0348, 0.0392, 0.0357, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:08:09,495 INFO [finetune.py:976] (2/7) Epoch 25, batch 2600, loss[loss=0.1828, simple_loss=0.2669, pruned_loss=0.04936, over 4813.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2441, pruned_loss=0.04907, over 953404.79 frames. ], batch size: 40, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:08:12,529 INFO [optim.py:369] (2/7) 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 22:08:35,662 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140103.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:08:39,909 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140110.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:08:42,879 INFO [finetune.py:976] (2/7) Epoch 25, batch 2650, loss[loss=0.2136, simple_loss=0.2798, pruned_loss=0.07366, over 4763.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2439, pruned_loss=0.04873, over 953766.43 frames. ], batch size: 27, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:08:49,431 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140125.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:04,025 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140160.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:14,361 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140162.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:15,626 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 25, batch 2700, loss[loss=0.1498, simple_loss=0.214, pruned_loss=0.04283, over 4676.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2418, pruned_loss=0.04776, over 950657.75 frames. ], batch size: 23, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:09:19,144 INFO [optim.py:369] (2/7) 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 22:09:19,890 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140171.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:09:55,085 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140206.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:10:00,543 INFO [finetune.py:976] (2/7) Epoch 25, batch 2750, loss[loss=0.1386, simple_loss=0.2175, pruned_loss=0.02983, over 4757.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2403, pruned_loss=0.04779, over 951067.31 frames. ], batch size: 26, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:10:18,553 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140227.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:10:32,149 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140238.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:11:02,661 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3822, 2.8749, 2.4641, 2.7462, 2.1182, 2.5440, 2.6454, 2.0720], + device='cuda:2'), covar=tensor([0.1923, 0.1012, 0.0683, 0.1107, 0.2878, 0.0965, 0.1791, 0.2257], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0299, 0.0212, 0.0275, 0.0310, 0.0253, 0.0247, 0.0262], + device='cuda:2'), 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:2') +2023-04-27 22:11:03,252 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140259.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:11:12,263 INFO [finetune.py:976] (2/7) Epoch 25, batch 2800, loss[loss=0.1347, simple_loss=0.2069, pruned_loss=0.03122, over 4897.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.237, pruned_loss=0.0467, over 952524.72 frames. ], batch size: 32, lr: 3.01e-03, grad_scale: 16.0 +2023-04-27 22:11:15,316 INFO [optim.py:369] (2/7) 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 22:11:20,365 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140275.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:11:21,690 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1043, 0.6331, 0.9053, 0.8149, 1.2400, 0.9666, 0.8921, 0.9500], + device='cuda:2'), covar=tensor([0.1820, 0.1733, 0.1778, 0.1653, 0.1044, 0.1384, 0.1676, 0.2146], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0310, 0.0353, 0.0288, 0.0330, 0.0308, 0.0300, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.4163e-05, 6.3754e-05, 7.4252e-05, 5.7767e-05, 6.7786e-05, 6.4487e-05, + 6.2280e-05, 7.9517e-05], device='cuda:2') +2023-04-27 22:11:32,990 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 25, batch 2850, loss[loss=0.1477, simple_loss=0.2202, pruned_loss=0.03754, over 4824.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2347, pruned_loss=0.04557, over 954864.23 frames. ], batch size: 30, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:12:17,262 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140320.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:12:48,662 INFO [finetune.py:976] (2/7) Epoch 25, batch 2900, loss[loss=0.1982, simple_loss=0.2664, pruned_loss=0.06502, over 4883.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2388, pruned_loss=0.04725, over 955502.66 frames. ], batch size: 32, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:12:51,722 INFO [optim.py:369] (2/7) 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 22:13:00,439 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4116, 1.9619, 1.8752, 2.0110, 1.9016, 1.9411, 1.9200, 1.8861], + device='cuda:2'), covar=tensor([0.4952, 0.5258, 0.4813, 0.4230, 0.5443, 0.6754, 0.4940, 0.5034], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0374, 0.0326, 0.0339, 0.0349, 0.0393, 0.0358, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:13:02,710 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140387.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:20,920 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6890, 2.1910, 1.7861, 1.6764, 1.2681, 1.2808, 1.7843, 1.2046], + device='cuda:2'), covar=tensor([0.1730, 0.1347, 0.1352, 0.1669, 0.2341, 0.1928, 0.0978, 0.2088], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0211, 0.0169, 0.0204, 0.0200, 0.0187, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 22:13:22,623 INFO [finetune.py:976] (2/7) Epoch 25, batch 2950, loss[loss=0.1636, simple_loss=0.2421, pruned_loss=0.04254, over 4875.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2418, pruned_loss=0.0482, over 955678.53 frames. ], batch size: 34, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:13:28,750 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140425.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:42,312 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140445.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:44,116 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140448.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:52,226 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140459.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:52,863 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140460.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:53,451 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.5585, 3.4843, 2.6474, 4.1381, 3.5769, 3.5816, 1.4848, 3.5475], + device='cuda:2'), covar=tensor([0.1858, 0.1344, 0.2913, 0.1863, 0.2334, 0.1751, 0.5650, 0.2397], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0218, 0.0251, 0.0306, 0.0298, 0.0248, 0.0273, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:13:54,556 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140462.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:13:56,279 INFO [finetune.py:976] (2/7) Epoch 25, batch 3000, loss[loss=0.153, simple_loss=0.2184, pruned_loss=0.04382, over 4719.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2424, pruned_loss=0.04852, over 953470.66 frames. ], batch size: 23, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:13:56,279 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 22:14:07,208 INFO [finetune.py:1010] (2/7) Epoch 25, validation: loss=0.1531, simple_loss=0.2225, pruned_loss=0.04184, over 2265189.00 frames. +2023-04-27 22:14:07,208 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 22:14:07,897 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140466.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:08,015 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-27 22:14:10,183 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140473.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:21,877 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140489.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:24,278 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140493.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:27,442 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-04-27 22:14:33,158 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140510.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:14:39,581 INFO [finetune.py:976] (2/7) Epoch 25, batch 3050, loss[loss=0.1882, simple_loss=0.2493, pruned_loss=0.06353, over 4810.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2439, pruned_loss=0.04873, over 953760.13 frames. ], batch size: 33, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:14:53,390 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1430, 0.7743, 0.8958, 0.8284, 1.2568, 1.0086, 0.9598, 0.9902], + device='cuda:2'), covar=tensor([0.1710, 0.1443, 0.2042, 0.1494, 0.1082, 0.1453, 0.1752, 0.2413], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0311, 0.0354, 0.0289, 0.0331, 0.0309, 0.0301, 0.0377], + device='cuda:2'), out_proj_covar=tensor([6.4560e-05, 6.3872e-05, 7.4357e-05, 5.8005e-05, 6.7993e-05, 6.4624e-05, + 6.2477e-05, 7.9892e-05], device='cuda:2') +2023-04-27 22:15:02,559 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140554.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:15:12,553 INFO [finetune.py:976] (2/7) Epoch 25, batch 3100, loss[loss=0.1648, simple_loss=0.2431, pruned_loss=0.04329, over 4818.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2417, pruned_loss=0.04796, over 953289.51 frames. ], batch size: 39, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:15:12,753 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 22:15:16,051 INFO [optim.py:369] (2/7) 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 22:15:16,201 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140570.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:16:07,368 INFO [finetune.py:976] (2/7) Epoch 25, batch 3150, loss[loss=0.1359, simple_loss=0.2194, pruned_loss=0.02616, over 4770.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2389, pruned_loss=0.04702, over 952923.83 frames. ], batch size: 28, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:16:07,437 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140615.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:16:40,291 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-27 22:17:13,347 INFO [finetune.py:976] (2/7) Epoch 25, batch 3200, loss[loss=0.1772, simple_loss=0.244, pruned_loss=0.05515, over 4815.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2357, pruned_loss=0.04635, over 953360.98 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:17:21,870 INFO [optim.py:369] (2/7) 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 22:18:08,550 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3268, 1.4692, 1.2434, 1.4301, 1.1471, 1.1645, 1.3501, 1.0979], + device='cuda:2'), covar=tensor([0.1702, 0.1563, 0.1207, 0.1505, 0.3849, 0.1500, 0.1629, 0.2147], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0301, 0.0213, 0.0277, 0.0311, 0.0255, 0.0249, 0.0265], + device='cuda:2'), out_proj_covar=tensor([1.1397e-04, 1.1899e-04, 8.3908e-05, 1.0902e-04, 1.2557e-04, 1.0059e-04, + 1.0032e-04, 1.0472e-04], device='cuda:2') +2023-04-27 22:18:19,768 INFO [finetune.py:976] (2/7) Epoch 25, batch 3250, loss[loss=0.1926, simple_loss=0.2602, pruned_loss=0.06253, over 4922.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2362, pruned_loss=0.04656, over 951684.75 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:18:26,971 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140723.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:18:55,995 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140743.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:19:16,131 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 25, batch 3300, loss[loss=0.2543, simple_loss=0.3193, pruned_loss=0.09462, over 4821.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2396, pruned_loss=0.04735, over 949430.46 frames. ], batch size: 51, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:19:20,935 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:19:48,015 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140784.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:19,459 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140807.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:20,737 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140809.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:23,779 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 25, batch 3350, loss[loss=0.1871, simple_loss=0.2709, pruned_loss=0.05162, over 4823.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2422, pruned_loss=0.04799, over 952500.34 frames. ], batch size: 33, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:20:46,744 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140845.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:20:59,016 INFO [finetune.py:976] (2/7) Epoch 25, batch 3400, loss[loss=0.1621, simple_loss=0.2392, pruned_loss=0.04252, over 4868.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2442, pruned_loss=0.04868, over 953358.13 frames. ], batch size: 34, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:20:59,097 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140865.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:02,012 INFO [optim.py:369] (2/7) 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 22:21:02,142 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140870.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:07,915 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9734, 1.4332, 1.5609, 1.6176, 2.0563, 1.7339, 1.4720, 1.3836], + device='cuda:2'), covar=tensor([0.1469, 0.1453, 0.1891, 0.1358, 0.0843, 0.1398, 0.2037, 0.2465], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0308, 0.0352, 0.0287, 0.0329, 0.0306, 0.0299, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.4367e-05, 6.3366e-05, 7.4042e-05, 5.7604e-05, 6.7613e-05, 6.4098e-05, + 6.2050e-05, 7.9258e-05], device='cuda:2') +2023-04-27 22:21:32,399 INFO [finetune.py:976] (2/7) Epoch 25, batch 3450, loss[loss=0.1227, simple_loss=0.1834, pruned_loss=0.03099, over 4191.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2433, pruned_loss=0.04823, over 953651.36 frames. ], batch size: 18, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:21:32,496 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140917.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:21:55,388 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140947.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:22:05,107 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140963.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:22:06,248 INFO [finetune.py:976] (2/7) Epoch 25, batch 3500, loss[loss=0.1891, simple_loss=0.2495, pruned_loss=0.06434, over 4422.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2427, pruned_loss=0.04846, over 954388.55 frames. ], batch size: 19, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:22:09,318 INFO [optim.py:369] (2/7) 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 22:22:11,874 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9083, 1.2424, 1.7080, 1.8061, 1.7562, 1.8491, 1.6903, 1.6762], + device='cuda:2'), covar=tensor([0.4103, 0.5072, 0.4072, 0.4403, 0.5370, 0.6880, 0.4297, 0.4393], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0376, 0.0328, 0.0340, 0.0350, 0.0392, 0.0359, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:22:14,760 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140978.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:22:35,390 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141008.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:22:39,553 INFO [finetune.py:976] (2/7) Epoch 25, batch 3550, loss[loss=0.1253, simple_loss=0.1948, pruned_loss=0.02793, over 4780.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2396, pruned_loss=0.04743, over 954605.80 frames. ], batch size: 29, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:22:58,168 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141043.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:23:01,422 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-27 22:23:13,443 INFO [finetune.py:976] (2/7) Epoch 25, batch 3600, loss[loss=0.2105, simple_loss=0.2741, pruned_loss=0.07349, over 4820.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2377, pruned_loss=0.04692, over 954877.51 frames. ], batch size: 39, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:23:16,477 INFO [optim.py:369] (2/7) 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 22:23:27,753 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141079.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:23:46,968 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141091.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:23:49,599 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-04-27 22:24:19,423 INFO [finetune.py:976] (2/7) Epoch 25, batch 3650, loss[loss=0.1367, simple_loss=0.2196, pruned_loss=0.02693, over 4815.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2386, pruned_loss=0.04703, over 953252.52 frames. ], batch size: 25, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:24:42,745 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141143.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:24:43,920 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141145.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:24:57,940 INFO [finetune.py:976] (2/7) Epoch 25, batch 3700, loss[loss=0.2126, simple_loss=0.2802, pruned_loss=0.07257, over 4819.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2422, pruned_loss=0.04824, over 953351.27 frames. ], batch size: 40, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:24:58,015 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141165.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:24:58,058 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:25:15,809 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141193.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:25:22,658 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141204.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:25:30,030 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141213.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:25:31,150 INFO [finetune.py:976] (2/7) Epoch 25, batch 3750, loss[loss=0.1713, simple_loss=0.241, pruned_loss=0.05076, over 4901.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2418, pruned_loss=0.04804, over 952650.54 frames. ], batch size: 43, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:26:14,183 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0152, 2.6227, 0.9636, 1.4559, 1.9661, 1.3026, 3.5051, 1.7342], + device='cuda:2'), covar=tensor([0.0749, 0.0605, 0.0834, 0.1271, 0.0591, 0.1022, 0.0230, 0.0657], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 22:26:37,433 INFO [finetune.py:976] (2/7) Epoch 25, batch 3800, loss[loss=0.1681, simple_loss=0.2397, pruned_loss=0.04825, over 4924.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2434, pruned_loss=0.04835, over 953172.29 frames. ], batch size: 33, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:26:45,872 INFO [optim.py:369] (2/7) 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 22:26:47,826 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141273.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:27:28,692 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141303.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:27:42,126 INFO [finetune.py:976] (2/7) Epoch 25, batch 3850, loss[loss=0.1216, simple_loss=0.1963, pruned_loss=0.02345, over 4810.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2421, pruned_loss=0.04808, over 952386.01 frames. ], batch size: 25, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:28:02,218 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141329.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:28:27,557 INFO [finetune.py:976] (2/7) Epoch 25, batch 3900, loss[loss=0.1625, simple_loss=0.2235, pruned_loss=0.05078, over 4716.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2398, pruned_loss=0.04735, over 953029.60 frames. ], batch size: 23, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:28:31,460 INFO [optim.py:369] (2/7) 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 22:28:37,516 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141379.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:28:44,200 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141390.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:00,435 INFO [finetune.py:976] (2/7) Epoch 25, batch 3950, loss[loss=0.164, simple_loss=0.2329, pruned_loss=0.04749, over 4897.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2377, pruned_loss=0.0467, over 953729.29 frames. ], batch size: 32, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:29:01,141 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2162, 1.2620, 3.7243, 3.4753, 3.2815, 3.5785, 3.4946, 3.3173], + device='cuda:2'), covar=tensor([0.7349, 0.5701, 0.1259, 0.1979, 0.1241, 0.1858, 0.1815, 0.1637], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0307, 0.0408, 0.0409, 0.0349, 0.0412, 0.0318, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:29:01,814 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141417.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 22:29:09,300 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141427.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:17,827 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141441.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:33,819 INFO [finetune.py:976] (2/7) Epoch 25, batch 4000, loss[loss=0.1678, simple_loss=0.2446, pruned_loss=0.04556, over 4738.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2372, pruned_loss=0.04676, over 952363.78 frames. ], batch size: 59, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:29:33,918 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 22:29:37,085 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-27 22:29:43,311 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141478.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 22:29:55,971 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141499.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:29:57,849 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141502.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:30:05,461 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 25, batch 4050, loss[loss=0.1276, simple_loss=0.1945, pruned_loss=0.0303, over 4718.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2389, pruned_loss=0.04687, over 953362.37 frames. ], batch size: 23, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:30:26,620 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3060, 1.9917, 2.5369, 2.6960, 2.3611, 2.1927, 2.3629, 2.3118], + device='cuda:2'), covar=tensor([0.4669, 0.7144, 0.6736, 0.5624, 0.5903, 0.9378, 0.9129, 0.9569], + device='cuda:2'), in_proj_covar=tensor([0.0436, 0.0419, 0.0510, 0.0505, 0.0465, 0.0499, 0.0502, 0.0515], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:30:39,499 INFO [finetune.py:976] (2/7) Epoch 25, batch 4100, loss[loss=0.1419, simple_loss=0.2033, pruned_loss=0.04025, over 4246.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2426, pruned_loss=0.04831, over 953074.55 frames. ], batch size: 18, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:30:42,488 INFO [optim.py:369] (2/7) 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 22:30:44,329 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141573.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:30:45,129 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-27 22:31:04,700 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 25, batch 4150, loss[loss=0.135, simple_loss=0.2105, pruned_loss=0.02982, over 3950.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2429, pruned_loss=0.04846, over 951742.70 frames. ], batch size: 17, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:31:21,627 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141621.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:31:30,719 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2358, 1.3822, 1.7138, 1.8138, 1.6338, 1.7123, 1.7038, 1.7245], + device='cuda:2'), covar=tensor([0.3837, 0.4977, 0.4266, 0.4016, 0.5482, 0.7194, 0.4691, 0.4393], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0373, 0.0327, 0.0339, 0.0348, 0.0392, 0.0359, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:31:51,125 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3816, 2.9601, 0.9901, 1.6931, 2.3520, 1.5410, 4.0022, 2.0650], + device='cuda:2'), covar=tensor([0.0621, 0.0882, 0.0967, 0.1213, 0.0517, 0.0914, 0.0271, 0.0585], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0046, 0.0046, 0.0048, 0.0050, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0007, 0.0008, 0.0008, 0.0010, 0.0007], + device='cuda:2') +2023-04-27 22:32:02,964 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141651.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:32:23,211 INFO [finetune.py:976] (2/7) Epoch 25, batch 4200, loss[loss=0.1644, simple_loss=0.2311, pruned_loss=0.04882, over 4887.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2442, pruned_loss=0.04834, over 953301.69 frames. ], batch size: 32, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:32:25,195 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3065, 1.5463, 1.8072, 1.9029, 1.7722, 1.8072, 1.8458, 1.8149], + device='cuda:2'), covar=tensor([0.3661, 0.5001, 0.4116, 0.4252, 0.5290, 0.6704, 0.4341, 0.4387], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0376, 0.0328, 0.0341, 0.0350, 0.0394, 0.0360, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:32:26,264 INFO [optim.py:369] (2/7) 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 22:32:32,696 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4061, 2.2548, 1.7912, 1.9183, 2.3160, 1.9530, 2.6328, 1.6434], + device='cuda:2'), covar=tensor([0.3502, 0.1794, 0.4512, 0.3267, 0.1668, 0.2324, 0.1525, 0.4138], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0353, 0.0428, 0.0353, 0.0383, 0.0376, 0.0370, 0.0425], + device='cuda:2'), out_proj_covar=tensor([9.9883e-05, 1.0533e-04, 1.2979e-04, 1.0599e-04, 1.1375e-04, 1.1196e-04, + 1.0868e-04, 1.2791e-04], device='cuda:2') +2023-04-27 22:32:46,770 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 22:32:47,760 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141685.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:32:48,481 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 22:33:28,597 INFO [finetune.py:976] (2/7) Epoch 25, batch 4250, loss[loss=0.1792, simple_loss=0.2479, pruned_loss=0.05528, over 4824.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2415, pruned_loss=0.0473, over 955601.86 frames. ], batch size: 39, lr: 3.00e-03, grad_scale: 16.0 +2023-04-27 22:33:52,182 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-27 22:34:07,579 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5650, 3.0917, 0.9911, 1.8093, 2.5023, 1.6975, 4.2589, 2.2975], + device='cuda:2'), covar=tensor([0.0580, 0.0713, 0.0894, 0.1210, 0.0470, 0.0924, 0.0215, 0.0526], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0050, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 22:34:28,655 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7963, 1.3591, 1.9472, 2.3044, 1.9185, 1.8039, 1.9007, 1.8448], + device='cuda:2'), covar=tensor([0.4639, 0.6461, 0.5910, 0.5657, 0.5787, 0.7105, 0.7281, 0.8074], + device='cuda:2'), in_proj_covar=tensor([0.0438, 0.0419, 0.0512, 0.0507, 0.0466, 0.0500, 0.0503, 0.0515], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:34:30,331 INFO [finetune.py:976] (2/7) Epoch 25, batch 4300, loss[loss=0.1494, simple_loss=0.2256, pruned_loss=0.03663, over 4819.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.24, pruned_loss=0.04738, over 956802.35 frames. ], batch size: 40, lr: 2.99e-03, grad_scale: 16.0 +2023-04-27 22:34:39,435 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141773.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 22:35:15,326 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141797.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:35:16,545 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141799.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:35:26,143 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6958, 1.5571, 4.6781, 4.3847, 4.0352, 4.4666, 4.3124, 4.0789], + device='cuda:2'), covar=tensor([0.7104, 0.5497, 0.0926, 0.1547, 0.1232, 0.1422, 0.1089, 0.1765], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0310, 0.0411, 0.0412, 0.0351, 0.0415, 0.0321, 0.0371], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:35:35,010 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9899, 2.5027, 1.0329, 1.4778, 1.9615, 1.1808, 3.3073, 1.7529], + device='cuda:2'), covar=tensor([0.0685, 0.0581, 0.0764, 0.1196, 0.0476, 0.1023, 0.0214, 0.0587], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 22:35:36,800 INFO [finetune.py:976] (2/7) Epoch 25, batch 4350, loss[loss=0.1616, simple_loss=0.2346, pruned_loss=0.04426, over 4740.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2377, pruned_loss=0.0471, over 956253.35 frames. ], batch size: 59, lr: 2.99e-03, grad_scale: 16.0 +2023-04-27 22:36:20,220 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141847.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:36:40,498 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9830, 1.0428, 1.1961, 1.1479, 0.9624, 0.8957, 0.9854, 0.5199], + device='cuda:2'), covar=tensor([0.0544, 0.0566, 0.0426, 0.0530, 0.0752, 0.1158, 0.0459, 0.0633], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0068, 0.0067, 0.0068, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 22:36:42,215 INFO [finetune.py:976] (2/7) Epoch 25, batch 4400, loss[loss=0.1465, simple_loss=0.2054, pruned_loss=0.04382, over 4704.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2387, pruned_loss=0.04772, over 954693.31 frames. ], batch size: 23, lr: 2.99e-03, grad_scale: 16.0 +2023-04-27 22:36:50,565 INFO [optim.py:369] (2/7) 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 22:37:24,415 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1054, 2.4797, 0.7545, 1.4652, 1.4014, 1.8434, 1.6012, 0.7899], + device='cuda:2'), covar=tensor([0.1484, 0.1140, 0.1800, 0.1290, 0.1179, 0.0891, 0.1454, 0.1972], + device='cuda:2'), in_proj_covar=tensor([0.0119, 0.0242, 0.0138, 0.0122, 0.0134, 0.0154, 0.0119, 0.0121], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 22:37:46,057 INFO [finetune.py:976] (2/7) Epoch 25, batch 4450, loss[loss=0.1882, simple_loss=0.2439, pruned_loss=0.06626, over 4795.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.243, pruned_loss=0.04895, over 955087.63 frames. ], batch size: 25, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:37:46,784 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141916.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:38:04,883 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141928.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:38:06,109 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-27 22:38:31,145 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1720, 1.1371, 3.8530, 3.5704, 3.4265, 3.7212, 3.7157, 3.4190], + device='cuda:2'), covar=tensor([0.7742, 0.6603, 0.1287, 0.2083, 0.1325, 0.2033, 0.1434, 0.1614], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0309, 0.0411, 0.0411, 0.0350, 0.0415, 0.0320, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:38:41,077 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-27 22:38:51,129 INFO [finetune.py:976] (2/7) Epoch 25, batch 4500, loss[loss=0.2395, simple_loss=0.3024, pruned_loss=0.08829, over 4848.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2445, pruned_loss=0.04933, over 954476.41 frames. ], batch size: 47, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:38:59,405 INFO [optim.py:369] (2/7) 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 22:39:09,403 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141977.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:39:21,463 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141985.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:39:23,979 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141989.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:39:57,184 INFO [finetune.py:976] (2/7) Epoch 25, batch 4550, loss[loss=0.1557, simple_loss=0.2383, pruned_loss=0.03649, over 4906.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2446, pruned_loss=0.04892, over 955438.01 frames. ], batch size: 38, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:40:19,390 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142033.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:40:30,489 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7030, 1.1781, 1.3718, 1.3169, 1.7841, 1.5159, 1.2469, 1.3296], + device='cuda:2'), covar=tensor([0.1617, 0.1492, 0.1812, 0.1435, 0.0973, 0.1370, 0.1736, 0.2167], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0307, 0.0351, 0.0285, 0.0328, 0.0304, 0.0298, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.4086e-05, 6.3128e-05, 7.3727e-05, 5.7181e-05, 6.7211e-05, 6.3629e-05, + 6.1792e-05, 7.8982e-05], device='cuda:2') +2023-04-27 22:40:50,974 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4687, 1.2206, 4.4192, 4.1603, 3.7999, 4.2354, 4.0450, 3.8472], + device='cuda:2'), covar=tensor([0.7468, 0.6200, 0.0973, 0.1601, 0.1043, 0.1532, 0.1386, 0.1722], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0308, 0.0408, 0.0411, 0.0349, 0.0414, 0.0319, 0.0369], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:41:01,791 INFO [finetune.py:976] (2/7) Epoch 25, batch 4600, loss[loss=0.1176, simple_loss=0.1968, pruned_loss=0.01923, over 4757.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2442, pruned_loss=0.04815, over 955399.27 frames. ], batch size: 28, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:41:09,215 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-04-27 22:41:10,136 INFO [optim.py:369] (2/7) 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 22:41:12,055 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142073.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 22:41:43,179 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142097.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:42:06,628 INFO [finetune.py:976] (2/7) Epoch 25, batch 4650, loss[loss=0.2147, simple_loss=0.2732, pruned_loss=0.07806, over 4931.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2418, pruned_loss=0.04778, over 956000.85 frames. ], batch size: 38, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:42:15,790 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142121.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 22:42:28,067 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9184, 0.9577, 1.1742, 1.1573, 0.9451, 0.8989, 0.9745, 0.5278], + device='cuda:2'), covar=tensor([0.0581, 0.0617, 0.0409, 0.0560, 0.0718, 0.1224, 0.0453, 0.0638], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0068, 0.0067, 0.0069, 0.0075, 0.0095, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 22:42:47,434 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 25, batch 4700, loss[loss=0.156, simple_loss=0.2153, pruned_loss=0.04842, over 4822.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2391, pruned_loss=0.04727, over 958407.39 frames. ], batch size: 41, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:43:19,816 INFO [optim.py:369] (2/7) 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 22:44:17,215 INFO [finetune.py:976] (2/7) Epoch 25, batch 4750, loss[loss=0.1903, simple_loss=0.257, pruned_loss=0.06176, over 4927.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2381, pruned_loss=0.04714, over 958078.45 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:45:29,862 INFO [finetune.py:976] (2/7) Epoch 25, batch 4800, loss[loss=0.1555, simple_loss=0.2249, pruned_loss=0.04301, over 4905.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2407, pruned_loss=0.0479, over 955910.95 frames. ], batch size: 37, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:45:34,012 INFO [optim.py:369] (2/7) 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 22:45:40,862 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142284.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:46:36,154 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4040, 1.7000, 1.8244, 1.9319, 1.8360, 1.9001, 1.8587, 1.8671], + device='cuda:2'), covar=tensor([0.3967, 0.5557, 0.4412, 0.4326, 0.5317, 0.6513, 0.5134, 0.4843], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0374, 0.0327, 0.0340, 0.0348, 0.0392, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:46:36,611 INFO [finetune.py:976] (2/7) Epoch 25, batch 4850, loss[loss=0.1624, simple_loss=0.2451, pruned_loss=0.03986, over 4712.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2435, pruned_loss=0.04865, over 954948.24 frames. ], batch size: 59, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:47:42,706 INFO [finetune.py:976] (2/7) Epoch 25, batch 4900, loss[loss=0.2248, simple_loss=0.2874, pruned_loss=0.08115, over 4144.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.245, pruned_loss=0.04923, over 954631.55 frames. ], batch size: 65, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:47:51,886 INFO [optim.py:369] (2/7) 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 22:48:49,024 INFO [finetune.py:976] (2/7) Epoch 25, batch 4950, loss[loss=0.1596, simple_loss=0.2412, pruned_loss=0.03898, over 4809.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2459, pruned_loss=0.04945, over 954985.00 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:49:58,430 INFO [finetune.py:976] (2/7) Epoch 25, batch 5000, loss[loss=0.1249, simple_loss=0.1905, pruned_loss=0.02964, over 4812.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2438, pruned_loss=0.04898, over 955661.43 frames. ], batch size: 25, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:50:01,485 INFO [optim.py:369] (2/7) 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 22:51:03,145 INFO [finetune.py:976] (2/7) Epoch 25, batch 5050, loss[loss=0.1774, simple_loss=0.2472, pruned_loss=0.05374, over 4828.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2419, pruned_loss=0.04851, over 957108.89 frames. ], batch size: 40, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:51:28,732 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 22:51:36,356 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 22:51:38,000 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-27 22:51:48,460 INFO [finetune.py:976] (2/7) Epoch 25, batch 5100, loss[loss=0.1784, simple_loss=0.2354, pruned_loss=0.06067, over 4901.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2392, pruned_loss=0.04797, over 958063.05 frames. ], batch size: 32, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:51:51,977 INFO [optim.py:369] (2/7) 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 22:51:53,290 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142572.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:51:53,938 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142584.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:21,634 INFO [finetune.py:976] (2/7) Epoch 25, batch 5150, loss[loss=0.124, simple_loss=0.1775, pruned_loss=0.0352, over 3998.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.238, pruned_loss=0.04724, over 955549.00 frames. ], batch size: 17, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:52:26,238 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142620.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:33,588 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142632.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:35,342 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142634.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:52:56,102 INFO [finetune.py:976] (2/7) Epoch 25, batch 5200, loss[loss=0.1695, simple_loss=0.2418, pruned_loss=0.04862, over 4829.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2407, pruned_loss=0.0476, over 955762.68 frames. ], batch size: 30, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:53:00,177 INFO [optim.py:369] (2/7) 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 22:53:14,594 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2529, 1.4282, 1.3467, 1.6598, 1.5645, 1.8013, 1.3870, 3.4603], + device='cuda:2'), covar=tensor([0.0602, 0.0819, 0.0814, 0.1213, 0.0660, 0.0610, 0.0775, 0.0126], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 22:53:25,543 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1819, 2.6604, 2.3103, 2.5182, 1.8644, 2.2421, 2.4968, 1.8029], + device='cuda:2'), covar=tensor([0.1965, 0.1068, 0.0717, 0.1108, 0.3227, 0.1025, 0.1729, 0.2372], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0303, 0.0214, 0.0278, 0.0314, 0.0257, 0.0250, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1388e-04, 1.1954e-04, 8.4314e-05, 1.0943e-04, 1.2659e-04, 1.0131e-04, + 1.0072e-04, 1.0543e-04], device='cuda:2') +2023-04-27 22:53:26,748 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2074, 2.4937, 0.7523, 1.5687, 1.4898, 1.8242, 1.6500, 0.7735], + device='cuda:2'), covar=tensor([0.1326, 0.1167, 0.1868, 0.1236, 0.1057, 0.0949, 0.1394, 0.1794], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0137, 0.0121, 0.0132, 0.0153, 0.0117, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 22:53:29,661 INFO [finetune.py:976] (2/7) Epoch 25, batch 5250, loss[loss=0.1836, simple_loss=0.2545, pruned_loss=0.05636, over 4883.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2436, pruned_loss=0.04826, over 956352.97 frames. ], batch size: 32, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:53:36,237 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=142724.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:53:38,692 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0122, 2.4243, 2.1024, 2.3384, 1.8010, 2.0373, 1.9345, 1.6245], + device='cuda:2'), covar=tensor([0.1895, 0.1225, 0.0781, 0.1171, 0.3398, 0.1106, 0.2023, 0.2605], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0303, 0.0214, 0.0278, 0.0315, 0.0257, 0.0250, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1401e-04, 1.1967e-04, 8.4372e-05, 1.0954e-04, 1.2676e-04, 1.0138e-04, + 1.0082e-04, 1.0559e-04], device='cuda:2') +2023-04-27 22:54:08,100 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0337, 2.4318, 2.1347, 2.4044, 1.7635, 2.1241, 1.9781, 1.6945], + device='cuda:2'), covar=tensor([0.1752, 0.1169, 0.0770, 0.1093, 0.3120, 0.1132, 0.1861, 0.2285], + device='cuda:2'), in_proj_covar=tensor([0.0286, 0.0304, 0.0215, 0.0279, 0.0315, 0.0258, 0.0251, 0.0268], + device='cuda:2'), out_proj_covar=tensor([1.1425e-04, 1.1980e-04, 8.4508e-05, 1.0973e-04, 1.2705e-04, 1.0148e-04, + 1.0102e-04, 1.0578e-04], device='cuda:2') +2023-04-27 22:54:19,563 INFO [finetune.py:976] (2/7) Epoch 25, batch 5300, loss[loss=0.1688, simple_loss=0.239, pruned_loss=0.04936, over 4830.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2431, pruned_loss=0.04753, over 955912.45 frames. ], batch size: 47, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:54:22,611 INFO [optim.py:369] (2/7) 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 22:54:44,822 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142785.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:55:07,526 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0196, 3.9687, 2.8731, 4.6418, 3.9564, 4.0479, 1.9201, 3.8725], + device='cuda:2'), covar=tensor([0.1418, 0.1026, 0.2917, 0.1258, 0.3110, 0.1485, 0.5170, 0.2321], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0218, 0.0251, 0.0305, 0.0299, 0.0248, 0.0273, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:55:27,672 INFO [finetune.py:976] (2/7) Epoch 25, batch 5350, loss[loss=0.1502, simple_loss=0.2271, pruned_loss=0.03667, over 4771.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2433, pruned_loss=0.04674, over 956585.97 frames. ], batch size: 26, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:55:48,745 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3608, 3.0182, 2.6264, 2.7677, 2.1449, 2.5398, 2.6267, 1.9561], + device='cuda:2'), covar=tensor([0.1971, 0.1099, 0.0648, 0.1243, 0.2758, 0.1129, 0.1824, 0.2399], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0302, 0.0214, 0.0277, 0.0314, 0.0257, 0.0249, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1374e-04, 1.1930e-04, 8.4198e-05, 1.0912e-04, 1.2663e-04, 1.0111e-04, + 1.0057e-04, 1.0530e-04], device='cuda:2') +2023-04-27 22:56:34,014 INFO [finetune.py:976] (2/7) Epoch 25, batch 5400, loss[loss=0.1334, simple_loss=0.2121, pruned_loss=0.02735, over 4792.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2416, pruned_loss=0.04677, over 955278.35 frames. ], batch size: 45, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:56:42,412 INFO [optim.py:369] (2/7) 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 22:57:08,649 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6883, 1.7081, 0.8159, 1.3440, 1.6795, 1.5333, 1.4172, 1.4365], + device='cuda:2'), covar=tensor([0.0449, 0.0322, 0.0340, 0.0498, 0.0264, 0.0467, 0.0453, 0.0497], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 22:57:16,477 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9595, 2.2646, 2.1921, 2.3746, 2.1283, 2.0824, 2.2857, 2.1702], + device='cuda:2'), covar=tensor([0.3303, 0.5324, 0.4276, 0.3939, 0.5222, 0.6711, 0.5776, 0.4858], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0376, 0.0329, 0.0341, 0.0350, 0.0394, 0.0361, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 22:57:18,639 INFO [finetune.py:976] (2/7) Epoch 25, batch 5450, loss[loss=0.1873, simple_loss=0.2536, pruned_loss=0.06055, over 4927.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2396, pruned_loss=0.04671, over 957164.59 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:57:27,218 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142929.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:57:38,005 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9927, 1.8860, 1.7031, 1.5123, 1.9038, 1.6719, 2.1216, 1.5029], + device='cuda:2'), covar=tensor([0.2926, 0.1344, 0.3304, 0.2418, 0.1246, 0.1722, 0.1557, 0.3765], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0356, 0.0430, 0.0354, 0.0386, 0.0379, 0.0373, 0.0429], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:57:52,210 INFO [finetune.py:976] (2/7) Epoch 25, batch 5500, loss[loss=0.1431, simple_loss=0.2234, pruned_loss=0.03138, over 4751.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2365, pruned_loss=0.04606, over 955937.42 frames. ], batch size: 54, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:57:55,646 INFO [optim.py:369] (2/7) 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 22:58:03,709 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7743, 1.3668, 1.8074, 2.2711, 1.9213, 1.7324, 1.7787, 1.6918], + device='cuda:2'), covar=tensor([0.4357, 0.6509, 0.6004, 0.5153, 0.5459, 0.7390, 0.7567, 0.9694], + device='cuda:2'), in_proj_covar=tensor([0.0440, 0.0420, 0.0514, 0.0506, 0.0468, 0.0502, 0.0504, 0.0517], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:58:26,144 INFO [finetune.py:976] (2/7) Epoch 25, batch 5550, loss[loss=0.2129, simple_loss=0.2768, pruned_loss=0.07448, over 4126.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2388, pruned_loss=0.04721, over 955229.07 frames. ], batch size: 65, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:58:32,924 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4161, 1.4541, 1.7823, 1.8280, 1.3851, 1.2509, 1.4758, 0.9859], + device='cuda:2'), covar=tensor([0.0567, 0.0571, 0.0388, 0.0548, 0.0690, 0.1023, 0.0560, 0.0598], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0073, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 22:58:57,671 INFO [finetune.py:976] (2/7) Epoch 25, batch 5600, loss[loss=0.1931, simple_loss=0.2726, pruned_loss=0.05673, over 4845.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2425, pruned_loss=0.04823, over 953816.65 frames. ], batch size: 47, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 22:59:00,542 INFO [optim.py:369] (2/7) 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 22:59:06,455 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143080.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 22:59:16,960 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-27 22:59:24,990 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6891, 1.3672, 1.7654, 1.9991, 1.7649, 1.7059, 1.7738, 1.7425], + device='cuda:2'), covar=tensor([0.5099, 0.6882, 0.6710, 0.7133, 0.6146, 0.8436, 0.8828, 0.9466], + device='cuda:2'), in_proj_covar=tensor([0.0440, 0.0420, 0.0515, 0.0507, 0.0468, 0.0502, 0.0505, 0.0517], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 22:59:27,655 INFO [finetune.py:976] (2/7) Epoch 25, batch 5650, loss[loss=0.1608, simple_loss=0.2508, pruned_loss=0.03535, over 4761.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2448, pruned_loss=0.04852, over 952820.24 frames. ], batch size: 54, lr: 2.99e-03, grad_scale: 32.0 +2023-04-27 23:00:05,937 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.2466, 2.7703, 2.6001, 2.6768, 2.5227, 2.6164, 2.6354, 2.6256], + device='cuda:2'), covar=tensor([0.3439, 0.5312, 0.4243, 0.4486, 0.4999, 0.6260, 0.5537, 0.4410], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0375, 0.0328, 0.0341, 0.0349, 0.0395, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:00:24,806 INFO [finetune.py:976] (2/7) Epoch 25, batch 5700, loss[loss=0.1522, simple_loss=0.2152, pruned_loss=0.04458, over 3971.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2398, pruned_loss=0.04761, over 932148.52 frames. ], batch size: 17, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:00:27,768 INFO [optim.py:369] (2/7) 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 23:00:44,128 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3602, 2.6044, 1.7585, 2.1510, 2.8327, 2.1440, 2.1780, 2.1920], + device='cuda:2'), covar=tensor([0.0416, 0.0288, 0.0234, 0.0457, 0.0204, 0.0445, 0.0405, 0.0456], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0052, 0.0038, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 23:00:49,012 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-27 23:00:50,705 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6622, 1.0987, 1.4714, 1.5299, 1.5026, 1.5562, 1.4987, 1.4881], + device='cuda:2'), covar=tensor([0.3329, 0.4045, 0.3500, 0.3828, 0.4811, 0.6312, 0.3836, 0.3653], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0375, 0.0328, 0.0341, 0.0349, 0.0395, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:01:04,602 INFO [finetune.py:976] (2/7) Epoch 26, batch 0, loss[loss=0.138, simple_loss=0.2069, pruned_loss=0.03457, over 4420.00 frames. ], tot_loss[loss=0.138, simple_loss=0.2069, pruned_loss=0.03457, over 4420.00 frames. ], batch size: 19, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:01:04,602 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 23:01:07,285 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8347, 2.1191, 1.9070, 2.0840, 1.6238, 1.8430, 1.7713, 1.4149], + device='cuda:2'), covar=tensor([0.1632, 0.1113, 0.0679, 0.1012, 0.3231, 0.1008, 0.1582, 0.2112], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0302, 0.0213, 0.0276, 0.0313, 0.0256, 0.0248, 0.0266], + device='cuda:2'), out_proj_covar=tensor([1.1349e-04, 1.1920e-04, 8.3963e-05, 1.0865e-04, 1.2629e-04, 1.0072e-04, + 1.0005e-04, 1.0505e-04], device='cuda:2') +2023-04-27 23:01:09,333 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3966, 1.3030, 1.6597, 1.6365, 1.3035, 1.2622, 1.3951, 0.8629], + device='cuda:2'), covar=tensor([0.0547, 0.0643, 0.0424, 0.0560, 0.0766, 0.1122, 0.0468, 0.0558], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0068, 0.0073, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:01:12,280 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2465, 1.5336, 1.8016, 1.9161, 1.8560, 1.9467, 1.7737, 1.8174], + device='cuda:2'), covar=tensor([0.4013, 0.5840, 0.4662, 0.4694, 0.5825, 0.7251, 0.5978, 0.5272], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0375, 0.0328, 0.0341, 0.0349, 0.0395, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:01:26,497 INFO [finetune.py:1010] (2/7) Epoch 26, validation: loss=0.1543, simple_loss=0.2237, pruned_loss=0.04251, over 2265189.00 frames. +2023-04-27 23:01:26,497 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 23:01:42,068 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 23:01:43,210 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1865, 2.4615, 2.1583, 2.4083, 1.7944, 2.2181, 2.1195, 1.7084], + device='cuda:2'), covar=tensor([0.1691, 0.1066, 0.0780, 0.1086, 0.3056, 0.0968, 0.1773, 0.2278], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0302, 0.0213, 0.0276, 0.0313, 0.0256, 0.0248, 0.0266], + device='cuda:2'), out_proj_covar=tensor([1.1334e-04, 1.1906e-04, 8.3882e-05, 1.0857e-04, 1.2624e-04, 1.0060e-04, + 9.9936e-05, 1.0497e-04], device='cuda:2') +2023-04-27 23:02:16,889 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143229.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:02:34,936 INFO [finetune.py:976] (2/7) Epoch 26, batch 50, loss[loss=0.1389, simple_loss=0.21, pruned_loss=0.03386, over 4814.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2428, pruned_loss=0.04662, over 215762.81 frames. ], batch size: 25, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:03:09,819 INFO [optim.py:369] (2/7) 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 23:03:19,707 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143277.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:03:40,347 INFO [finetune.py:976] (2/7) Epoch 26, batch 100, loss[loss=0.1654, simple_loss=0.2322, pruned_loss=0.04931, over 4821.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2377, pruned_loss=0.0465, over 379652.34 frames. ], batch size: 30, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:03:54,535 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7305, 3.1017, 1.0775, 2.0330, 2.3420, 1.7676, 4.3887, 2.4116], + device='cuda:2'), covar=tensor([0.0566, 0.0664, 0.0821, 0.1162, 0.0528, 0.0931, 0.0210, 0.0541], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 23:04:19,089 INFO [finetune.py:976] (2/7) Epoch 26, batch 150, loss[loss=0.1559, simple_loss=0.2212, pruned_loss=0.04533, over 4869.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2331, pruned_loss=0.04531, over 508507.29 frames. ], batch size: 31, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:04:25,875 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-04-27 23:04:37,628 INFO [optim.py:369] (2/7) 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 23:04:43,892 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143380.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:04:52,922 INFO [finetune.py:976] (2/7) Epoch 26, batch 200, loss[loss=0.1768, simple_loss=0.2307, pruned_loss=0.06143, over 4795.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2312, pruned_loss=0.04451, over 608352.82 frames. ], batch size: 25, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:04:59,041 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143428.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:05:16,660 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4366, 1.8158, 2.3355, 2.7764, 2.2370, 1.7858, 1.5504, 2.0677], + device='cuda:2'), covar=tensor([0.2983, 0.3019, 0.1556, 0.1983, 0.2491, 0.2577, 0.3731, 0.1959], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0315, 0.0221, 0.0235, 0.0227, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 23:05:31,680 INFO [finetune.py:976] (2/7) Epoch 26, batch 250, loss[loss=0.1491, simple_loss=0.2144, pruned_loss=0.04186, over 4797.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2351, pruned_loss=0.04541, over 686244.71 frames. ], batch size: 25, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:05:44,886 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143461.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:05:50,285 INFO [optim.py:369] (2/7) 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 23:05:58,104 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=143473.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:05:58,134 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1127, 2.6933, 2.3085, 2.0798, 1.5719, 1.6105, 2.4629, 1.6419], + device='cuda:2'), covar=tensor([0.1637, 0.1490, 0.1244, 0.1645, 0.2046, 0.1710, 0.0759, 0.1821], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0210, 0.0168, 0.0203, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 23:06:15,190 INFO [finetune.py:976] (2/7) Epoch 26, batch 300, loss[loss=0.1846, simple_loss=0.262, pruned_loss=0.05365, over 4932.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2404, pruned_loss=0.04652, over 748355.59 frames. ], batch size: 33, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:06:19,507 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-27 23:06:43,062 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143534.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 23:06:48,238 INFO [finetune.py:976] (2/7) Epoch 26, batch 350, loss[loss=0.188, simple_loss=0.2554, pruned_loss=0.06026, over 4143.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2438, pruned_loss=0.04864, over 793071.85 frames. ], batch size: 65, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:07:08,173 INFO [optim.py:369] (2/7) 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 23:07:22,102 INFO [finetune.py:976] (2/7) Epoch 26, batch 400, loss[loss=0.1918, simple_loss=0.258, pruned_loss=0.0628, over 4810.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2441, pruned_loss=0.04866, over 828668.07 frames. ], batch size: 41, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:07:42,362 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-04-27 23:07:55,516 INFO [finetune.py:976] (2/7) Epoch 26, batch 450, loss[loss=0.1848, simple_loss=0.2574, pruned_loss=0.05609, over 4832.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.242, pruned_loss=0.04809, over 857052.23 frames. ], batch size: 47, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:07:58,554 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1497, 1.8080, 2.0137, 2.4326, 2.5159, 2.0404, 1.7880, 2.2518], + device='cuda:2'), covar=tensor([0.0816, 0.1165, 0.0792, 0.0595, 0.0563, 0.0810, 0.0722, 0.0526], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0203, 0.0185, 0.0171, 0.0177, 0.0177, 0.0150, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:08:20,552 INFO [optim.py:369] (2/7) 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 23:08:43,964 INFO [finetune.py:976] (2/7) Epoch 26, batch 500, loss[loss=0.1754, simple_loss=0.238, pruned_loss=0.05637, over 4748.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2391, pruned_loss=0.04732, over 878136.57 frames. ], batch size: 59, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:09:17,389 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5993, 1.5324, 1.8892, 1.9824, 1.5212, 1.3429, 1.5927, 0.9904], + device='cuda:2'), covar=tensor([0.0532, 0.0618, 0.0368, 0.0546, 0.0641, 0.1139, 0.0584, 0.0648], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0073, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:09:27,561 INFO [finetune.py:976] (2/7) Epoch 26, batch 550, loss[loss=0.1816, simple_loss=0.2575, pruned_loss=0.05287, over 4829.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2367, pruned_loss=0.04683, over 895907.66 frames. ], batch size: 39, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:09:31,250 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5520, 1.3039, 4.2500, 3.9510, 3.6740, 4.0737, 4.0048, 3.7425], + device='cuda:2'), covar=tensor([0.7494, 0.6225, 0.1160, 0.1961, 0.1186, 0.1668, 0.1552, 0.1566], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0308, 0.0408, 0.0411, 0.0349, 0.0415, 0.0318, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:09:37,697 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143756.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 23:09:47,112 INFO [optim.py:369] (2/7) 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 23:10:00,517 INFO [finetune.py:976] (2/7) Epoch 26, batch 600, loss[loss=0.1928, simple_loss=0.2711, pruned_loss=0.05719, over 4860.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2379, pruned_loss=0.04737, over 909696.61 frames. ], batch size: 44, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:10:25,722 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143829.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:10:33,615 INFO [finetune.py:976] (2/7) Epoch 26, batch 650, loss[loss=0.1631, simple_loss=0.2482, pruned_loss=0.03905, over 4730.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2398, pruned_loss=0.04743, over 921577.34 frames. ], batch size: 59, lr: 2.98e-03, grad_scale: 32.0 +2023-04-27 23:11:14,499 INFO [optim.py:369] (2/7) 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 23:11:40,022 INFO [finetune.py:976] (2/7) Epoch 26, batch 700, loss[loss=0.1709, simple_loss=0.2555, pruned_loss=0.04312, over 4844.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2415, pruned_loss=0.04758, over 929911.40 frames. ], batch size: 49, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:11:41,488 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-27 23:12:45,799 INFO [finetune.py:976] (2/7) Epoch 26, batch 750, loss[loss=0.1652, simple_loss=0.2484, pruned_loss=0.04094, over 4901.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.242, pruned_loss=0.04767, over 936251.05 frames. ], batch size: 36, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:13:05,040 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4102, 1.8020, 2.1916, 2.8956, 2.3188, 1.7716, 1.9325, 2.2496], + device='cuda:2'), covar=tensor([0.2823, 0.3161, 0.1535, 0.2166, 0.2293, 0.2442, 0.3524, 0.1905], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0227, 0.0312, 0.0219, 0.0234, 0.0226, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 23:13:26,583 INFO [optim.py:369] (2/7) 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 23:13:56,909 INFO [finetune.py:976] (2/7) Epoch 26, batch 800, loss[loss=0.1542, simple_loss=0.2359, pruned_loss=0.03628, over 4764.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2411, pruned_loss=0.04675, over 940468.05 frames. ], batch size: 26, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:13:57,039 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4693, 1.3841, 1.6396, 1.7506, 1.3931, 1.1180, 1.4161, 0.8646], + device='cuda:2'), covar=tensor([0.0527, 0.0442, 0.0434, 0.0443, 0.0561, 0.1282, 0.0536, 0.0635], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0074, 0.0094, 0.0073, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:14:52,476 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9774, 1.4292, 1.7001, 2.3216, 2.3621, 1.8084, 1.5520, 2.0067], + device='cuda:2'), covar=tensor([0.0800, 0.1554, 0.1024, 0.0512, 0.0596, 0.0882, 0.0814, 0.0620], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0202, 0.0185, 0.0171, 0.0177, 0.0177, 0.0150, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:15:01,725 INFO [finetune.py:976] (2/7) Epoch 26, batch 850, loss[loss=0.1546, simple_loss=0.2175, pruned_loss=0.04583, over 4904.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.239, pruned_loss=0.04628, over 943073.54 frames. ], batch size: 32, lr: 2.98e-03, grad_scale: 64.0 +2023-04-27 23:15:15,467 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144056.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:15:35,338 INFO [optim.py:369] (2/7) 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 23:16:05,579 INFO [finetune.py:976] (2/7) Epoch 26, batch 900, loss[loss=0.1369, simple_loss=0.2136, pruned_loss=0.03013, over 4898.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2363, pruned_loss=0.04546, over 946399.37 frames. ], batch size: 32, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:16:08,974 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-27 23:16:12,932 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144104.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:16:17,985 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-04-27 23:16:22,994 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0387, 1.0835, 1.1525, 1.1753, 0.9875, 0.8848, 1.0520, 0.7830], + device='cuda:2'), covar=tensor([0.0547, 0.0598, 0.0489, 0.0504, 0.0655, 0.1111, 0.0429, 0.0549], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0073, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:16:34,942 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144129.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:16:41,251 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6762, 3.3289, 2.8930, 3.0838, 2.4423, 2.9273, 3.0373, 2.5693], + device='cuda:2'), covar=tensor([0.1853, 0.1064, 0.0696, 0.1100, 0.2588, 0.1045, 0.1563, 0.2016], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0301, 0.0213, 0.0276, 0.0313, 0.0255, 0.0249, 0.0265], + device='cuda:2'), out_proj_covar=tensor([1.1395e-04, 1.1845e-04, 8.3950e-05, 1.0873e-04, 1.2625e-04, 1.0030e-04, + 1.0024e-04, 1.0456e-04], device='cuda:2') +2023-04-27 23:16:53,899 INFO [finetune.py:976] (2/7) Epoch 26, batch 950, loss[loss=0.2108, simple_loss=0.2659, pruned_loss=0.0779, over 4928.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2351, pruned_loss=0.04567, over 949373.03 frames. ], batch size: 37, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:17:27,097 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2143, 1.4769, 1.2923, 1.7939, 1.6481, 1.8151, 1.4579, 3.2337], + device='cuda:2'), covar=tensor([0.0608, 0.0777, 0.0796, 0.1099, 0.0575, 0.0511, 0.0687, 0.0149], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 23:17:28,112 INFO [optim.py:369] (2/7) 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 23:17:36,053 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7276, 1.4347, 1.2986, 1.6084, 1.8675, 1.5444, 1.3740, 1.2626], + device='cuda:2'), covar=tensor([0.1831, 0.1489, 0.1946, 0.1487, 0.1040, 0.1822, 0.2165, 0.2429], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0308, 0.0350, 0.0286, 0.0329, 0.0303, 0.0297, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4177e-05, 6.3268e-05, 7.3400e-05, 5.7218e-05, 6.7499e-05, 6.3342e-05, + 6.1505e-05, 7.9160e-05], device='cuda:2') +2023-04-27 23:17:37,856 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144177.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:17:59,696 INFO [finetune.py:976] (2/7) Epoch 26, batch 1000, loss[loss=0.174, simple_loss=0.2459, pruned_loss=0.05106, over 4895.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2378, pruned_loss=0.04686, over 950760.25 frames. ], batch size: 32, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:18:32,527 INFO [finetune.py:976] (2/7) Epoch 26, batch 1050, loss[loss=0.163, simple_loss=0.2365, pruned_loss=0.04475, over 4825.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2393, pruned_loss=0.04696, over 951727.27 frames. ], batch size: 30, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:18:45,247 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8289, 2.4411, 2.0452, 1.8720, 1.3621, 1.4362, 2.0909, 1.3238], + device='cuda:2'), covar=tensor([0.1548, 0.1339, 0.1181, 0.1566, 0.2129, 0.1763, 0.0827, 0.1895], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0210, 0.0168, 0.0204, 0.0200, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 23:18:51,250 INFO [optim.py:369] (2/7) 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 23:19:06,554 INFO [finetune.py:976] (2/7) Epoch 26, batch 1100, loss[loss=0.1326, simple_loss=0.196, pruned_loss=0.03457, over 3994.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2414, pruned_loss=0.04721, over 952503.05 frames. ], batch size: 17, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:19:39,801 INFO [finetune.py:976] (2/7) Epoch 26, batch 1150, loss[loss=0.1664, simple_loss=0.2478, pruned_loss=0.04252, over 4918.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.242, pruned_loss=0.04712, over 953530.40 frames. ], batch size: 29, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:19:46,926 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7891, 3.1258, 1.0052, 1.9183, 2.9294, 1.7346, 4.5559, 2.3727], + device='cuda:2'), covar=tensor([0.0576, 0.0731, 0.0848, 0.1234, 0.0445, 0.0944, 0.0240, 0.0570], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 23:19:58,623 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6578, 3.4890, 0.9250, 1.9132, 1.8633, 2.5342, 1.9474, 1.0433], + device='cuda:2'), covar=tensor([0.1246, 0.0796, 0.1900, 0.1162, 0.0998, 0.0902, 0.1417, 0.2008], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0237, 0.0135, 0.0120, 0.0131, 0.0151, 0.0116, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:19:58,645 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6716, 1.6547, 0.7847, 1.3315, 1.7669, 1.4868, 1.3889, 1.4967], + device='cuda:2'), covar=tensor([0.0469, 0.0351, 0.0340, 0.0535, 0.0272, 0.0496, 0.0478, 0.0546], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-27 23:19:59,723 INFO [optim.py:369] (2/7) 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 23:20:13,107 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.4694, 3.3459, 2.4983, 3.9437, 3.4412, 3.4134, 1.5608, 3.3857], + device='cuda:2'), covar=tensor([0.1957, 0.1381, 0.3320, 0.2208, 0.2916, 0.2011, 0.5834, 0.2569], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0219, 0.0253, 0.0307, 0.0300, 0.0248, 0.0274, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:20:14,235 INFO [finetune.py:976] (2/7) Epoch 26, batch 1200, loss[loss=0.1667, simple_loss=0.2331, pruned_loss=0.05015, over 4931.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2409, pruned_loss=0.0467, over 955028.93 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:20:32,358 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144411.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:21:13,638 INFO [finetune.py:976] (2/7) Epoch 26, batch 1250, loss[loss=0.1821, simple_loss=0.2577, pruned_loss=0.0532, over 4887.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2389, pruned_loss=0.04632, over 954322.82 frames. ], batch size: 32, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:21:45,828 INFO [optim.py:369] (2/7) 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 23:21:51,777 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144472.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:22:16,358 INFO [finetune.py:976] (2/7) Epoch 26, batch 1300, loss[loss=0.2062, simple_loss=0.2632, pruned_loss=0.07464, over 4897.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2371, pruned_loss=0.0461, over 955462.55 frames. ], batch size: 35, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:23:06,076 INFO [finetune.py:976] (2/7) Epoch 26, batch 1350, loss[loss=0.1562, simple_loss=0.2451, pruned_loss=0.0336, over 4820.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.237, pruned_loss=0.04604, over 952793.78 frames. ], batch size: 40, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:23:09,633 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0061, 2.4485, 2.1637, 2.3957, 1.7316, 2.1732, 2.0981, 1.6844], + device='cuda:2'), covar=tensor([0.1919, 0.1282, 0.0817, 0.1182, 0.3181, 0.1166, 0.1816, 0.2522], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0299, 0.0212, 0.0274, 0.0312, 0.0254, 0.0247, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1339e-04, 1.1781e-04, 8.3549e-05, 1.0802e-04, 1.2573e-04, 9.9741e-05, + 9.9468e-05, 1.0403e-04], device='cuda:2') +2023-04-27 23:23:26,215 INFO [optim.py:369] (2/7) 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 23:23:39,998 INFO [finetune.py:976] (2/7) Epoch 26, batch 1400, loss[loss=0.1563, simple_loss=0.232, pruned_loss=0.04033, over 4896.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2396, pruned_loss=0.0472, over 952600.81 frames. ], batch size: 35, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:23:54,337 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4807, 1.4337, 1.7916, 1.8283, 1.3277, 1.2195, 1.5363, 1.0952], + device='cuda:2'), covar=tensor([0.0568, 0.0562, 0.0393, 0.0556, 0.0766, 0.1094, 0.0485, 0.0539], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0067, 0.0069, 0.0074, 0.0095, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:23:57,848 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3095, 1.8244, 2.2396, 2.5860, 2.2385, 1.7518, 1.4250, 2.0497], + device='cuda:2'), covar=tensor([0.3020, 0.2799, 0.1575, 0.2022, 0.2313, 0.2445, 0.3647, 0.1870], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0227, 0.0312, 0.0221, 0.0234, 0.0226, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 23:23:59,012 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3135, 1.6207, 1.4917, 1.7917, 1.8047, 2.0942, 1.5589, 3.7068], + device='cuda:2'), covar=tensor([0.0579, 0.0805, 0.0786, 0.1187, 0.0593, 0.0433, 0.0691, 0.0112], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 23:24:12,867 INFO [finetune.py:976] (2/7) Epoch 26, batch 1450, loss[loss=0.1835, simple_loss=0.2608, pruned_loss=0.05308, over 4807.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2414, pruned_loss=0.04767, over 952448.19 frames. ], batch size: 41, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:24:25,439 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-27 23:24:33,310 INFO [optim.py:369] (2/7) 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 23:24:46,127 INFO [finetune.py:976] (2/7) Epoch 26, batch 1500, loss[loss=0.1935, simple_loss=0.2631, pruned_loss=0.06192, over 4763.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2432, pruned_loss=0.04808, over 953567.15 frames. ], batch size: 59, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:24:53,194 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6403, 2.2916, 2.6924, 3.1439, 2.6156, 2.2150, 2.0449, 2.5248], + device='cuda:2'), covar=tensor([0.2948, 0.2572, 0.1389, 0.1766, 0.2238, 0.2254, 0.3314, 0.1519], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0227, 0.0312, 0.0221, 0.0235, 0.0226, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 23:25:09,289 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5518, 1.3293, 1.2847, 1.3239, 1.7357, 1.5046, 1.2607, 1.2726], + device='cuda:2'), covar=tensor([0.1520, 0.1152, 0.1498, 0.1420, 0.0762, 0.1159, 0.1549, 0.1848], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0309, 0.0353, 0.0288, 0.0330, 0.0306, 0.0299, 0.0376], + device='cuda:2'), out_proj_covar=tensor([6.4799e-05, 6.3532e-05, 7.4185e-05, 5.7847e-05, 6.7660e-05, 6.3939e-05, + 6.2063e-05, 7.9840e-05], device='cuda:2') +2023-04-27 23:25:20,128 INFO [finetune.py:976] (2/7) Epoch 26, batch 1550, loss[loss=0.133, simple_loss=0.2112, pruned_loss=0.02735, over 4847.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2429, pruned_loss=0.04826, over 951776.90 frames. ], batch size: 44, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:25:21,951 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4328, 1.3764, 1.4472, 1.0038, 1.4620, 1.2592, 1.7350, 1.3857], + device='cuda:2'), covar=tensor([0.3402, 0.1810, 0.4242, 0.2517, 0.1437, 0.1943, 0.1398, 0.4307], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0351, 0.0426, 0.0350, 0.0382, 0.0375, 0.0366, 0.0424], + device='cuda:2'), out_proj_covar=tensor([9.9805e-05, 1.0468e-04, 1.2882e-04, 1.0494e-04, 1.1339e-04, 1.1138e-04, + 1.0697e-04, 1.2753e-04], device='cuda:2') +2023-04-27 23:25:42,647 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 26, batch 1600, loss[loss=0.1358, simple_loss=0.2188, pruned_loss=0.02642, over 4823.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2408, pruned_loss=0.0475, over 952680.48 frames. ], batch size: 40, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:27:00,876 INFO [finetune.py:976] (2/7) Epoch 26, batch 1650, loss[loss=0.1533, simple_loss=0.2188, pruned_loss=0.04384, over 4793.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.238, pruned_loss=0.04649, over 953898.07 frames. ], batch size: 29, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:27:10,828 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5481, 1.3747, 4.3437, 4.0990, 3.7732, 4.1801, 4.0824, 3.7916], + device='cuda:2'), covar=tensor([0.7392, 0.6290, 0.0955, 0.1646, 0.1141, 0.1786, 0.1282, 0.1549], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0308, 0.0408, 0.0409, 0.0349, 0.0415, 0.0318, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:27:20,928 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 26, batch 1700, loss[loss=0.1787, simple_loss=0.233, pruned_loss=0.0622, over 4225.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2359, pruned_loss=0.04587, over 954363.71 frames. ], batch size: 65, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:28:30,626 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-27 23:28:34,818 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5490, 1.3964, 1.2812, 1.4032, 1.7853, 1.4779, 1.2724, 1.2112], + device='cuda:2'), covar=tensor([0.1654, 0.1115, 0.1701, 0.1162, 0.0790, 0.1400, 0.1586, 0.2055], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0306, 0.0349, 0.0285, 0.0328, 0.0304, 0.0297, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.4188e-05, 6.2965e-05, 7.3383e-05, 5.7196e-05, 6.7251e-05, 6.3461e-05, + 6.1564e-05, 7.9364e-05], device='cuda:2') +2023-04-27 23:28:44,464 INFO [finetune.py:976] (2/7) Epoch 26, batch 1750, loss[loss=0.1586, simple_loss=0.2482, pruned_loss=0.03452, over 4812.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2394, pruned_loss=0.04714, over 955575.45 frames. ], batch size: 41, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:28:45,291 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-27 23:29:07,729 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4729, 2.3349, 2.2339, 2.0770, 2.5672, 2.0890, 3.0816, 1.8653], + device='cuda:2'), covar=tensor([0.2768, 0.1447, 0.3010, 0.2102, 0.1259, 0.2189, 0.0967, 0.3466], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0349, 0.0423, 0.0349, 0.0380, 0.0372, 0.0363, 0.0420], + device='cuda:2'), out_proj_covar=tensor([9.9199e-05, 1.0381e-04, 1.2790e-04, 1.0461e-04, 1.1285e-04, 1.1075e-04, + 1.0626e-04, 1.2629e-04], device='cuda:2') +2023-04-27 23:29:25,331 INFO [optim.py:369] (2/7) 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 23:29:36,261 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1050, 0.8112, 0.8432, 0.8671, 1.2349, 0.9624, 0.9435, 0.9106], + device='cuda:2'), covar=tensor([0.1793, 0.1614, 0.2282, 0.1790, 0.1285, 0.1664, 0.1997, 0.2682], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0306, 0.0350, 0.0285, 0.0328, 0.0304, 0.0297, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.4115e-05, 6.2989e-05, 7.3501e-05, 5.7164e-05, 6.7313e-05, 6.3437e-05, + 6.1611e-05, 7.9403e-05], device='cuda:2') +2023-04-27 23:29:50,047 INFO [finetune.py:976] (2/7) Epoch 26, batch 1800, loss[loss=0.1719, simple_loss=0.2506, pruned_loss=0.04664, over 4740.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2425, pruned_loss=0.04784, over 955544.68 frames. ], batch size: 27, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:29:53,383 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144996.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:29:56,771 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-04-27 23:30:15,058 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:24,556 INFO [finetune.py:976] (2/7) Epoch 26, batch 1850, loss[loss=0.1214, simple_loss=0.1925, pruned_loss=0.02512, over 4459.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2421, pruned_loss=0.04746, over 955013.12 frames. ], batch size: 19, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:30:34,393 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145057.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:30:41,226 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145067.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:44,630 INFO [optim.py:369] (2/7) 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 23:30:55,905 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145088.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:30:58,303 INFO [finetune.py:976] (2/7) Epoch 26, batch 1900, loss[loss=0.1693, simple_loss=0.2557, pruned_loss=0.04145, over 4843.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2436, pruned_loss=0.04775, over 957129.63 frames. ], batch size: 44, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:31:13,799 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145115.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:31:17,405 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-04-27 23:31:22,137 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 26, batch 1950, loss[loss=0.1919, simple_loss=0.2587, pruned_loss=0.06256, over 4827.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2421, pruned_loss=0.04705, over 957679.24 frames. ], batch size: 39, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:32:03,281 INFO [optim.py:369] (2/7) 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 23:32:23,668 INFO [finetune.py:976] (2/7) Epoch 26, batch 2000, loss[loss=0.1281, simple_loss=0.2076, pruned_loss=0.02428, over 4861.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2403, pruned_loss=0.04679, over 957536.75 frames. ], batch size: 34, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:32:48,924 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7220, 1.4008, 1.4058, 1.4981, 1.9079, 1.5694, 1.3861, 1.3341], + device='cuda:2'), covar=tensor([0.1548, 0.1391, 0.1547, 0.1466, 0.0907, 0.1526, 0.1675, 0.1917], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0306, 0.0348, 0.0284, 0.0326, 0.0303, 0.0295, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.3610e-05, 6.2920e-05, 7.3164e-05, 5.6831e-05, 6.6880e-05, 6.3297e-05, + 6.1141e-05, 7.8985e-05], device='cuda:2') +2023-04-27 23:33:03,263 INFO [finetune.py:976] (2/7) Epoch 26, batch 2050, loss[loss=0.1634, simple_loss=0.2389, pruned_loss=0.04391, over 4813.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2377, pruned_loss=0.04609, over 954955.82 frames. ], batch size: 41, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:33:42,766 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5482, 1.7309, 1.3899, 1.2069, 1.2015, 1.1565, 1.4389, 1.1238], + device='cuda:2'), covar=tensor([0.1616, 0.1270, 0.1460, 0.1573, 0.2337, 0.1955, 0.1022, 0.2076], + device='cuda:2'), in_proj_covar=tensor([0.0196, 0.0209, 0.0168, 0.0203, 0.0199, 0.0185, 0.0156, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 23:33:43,224 INFO [optim.py:369] (2/7) 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 23:33:58,466 INFO [finetune.py:976] (2/7) Epoch 26, batch 2100, loss[loss=0.1722, simple_loss=0.2499, pruned_loss=0.0473, over 4916.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2376, pruned_loss=0.04616, over 955654.98 frames. ], batch size: 36, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:34:02,161 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8959, 2.1826, 2.1187, 2.2510, 1.9699, 2.1129, 2.1689, 2.0847], + device='cuda:2'), covar=tensor([0.3546, 0.5865, 0.4602, 0.4183, 0.5555, 0.6720, 0.5612, 0.5545], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0376, 0.0330, 0.0341, 0.0351, 0.0395, 0.0362, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:34:32,515 INFO [finetune.py:976] (2/7) Epoch 26, batch 2150, loss[loss=0.1363, simple_loss=0.2119, pruned_loss=0.03033, over 4769.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2409, pruned_loss=0.04741, over 956561.78 frames. ], batch size: 26, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:34:39,185 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145352.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 23:34:51,070 INFO [optim.py:369] (2/7) 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 23:34:51,204 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145383.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:35:00,503 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0823, 1.5674, 5.3182, 4.9574, 4.5660, 5.1284, 4.7173, 4.6821], + device='cuda:2'), covar=tensor([0.6718, 0.5885, 0.0915, 0.1827, 0.1086, 0.1113, 0.0966, 0.1534], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0308, 0.0408, 0.0410, 0.0348, 0.0415, 0.0319, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:35:10,686 INFO [finetune.py:976] (2/7) Epoch 26, batch 2200, loss[loss=0.171, simple_loss=0.2555, pruned_loss=0.04332, over 4806.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2427, pruned_loss=0.04762, over 957230.38 frames. ], batch size: 51, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:35:30,051 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145422.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:35:37,105 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145432.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:35:44,059 INFO [finetune.py:976] (2/7) Epoch 26, batch 2250, loss[loss=0.1903, simple_loss=0.2557, pruned_loss=0.06239, over 4772.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2441, pruned_loss=0.04862, over 955581.83 frames. ], batch size: 26, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:36:03,179 INFO [optim.py:369] (2/7) 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 23:36:03,890 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9135, 2.2616, 0.9787, 1.2426, 1.6160, 1.1102, 2.4604, 1.3856], + device='cuda:2'), covar=tensor([0.0733, 0.0659, 0.0697, 0.1374, 0.0469, 0.1102, 0.0366, 0.0707], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-27 23:36:17,802 INFO [finetune.py:976] (2/7) Epoch 26, batch 2300, loss[loss=0.1424, simple_loss=0.2151, pruned_loss=0.03484, over 4865.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2437, pruned_loss=0.04813, over 957111.68 frames. ], batch size: 31, lr: 2.97e-03, grad_scale: 32.0 +2023-04-27 23:36:19,722 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3906, 1.5594, 1.4913, 1.8126, 1.6729, 1.8695, 1.4705, 3.0321], + device='cuda:2'), covar=tensor([0.0673, 0.0867, 0.0841, 0.1122, 0.0650, 0.0730, 0.0831, 0.0232], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0013, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-27 23:36:49,874 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2718, 2.5067, 0.8473, 1.5079, 1.4805, 1.9284, 1.6055, 0.8167], + device='cuda:2'), covar=tensor([0.1258, 0.0970, 0.1710, 0.1200, 0.1092, 0.0853, 0.1368, 0.1719], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0237, 0.0134, 0.0120, 0.0130, 0.0151, 0.0116, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:36:51,030 INFO [finetune.py:976] (2/7) Epoch 26, batch 2350, loss[loss=0.1665, simple_loss=0.2309, pruned_loss=0.051, over 4712.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2414, pruned_loss=0.04773, over 955320.25 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:37:10,060 INFO [optim.py:369] (2/7) 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 23:37:40,488 INFO [finetune.py:976] (2/7) Epoch 26, batch 2400, loss[loss=0.14, simple_loss=0.2068, pruned_loss=0.03663, over 4793.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2388, pruned_loss=0.04745, over 955012.11 frames. ], batch size: 29, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:37:41,822 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145594.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:38:03,936 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-04-27 23:38:46,679 INFO [finetune.py:976] (2/7) Epoch 26, batch 2450, loss[loss=0.1898, simple_loss=0.2586, pruned_loss=0.06049, over 4890.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.236, pruned_loss=0.04657, over 955252.37 frames. ], batch size: 35, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:38:47,933 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145643.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:39:00,512 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145652.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:39:08,349 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145655.0, num_to_drop=1, layers_to_drop={2} +2023-04-27 23:39:30,139 INFO [optim.py:369] (2/7) 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 23:39:42,992 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145683.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:39:48,405 INFO [finetune.py:976] (2/7) Epoch 26, batch 2500, loss[loss=0.2056, simple_loss=0.279, pruned_loss=0.06608, over 4809.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.237, pruned_loss=0.04635, over 954145.54 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:39:54,790 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145700.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:39:57,802 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145704.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:09,662 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145722.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:12,703 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145727.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:15,130 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145731.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:20,104 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-04-27 23:40:22,234 INFO [finetune.py:976] (2/7) Epoch 26, batch 2550, loss[loss=0.1898, simple_loss=0.2637, pruned_loss=0.05791, over 4863.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2412, pruned_loss=0.04718, over 954894.67 frames. ], batch size: 34, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:40:41,743 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145770.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:40:42,277 INFO [optim.py:369] (2/7) 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 23:40:56,106 INFO [finetune.py:976] (2/7) Epoch 26, batch 2600, loss[loss=0.1644, simple_loss=0.2455, pruned_loss=0.04166, over 4760.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2425, pruned_loss=0.04762, over 954548.78 frames. ], batch size: 27, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:41:09,799 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-04-27 23:41:14,621 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-27 23:41:16,298 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145821.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:41:21,609 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5465, 1.1314, 4.1830, 3.6121, 3.7034, 3.9522, 3.8185, 3.4888], + device='cuda:2'), covar=tensor([0.9638, 0.8509, 0.1579, 0.3154, 0.2082, 0.3144, 0.2598, 0.2815], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0309, 0.0410, 0.0409, 0.0350, 0.0417, 0.0320, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:41:29,915 INFO [finetune.py:976] (2/7) Epoch 26, batch 2650, loss[loss=0.1645, simple_loss=0.2397, pruned_loss=0.0447, over 4830.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2431, pruned_loss=0.04766, over 954193.67 frames. ], batch size: 47, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:41:33,096 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8621, 1.4655, 1.7034, 1.7031, 1.7270, 1.3989, 0.8152, 1.4933], + device='cuda:2'), covar=tensor([0.3022, 0.2877, 0.1596, 0.1976, 0.2102, 0.2354, 0.3903, 0.1758], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0246, 0.0228, 0.0314, 0.0221, 0.0235, 0.0227, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 23:41:49,916 INFO [optim.py:369] (2/7) 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 23:41:57,240 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145882.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:42:03,155 INFO [finetune.py:976] (2/7) Epoch 26, batch 2700, loss[loss=0.2, simple_loss=0.2712, pruned_loss=0.06439, over 4686.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2425, pruned_loss=0.04699, over 955256.91 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:42:14,448 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-27 23:42:36,792 INFO [finetune.py:976] (2/7) Epoch 26, batch 2750, loss[loss=0.1656, simple_loss=0.2321, pruned_loss=0.04956, over 4787.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2414, pruned_loss=0.04743, over 956000.09 frames. ], batch size: 26, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:42:42,208 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145950.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:42:56,327 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6612, 1.5185, 1.8710, 2.0138, 1.5011, 1.3898, 1.6277, 1.0634], + device='cuda:2'), covar=tensor([0.0477, 0.0715, 0.0391, 0.0524, 0.0704, 0.1023, 0.0595, 0.0577], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0068, 0.0067, 0.0069, 0.0075, 0.0095, 0.0073, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:42:56,787 INFO [optim.py:369] (2/7) 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 23:43:04,376 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-27 23:43:10,066 INFO [finetune.py:976] (2/7) Epoch 26, batch 2800, loss[loss=0.1439, simple_loss=0.2061, pruned_loss=0.04083, over 4821.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2367, pruned_loss=0.04534, over 955710.69 frames. ], batch size: 51, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:43:14,390 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145999.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:43:32,461 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-04-27 23:43:35,186 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146027.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:43:40,248 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-27 23:43:44,744 INFO [finetune.py:976] (2/7) Epoch 26, batch 2850, loss[loss=0.2049, simple_loss=0.2736, pruned_loss=0.0681, over 4892.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2362, pruned_loss=0.04551, over 956706.74 frames. ], batch size: 35, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:44:06,241 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1623, 0.7820, 0.9572, 0.7905, 1.2204, 1.0037, 0.9084, 0.9692], + device='cuda:2'), covar=tensor([0.1804, 0.1624, 0.2044, 0.1795, 0.1011, 0.1451, 0.1501, 0.2202], + device='cuda:2'), in_proj_covar=tensor([0.0311, 0.0307, 0.0348, 0.0284, 0.0325, 0.0302, 0.0297, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.3470e-05, 6.3253e-05, 7.3149e-05, 5.6773e-05, 6.6632e-05, 6.2975e-05, + 6.1431e-05, 7.8762e-05], device='cuda:2') +2023-04-27 23:44:22,216 INFO [optim.py:369] (2/7) 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 23:44:25,239 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 26, batch 2900, loss[loss=0.1868, simple_loss=0.267, pruned_loss=0.05331, over 4825.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2386, pruned_loss=0.04597, over 955108.65 frames. ], batch size: 51, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:45:49,355 INFO [finetune.py:976] (2/7) Epoch 26, batch 2950, loss[loss=0.2165, simple_loss=0.2807, pruned_loss=0.07613, over 4827.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2416, pruned_loss=0.04746, over 954288.81 frames. ], batch size: 39, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:45:59,491 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146149.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:46:11,515 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3274, 1.5441, 1.3170, 1.5407, 1.2434, 1.3090, 1.2877, 1.1115], + device='cuda:2'), covar=tensor([0.1709, 0.1397, 0.0944, 0.1079, 0.3502, 0.1108, 0.1674, 0.2039], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0299, 0.0215, 0.0276, 0.0314, 0.0254, 0.0248, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1373e-04, 1.1792e-04, 8.4636e-05, 1.0860e-04, 1.2657e-04, 9.9862e-05, + 9.9947e-05, 1.0382e-04], device='cuda:2') +2023-04-27 23:46:13,680 INFO [optim.py:369] (2/7) 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 23:46:17,216 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9088, 2.2396, 1.8463, 2.2430, 1.5550, 1.8921, 1.8670, 1.5420], + device='cuda:2'), covar=tensor([0.1972, 0.1414, 0.0819, 0.0988, 0.3492, 0.1164, 0.1921, 0.2450], + device='cuda:2'), in_proj_covar=tensor([0.0285, 0.0299, 0.0215, 0.0276, 0.0314, 0.0254, 0.0248, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1368e-04, 1.1784e-04, 8.4600e-05, 1.0853e-04, 1.2649e-04, 9.9761e-05, + 9.9895e-05, 1.0372e-04], device='cuda:2') +2023-04-27 23:46:18,906 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146177.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:46:28,438 INFO [finetune.py:976] (2/7) Epoch 26, batch 3000, loss[loss=0.2006, simple_loss=0.2671, pruned_loss=0.06711, over 4085.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2435, pruned_loss=0.04768, over 952623.36 frames. ], batch size: 65, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:46:28,438 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-27 23:46:38,926 INFO [finetune.py:1010] (2/7) Epoch 26, validation: loss=0.1526, simple_loss=0.2216, pruned_loss=0.04183, over 2265189.00 frames. +2023-04-27 23:46:38,926 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6400MB +2023-04-27 23:46:50,693 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146210.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:46:53,218 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-04-27 23:47:11,568 INFO [finetune.py:976] (2/7) Epoch 26, batch 3050, loss[loss=0.1767, simple_loss=0.2496, pruned_loss=0.05189, over 4815.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2438, pruned_loss=0.04747, over 953689.12 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 64.0 +2023-04-27 23:47:17,395 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146250.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:47:27,535 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0911, 2.3855, 1.9763, 2.3368, 1.6979, 2.0783, 2.0972, 1.5522], + device='cuda:2'), covar=tensor([0.1733, 0.1287, 0.0865, 0.1172, 0.2965, 0.1036, 0.1850, 0.2379], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0299, 0.0215, 0.0275, 0.0313, 0.0253, 0.0247, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1342e-04, 1.1760e-04, 8.4519e-05, 1.0844e-04, 1.2605e-04, 9.9637e-05, + 9.9717e-05, 1.0362e-04], device='cuda:2') +2023-04-27 23:47:31,020 INFO [optim.py:369] (2/7) 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 23:47:45,139 INFO [finetune.py:976] (2/7) Epoch 26, batch 3100, loss[loss=0.1539, simple_loss=0.2179, pruned_loss=0.04497, over 4915.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2415, pruned_loss=0.04697, over 951133.12 frames. ], batch size: 46, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:47:49,793 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146298.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:47:50,445 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146299.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:48:18,962 INFO [finetune.py:976] (2/7) Epoch 26, batch 3150, loss[loss=0.1885, simple_loss=0.2471, pruned_loss=0.06501, over 4820.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2391, pruned_loss=0.04645, over 952540.99 frames. ], batch size: 33, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:48:22,566 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 23:48:51,066 INFO [finetune.py:976] (2/7) Epoch 26, batch 3200, loss[loss=0.2011, simple_loss=0.2562, pruned_loss=0.073, over 4917.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2358, pruned_loss=0.04574, over 952578.94 frames. ], batch size: 37, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:48:51,288 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-04-27 23:48:59,729 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.1507, 4.1578, 3.0384, 4.8188, 4.1101, 4.2233, 1.4803, 4.1360], + device='cuda:2'), covar=tensor([0.1747, 0.1164, 0.3843, 0.1016, 0.5268, 0.1650, 0.6308, 0.2274], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0220, 0.0253, 0.0306, 0.0301, 0.0247, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:49:05,531 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8000, 1.4688, 2.0067, 2.2666, 1.8837, 1.8168, 1.9510, 1.8217], + device='cuda:2'), covar=tensor([0.4363, 0.6433, 0.5506, 0.5466, 0.5495, 0.6937, 0.6962, 0.8871], + device='cuda:2'), in_proj_covar=tensor([0.0441, 0.0422, 0.0515, 0.0507, 0.0470, 0.0506, 0.0506, 0.0518], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:49:24,697 INFO [finetune.py:976] (2/7) Epoch 26, batch 3250, loss[loss=0.1412, simple_loss=0.2233, pruned_loss=0.02953, over 4823.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2356, pruned_loss=0.04553, over 954577.75 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:49:45,352 INFO [optim.py:369] (2/7) 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 23:49:46,059 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3781, 3.1602, 0.7228, 1.5931, 1.5561, 2.2922, 1.7483, 1.1441], + device='cuda:2'), covar=tensor([0.1868, 0.1810, 0.2705, 0.1974, 0.1556, 0.1419, 0.1925, 0.2281], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0239, 0.0136, 0.0121, 0.0132, 0.0153, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:49:53,885 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146477.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:49:56,405 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8007, 2.3276, 1.8234, 1.7523, 1.3289, 1.3511, 1.9405, 1.2566], + device='cuda:2'), covar=tensor([0.1695, 0.1369, 0.1364, 0.1612, 0.2273, 0.1979, 0.0930, 0.2111], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0210, 0.0170, 0.0204, 0.0200, 0.0186, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-27 23:49:56,472 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-04-27 23:50:13,005 INFO [finetune.py:976] (2/7) Epoch 26, batch 3300, loss[loss=0.1479, simple_loss=0.2275, pruned_loss=0.03419, over 4772.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2397, pruned_loss=0.04678, over 955588.79 frames. ], batch size: 26, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:50:28,267 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146505.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:50:57,305 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146525.0, num_to_drop=1, layers_to_drop={0} +2023-04-27 23:51:19,415 INFO [finetune.py:976] (2/7) Epoch 26, batch 3350, loss[loss=0.129, simple_loss=0.1943, pruned_loss=0.03189, over 3998.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2412, pruned_loss=0.0468, over 955506.15 frames. ], batch size: 17, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:51:39,284 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0433, 2.5922, 2.2809, 2.6104, 2.0203, 2.2616, 2.3661, 1.7262], + device='cuda:2'), covar=tensor([0.2192, 0.1140, 0.0764, 0.1056, 0.2926, 0.1162, 0.2135, 0.2419], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0303, 0.0218, 0.0280, 0.0317, 0.0256, 0.0251, 0.0267], + device='cuda:2'), out_proj_covar=tensor([1.1525e-04, 1.1928e-04, 8.5685e-05, 1.1011e-04, 1.2767e-04, 1.0085e-04, + 1.0118e-04, 1.0533e-04], device='cuda:2') +2023-04-27 23:51:45,685 INFO [optim.py:369] (2/7) 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 23:51:48,973 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-04-27 23:51:51,407 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-04-27 23:51:57,899 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146591.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:51:58,398 INFO [finetune.py:976] (2/7) Epoch 26, batch 3400, loss[loss=0.1988, simple_loss=0.2689, pruned_loss=0.06433, over 4862.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.243, pruned_loss=0.04778, over 956963.71 frames. ], batch size: 34, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:52:27,906 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146626.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:52:43,201 INFO [finetune.py:976] (2/7) Epoch 26, batch 3450, loss[loss=0.1878, simple_loss=0.2547, pruned_loss=0.0604, over 4895.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.242, pruned_loss=0.04734, over 957678.50 frames. ], batch size: 35, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:52:47,441 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-04-27 23:52:49,828 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 23:53:13,321 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9632, 3.9714, 2.8851, 4.6142, 4.0401, 4.0307, 1.5479, 3.7714], + device='cuda:2'), covar=tensor([0.1605, 0.1293, 0.2920, 0.1310, 0.2426, 0.1673, 0.6138, 0.2364], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0220, 0.0254, 0.0306, 0.0300, 0.0247, 0.0274, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:53:13,978 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146687.0, num_to_drop=1, layers_to_drop={1} +2023-04-27 23:53:16,864 INFO [finetune.py:976] (2/7) Epoch 26, batch 3500, loss[loss=0.1846, simple_loss=0.2507, pruned_loss=0.05927, over 4849.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2393, pruned_loss=0.04676, over 958322.44 frames. ], batch size: 49, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:53:50,652 INFO [finetune.py:976] (2/7) Epoch 26, batch 3550, loss[loss=0.1849, simple_loss=0.2479, pruned_loss=0.06099, over 4935.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2373, pruned_loss=0.04656, over 957051.38 frames. ], batch size: 33, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:54:00,404 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1769, 2.0505, 2.6663, 2.8927, 1.8256, 1.7721, 2.1561, 1.2474], + device='cuda:2'), covar=tensor([0.0478, 0.0643, 0.0292, 0.0481, 0.0765, 0.0940, 0.0659, 0.0691], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0068, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-27 23:54:11,377 INFO [optim.py:369] (2/7) 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 23:54:14,460 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3194, 1.5723, 2.0686, 2.5484, 2.2098, 1.6368, 1.4176, 1.9418], + device='cuda:2'), covar=tensor([0.3442, 0.3944, 0.1958, 0.2743, 0.2733, 0.2911, 0.4536, 0.2178], + device='cuda:2'), in_proj_covar=tensor([0.0292, 0.0245, 0.0227, 0.0313, 0.0220, 0.0234, 0.0226, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-27 23:54:24,572 INFO [finetune.py:976] (2/7) Epoch 26, batch 3600, loss[loss=0.2164, simple_loss=0.2981, pruned_loss=0.0673, over 4088.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2361, pruned_loss=0.04628, over 955864.56 frames. ], batch size: 65, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:54:32,798 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146805.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:54:56,341 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-04-27 23:54:59,555 INFO [finetune.py:976] (2/7) Epoch 26, batch 3650, loss[loss=0.1873, simple_loss=0.2612, pruned_loss=0.05675, over 4806.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2387, pruned_loss=0.04688, over 955361.27 frames. ], batch size: 41, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:55:00,917 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146844.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:55:11,633 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 23:55:34,934 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 23:55:48,366 INFO [finetune.py:976] (2/7) Epoch 26, batch 3700, loss[loss=0.1623, simple_loss=0.2402, pruned_loss=0.04225, over 4925.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2416, pruned_loss=0.04767, over 952900.57 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:56:01,872 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146905.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:56:36,153 INFO [finetune.py:976] (2/7) Epoch 26, batch 3750, loss[loss=0.2149, simple_loss=0.2768, pruned_loss=0.07646, over 4824.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2433, pruned_loss=0.04863, over 952177.45 frames. ], batch size: 33, lr: 2.96e-03, grad_scale: 32.0 +2023-04-27 23:56:44,469 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146947.0, num_to_drop=0, layers_to_drop=set() +2023-04-27 23:57:06,090 INFO [optim.py:369] (2/7) 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 23:57:09,745 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7975, 1.2977, 1.8975, 2.2987, 1.8601, 1.7885, 1.8508, 1.7891], + device='cuda:2'), covar=tensor([0.4354, 0.6647, 0.5996, 0.5206, 0.5555, 0.7929, 0.7400, 0.8232], + device='cuda:2'), in_proj_covar=tensor([0.0440, 0.0422, 0.0516, 0.0508, 0.0470, 0.0505, 0.0507, 0.0520], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-27 23:57:13,806 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146982.0, num_to_drop=1, layers_to_drop={3} +2023-04-27 23:57:20,816 INFO [finetune.py:976] (2/7) Epoch 26, batch 3800, loss[loss=0.1636, simple_loss=0.2374, pruned_loss=0.04488, over 4811.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2441, pruned_loss=0.04916, over 952743.07 frames. ], batch size: 47, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:58:05,507 INFO [finetune.py:976] (2/7) Epoch 26, batch 3850, loss[loss=0.1349, simple_loss=0.2166, pruned_loss=0.02665, over 4760.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2423, pruned_loss=0.04835, over 953115.20 frames. ], batch size: 26, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:58:08,742 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-27 23:58:23,781 INFO [optim.py:369] (2/7) 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 23:58:39,180 INFO [finetune.py:976] (2/7) Epoch 26, batch 3900, loss[loss=0.179, simple_loss=0.2482, pruned_loss=0.05491, over 4720.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2389, pruned_loss=0.04704, over 953906.94 frames. ], batch size: 54, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:59:08,828 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-27 23:59:12,137 INFO [finetune.py:976] (2/7) Epoch 26, batch 3950, loss[loss=0.1827, simple_loss=0.2532, pruned_loss=0.05612, over 4819.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2367, pruned_loss=0.04592, over 954745.34 frames. ], batch size: 41, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:59:17,531 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2476, 1.4803, 1.7476, 1.8316, 1.7806, 1.8854, 1.8024, 1.7782], + device='cuda:2'), covar=tensor([0.3696, 0.4782, 0.3857, 0.3861, 0.4940, 0.6461, 0.4161, 0.4057], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0375, 0.0329, 0.0339, 0.0351, 0.0395, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-27 23:59:31,437 INFO [optim.py:369] (2/7) 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 23:59:35,984 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-04-27 23:59:45,578 INFO [finetune.py:976] (2/7) Epoch 26, batch 4000, loss[loss=0.1847, simple_loss=0.2615, pruned_loss=0.05396, over 4279.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2371, pruned_loss=0.04649, over 954655.68 frames. ], batch size: 65, lr: 2.95e-03, grad_scale: 32.0 +2023-04-27 23:59:51,633 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147200.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:00:02,098 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-28 00:00:18,694 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147241.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:00:19,194 INFO [finetune.py:976] (2/7) Epoch 26, batch 4050, loss[loss=0.1566, simple_loss=0.2296, pruned_loss=0.04182, over 4767.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2415, pruned_loss=0.0482, over 955286.59 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:00:23,238 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147247.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:00:34,047 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3423, 2.5922, 0.8488, 1.4122, 1.5878, 1.9476, 1.6017, 0.9688], + device='cuda:2'), covar=tensor([0.1428, 0.1114, 0.1793, 0.1434, 0.1262, 0.0970, 0.1685, 0.1611], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0137, 0.0122, 0.0133, 0.0154, 0.0118, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:00:44,478 INFO [optim.py:369] (2/7) 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 00:00:55,646 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147282.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:10,421 INFO [finetune.py:976] (2/7) Epoch 26, batch 4100, loss[loss=0.146, simple_loss=0.2197, pruned_loss=0.03616, over 4830.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2425, pruned_loss=0.04844, over 954085.66 frames. ], batch size: 49, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:01:12,317 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147295.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:23,295 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147302.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:34,378 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147310.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:01:54,624 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147330.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:02:13,693 INFO [finetune.py:976] (2/7) Epoch 26, batch 4150, loss[loss=0.1968, simple_loss=0.2644, pruned_loss=0.06457, over 4885.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2439, pruned_loss=0.04857, over 953296.90 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:02:28,204 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-28 00:02:56,062 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147371.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:02:56,541 INFO [optim.py:369] (2/7) 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 00:03:00,397 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3174, 1.7808, 1.5983, 2.2105, 2.3457, 1.8835, 1.9663, 1.6213], + device='cuda:2'), covar=tensor([0.1736, 0.1707, 0.2161, 0.1433, 0.1178, 0.2168, 0.1741, 0.2275], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0310, 0.0353, 0.0287, 0.0327, 0.0305, 0.0298, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.4441e-05, 6.3801e-05, 7.4137e-05, 5.7395e-05, 6.7015e-05, 6.3690e-05, + 6.1754e-05, 7.9367e-05], device='cuda:2') +2023-04-28 00:03:09,328 INFO [finetune.py:976] (2/7) Epoch 26, batch 4200, loss[loss=0.1665, simple_loss=0.2348, pruned_loss=0.04913, over 4822.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2439, pruned_loss=0.04824, over 955076.24 frames. ], batch size: 39, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:03:18,213 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3757, 1.3516, 1.5853, 1.6716, 1.3044, 1.2068, 1.3750, 0.8911], + device='cuda:2'), covar=tensor([0.0588, 0.0681, 0.0415, 0.0531, 0.0692, 0.1048, 0.0657, 0.0549], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:03:21,139 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147407.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:03:37,864 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147434.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:03:42,686 INFO [finetune.py:976] (2/7) Epoch 26, batch 4250, loss[loss=0.1552, simple_loss=0.2282, pruned_loss=0.04109, over 4826.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2429, pruned_loss=0.04817, over 952112.64 frames. ], batch size: 39, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:04:01,795 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 00:04:06,096 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2650, 1.8936, 2.1142, 2.4601, 2.4369, 1.9100, 1.7551, 2.2886], + device='cuda:2'), covar=tensor([0.0732, 0.1172, 0.0721, 0.0559, 0.0644, 0.0908, 0.0811, 0.0537], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0202, 0.0184, 0.0171, 0.0177, 0.0178, 0.0150, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:04:16,302 INFO [finetune.py:976] (2/7) Epoch 26, batch 4300, loss[loss=0.1533, simple_loss=0.2151, pruned_loss=0.04579, over 4906.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2402, pruned_loss=0.04784, over 954799.66 frames. ], batch size: 36, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:04:18,244 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147495.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:04:22,178 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147500.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:04:29,290 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0252, 1.1901, 1.7326, 1.8254, 1.7337, 1.8588, 1.7105, 1.6941], + device='cuda:2'), covar=tensor([0.3697, 0.4986, 0.4276, 0.4406, 0.5489, 0.6899, 0.4455, 0.4395], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0377, 0.0330, 0.0341, 0.0352, 0.0397, 0.0362, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:04:46,752 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-04-28 00:04:49,577 INFO [finetune.py:976] (2/7) Epoch 26, batch 4350, loss[loss=0.1519, simple_loss=0.2216, pruned_loss=0.04107, over 4730.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2371, pruned_loss=0.04667, over 954738.31 frames. ], batch size: 54, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:04:51,100 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-28 00:04:53,338 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147548.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:05:00,890 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3795, 1.7030, 1.7195, 2.0601, 2.0131, 2.0316, 1.6624, 4.4882], + device='cuda:2'), covar=tensor([0.0552, 0.0841, 0.0753, 0.1162, 0.0620, 0.0505, 0.0738, 0.0104], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 00:05:08,013 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1099, 0.7698, 0.9455, 0.7661, 1.2150, 0.9956, 0.8791, 0.9706], + device='cuda:2'), covar=tensor([0.2274, 0.1754, 0.2492, 0.2062, 0.1289, 0.1809, 0.1971, 0.2996], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0310, 0.0352, 0.0287, 0.0327, 0.0305, 0.0299, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.4333e-05, 6.3751e-05, 7.3924e-05, 5.7396e-05, 6.6824e-05, 6.3611e-05, + 6.1775e-05, 7.9443e-05], device='cuda:2') +2023-04-28 00:05:10,813 INFO [optim.py:369] (2/7) 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 00:05:12,725 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9994, 1.0478, 1.1551, 1.1712, 0.9906, 0.9094, 1.0078, 0.5348], + device='cuda:2'), covar=tensor([0.0579, 0.0613, 0.0483, 0.0499, 0.0710, 0.1201, 0.0464, 0.0683], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0068, 0.0073, 0.0093, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:05:23,016 INFO [finetune.py:976] (2/7) Epoch 26, batch 4400, loss[loss=0.1962, simple_loss=0.2842, pruned_loss=0.05414, over 4841.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2374, pruned_loss=0.04674, over 953661.86 frames. ], batch size: 49, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:05:23,826 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-04-28 00:05:26,139 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147597.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:05:47,720 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6531, 3.7992, 0.8488, 1.9917, 2.1620, 2.5849, 2.2192, 1.0937], + device='cuda:2'), covar=tensor([0.1282, 0.1081, 0.2011, 0.1256, 0.1004, 0.1051, 0.1405, 0.2146], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0239, 0.0136, 0.0122, 0.0132, 0.0153, 0.0118, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:05:56,084 INFO [finetune.py:976] (2/7) Epoch 26, batch 4450, loss[loss=0.2052, simple_loss=0.2844, pruned_loss=0.06304, over 4808.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2415, pruned_loss=0.04775, over 952575.99 frames. ], batch size: 45, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:06:01,709 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3496, 1.3149, 1.5643, 1.6075, 1.2907, 1.1494, 1.3940, 0.9945], + device='cuda:2'), covar=tensor([0.0548, 0.0461, 0.0413, 0.0484, 0.0634, 0.0869, 0.0463, 0.0522], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:06:17,539 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147659.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:06:22,788 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147666.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:06:31,475 INFO [optim.py:369] (2/7) 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 00:06:38,231 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 00:06:44,114 INFO [finetune.py:976] (2/7) Epoch 26, batch 4500, loss[loss=0.1714, simple_loss=0.2541, pruned_loss=0.04439, over 4762.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2432, pruned_loss=0.04836, over 951389.59 frames. ], batch size: 54, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:07:17,731 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147720.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:07:37,060 INFO [finetune.py:976] (2/7) Epoch 26, batch 4550, loss[loss=0.1593, simple_loss=0.2426, pruned_loss=0.03802, over 4747.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2446, pruned_loss=0.04883, over 950999.28 frames. ], batch size: 27, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:08:01,006 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 00:08:29,095 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8185, 1.3725, 1.8411, 2.3156, 1.8743, 1.8155, 1.8403, 1.7982], + device='cuda:2'), covar=tensor([0.4378, 0.6442, 0.6079, 0.5197, 0.5750, 0.7481, 0.7921, 0.8398], + device='cuda:2'), in_proj_covar=tensor([0.0444, 0.0424, 0.0518, 0.0509, 0.0471, 0.0507, 0.0509, 0.0521], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:08:31,442 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147790.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:08:32,582 INFO [finetune.py:976] (2/7) Epoch 26, batch 4600, loss[loss=0.1636, simple_loss=0.2402, pruned_loss=0.04349, over 4760.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2421, pruned_loss=0.04709, over 952067.08 frames. ], batch size: 26, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:08:46,312 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-04-28 00:08:47,339 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147815.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:05,600 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0664, 1.4344, 1.3045, 1.6518, 1.5128, 1.4243, 1.4158, 2.4201], + device='cuda:2'), covar=tensor([0.0634, 0.0810, 0.0801, 0.1182, 0.0653, 0.0500, 0.0709, 0.0217], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 00:09:06,109 INFO [finetune.py:976] (2/7) Epoch 26, batch 4650, loss[loss=0.1851, simple_loss=0.2554, pruned_loss=0.05739, over 4929.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2389, pruned_loss=0.04591, over 953818.85 frames. ], batch size: 38, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:09:08,108 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3013, 1.5041, 1.7259, 1.8199, 1.7281, 1.8943, 1.8149, 1.7860], + device='cuda:2'), covar=tensor([0.3009, 0.4170, 0.3524, 0.3891, 0.4843, 0.6048, 0.4060, 0.3919], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0374, 0.0328, 0.0339, 0.0350, 0.0395, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:09:11,683 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5431, 3.4816, 0.9546, 1.8814, 2.0260, 2.4185, 1.9362, 1.1337], + device='cuda:2'), covar=tensor([0.1325, 0.0949, 0.1901, 0.1199, 0.1005, 0.0951, 0.1552, 0.1844], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0237, 0.0135, 0.0120, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:09:15,424 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147857.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:24,993 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1641, 1.7171, 1.9571, 2.4662, 2.4616, 2.0727, 1.7041, 2.2714], + device='cuda:2'), covar=tensor([0.0748, 0.1391, 0.0824, 0.0534, 0.0543, 0.0786, 0.0781, 0.0554], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0202, 0.0184, 0.0171, 0.0177, 0.0178, 0.0151, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:09:25,964 INFO [optim.py:369] (2/7) 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 00:09:29,048 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147876.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:40,055 INFO [finetune.py:976] (2/7) Epoch 26, batch 4700, loss[loss=0.1719, simple_loss=0.235, pruned_loss=0.05436, over 4851.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2362, pruned_loss=0.04525, over 954960.00 frames. ], batch size: 44, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:09:43,170 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147911.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:09:55,897 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147918.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:12,812 INFO [finetune.py:976] (2/7) Epoch 26, batch 4750, loss[loss=0.1965, simple_loss=0.2591, pruned_loss=0.06697, over 4846.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2349, pruned_loss=0.04518, over 956211.45 frames. ], batch size: 49, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:10:15,201 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147945.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:28,064 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147966.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:31,606 INFO [optim.py:369] (2/7) 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 00:10:31,740 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147972.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:10:45,438 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8871, 1.1660, 5.0074, 4.7122, 4.3247, 4.7909, 4.3780, 4.3940], + device='cuda:2'), covar=tensor([0.6697, 0.6361, 0.0897, 0.1532, 0.0982, 0.1392, 0.1695, 0.1454], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0308, 0.0407, 0.0409, 0.0347, 0.0413, 0.0318, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:10:46,577 INFO [finetune.py:976] (2/7) Epoch 26, batch 4800, loss[loss=0.1174, simple_loss=0.1894, pruned_loss=0.02269, over 4734.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2369, pruned_loss=0.04577, over 957118.98 frames. ], batch size: 23, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:10:51,473 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6124, 1.0702, 1.3913, 1.2372, 1.7205, 1.3989, 1.2477, 1.3326], + device='cuda:2'), covar=tensor([0.1802, 0.1794, 0.1890, 0.1535, 0.0991, 0.1809, 0.1808, 0.2534], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0310, 0.0353, 0.0287, 0.0328, 0.0305, 0.0299, 0.0376], + device='cuda:2'), out_proj_covar=tensor([6.4306e-05, 6.3714e-05, 7.4087e-05, 5.7488e-05, 6.7026e-05, 6.3779e-05, + 6.1836e-05, 7.9760e-05], device='cuda:2') +2023-04-28 00:11:00,207 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0194, 2.7169, 2.0752, 2.1101, 1.4709, 1.3714, 2.2178, 1.5225], + device='cuda:2'), covar=tensor([0.1745, 0.1424, 0.1382, 0.1719, 0.2300, 0.2057, 0.0960, 0.2101], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0211, 0.0170, 0.0203, 0.0200, 0.0186, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 00:11:01,948 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148014.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:11:02,583 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148015.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:11:21,740 INFO [finetune.py:976] (2/7) Epoch 26, batch 4850, loss[loss=0.1484, simple_loss=0.2284, pruned_loss=0.03421, over 4726.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2396, pruned_loss=0.04642, over 956486.11 frames. ], batch size: 23, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:11:44,618 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148063.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:11:55,270 INFO [optim.py:369] (2/7) 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 00:12:18,061 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148090.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:12:24,896 INFO [finetune.py:976] (2/7) Epoch 26, batch 4900, loss[loss=0.1821, simple_loss=0.2411, pruned_loss=0.06157, over 4553.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2405, pruned_loss=0.04628, over 955743.16 frames. ], batch size: 20, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:12:47,897 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148111.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:13:08,517 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6708, 1.4789, 1.6254, 2.0166, 2.0410, 1.6442, 1.4194, 1.8488], + device='cuda:2'), covar=tensor([0.0714, 0.1164, 0.0708, 0.0488, 0.0552, 0.0872, 0.0754, 0.0492], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0202, 0.0184, 0.0171, 0.0177, 0.0177, 0.0152, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:13:10,404 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0347, 1.0619, 1.1583, 1.1985, 1.0406, 0.9622, 1.0040, 0.5842], + device='cuda:2'), covar=tensor([0.0526, 0.0595, 0.0481, 0.0566, 0.0683, 0.1178, 0.0447, 0.0612], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0068, 0.0067, 0.0069, 0.0074, 0.0094, 0.0073, 0.0064], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:13:22,194 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 26, batch 4950, loss[loss=0.1666, simple_loss=0.2498, pruned_loss=0.04165, over 4811.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2426, pruned_loss=0.04725, over 955273.02 frames. ], batch size: 39, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:13:52,205 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 00:13:55,279 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9037, 2.1988, 1.9463, 2.2382, 1.5740, 1.9126, 1.8739, 1.4935], + device='cuda:2'), covar=tensor([0.1637, 0.1124, 0.0741, 0.0819, 0.3173, 0.0889, 0.1600, 0.2122], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0298, 0.0214, 0.0276, 0.0313, 0.0253, 0.0248, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1307e-04, 1.1737e-04, 8.4275e-05, 1.0845e-04, 1.2604e-04, 9.9521e-05, + 1.0022e-04, 1.0340e-04], device='cuda:2') +2023-04-28 00:14:06,565 INFO [finetune.py:976] (2/7) Epoch 26, batch 5000, loss[loss=0.1821, simple_loss=0.2516, pruned_loss=0.05632, over 4780.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2412, pruned_loss=0.04645, over 955859.44 frames. ], batch size: 51, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:14:06,662 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9652, 2.3190, 1.4135, 1.7139, 2.4651, 1.7841, 1.7006, 1.9521], + device='cuda:2'), covar=tensor([0.0448, 0.0312, 0.0266, 0.0491, 0.0219, 0.0477, 0.0489, 0.0510], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0051, 0.0046, 0.0038, 0.0053, 0.0039, 0.0050, 0.0050, 0.0052], + device='cuda:2') +2023-04-28 00:14:21,041 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148213.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:14:33,575 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-28 00:14:39,768 INFO [finetune.py:976] (2/7) Epoch 26, batch 5050, loss[loss=0.1674, simple_loss=0.2431, pruned_loss=0.04586, over 4821.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.239, pruned_loss=0.04639, over 956298.46 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 32.0 +2023-04-28 00:14:56,593 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148267.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:14:59,522 INFO [optim.py:369] (2/7) 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 00:15:08,805 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148287.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:15:11,732 INFO [finetune.py:976] (2/7) Epoch 26, batch 5100, loss[loss=0.1992, simple_loss=0.2497, pruned_loss=0.07438, over 4831.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2355, pruned_loss=0.04526, over 956417.73 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 64.0 +2023-04-28 00:15:26,474 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9440, 2.2278, 2.1488, 2.2840, 2.0625, 2.1938, 2.2760, 2.1600], + device='cuda:2'), covar=tensor([0.3578, 0.5644, 0.4873, 0.4704, 0.5745, 0.6646, 0.5504, 0.5094], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0375, 0.0330, 0.0340, 0.0351, 0.0395, 0.0361, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:15:28,678 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148315.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:15:45,038 INFO [finetune.py:976] (2/7) Epoch 26, batch 5150, loss[loss=0.1425, simple_loss=0.217, pruned_loss=0.034, over 4741.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2355, pruned_loss=0.04566, over 954695.80 frames. ], batch size: 27, lr: 2.95e-03, grad_scale: 64.0 +2023-04-28 00:15:49,312 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148348.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:16:00,242 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 00:16:18,408 INFO [finetune.py:976] (2/7) Epoch 26, batch 5200, loss[loss=0.1733, simple_loss=0.2566, pruned_loss=0.04505, over 4815.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2399, pruned_loss=0.04697, over 956079.23 frames. ], batch size: 40, lr: 2.95e-03, grad_scale: 64.0 +2023-04-28 00:16:38,059 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-04-28 00:16:44,040 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148429.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:16:44,069 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4629, 1.7996, 1.9280, 1.9666, 1.8468, 1.8844, 1.9904, 1.9274], + device='cuda:2'), covar=tensor([0.4161, 0.5124, 0.3908, 0.3987, 0.5484, 0.6946, 0.4889, 0.4720], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0374, 0.0329, 0.0339, 0.0350, 0.0394, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:16:51,852 INFO [finetune.py:976] (2/7) Epoch 26, batch 5250, loss[loss=0.1453, simple_loss=0.2339, pruned_loss=0.02834, over 4834.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2426, pruned_loss=0.04797, over 954635.35 frames. ], batch size: 47, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:17:12,090 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 00:17:23,752 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 26, batch 5300, loss[loss=0.1484, simple_loss=0.2329, pruned_loss=0.03195, over 4839.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.244, pruned_loss=0.04836, over 955656.16 frames. ], batch size: 47, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:17:28,537 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148498.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:17:52,364 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148513.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:17:55,673 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-28 00:17:56,055 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148519.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:18:24,821 INFO [finetune.py:976] (2/7) Epoch 26, batch 5350, loss[loss=0.1982, simple_loss=0.2572, pruned_loss=0.06962, over 4773.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2444, pruned_loss=0.0485, over 954555.17 frames. ], batch size: 54, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:18:45,988 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148559.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:18:47,177 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148561.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:18:55,818 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-28 00:18:56,337 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148567.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:18:59,261 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 26, batch 5400, loss[loss=0.1603, simple_loss=0.22, pruned_loss=0.05026, over 4868.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2403, pruned_loss=0.04685, over 955728.80 frames. ], batch size: 31, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:19:50,885 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 00:19:53,633 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148615.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:20:04,088 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-04-28 00:20:11,714 INFO [finetune.py:976] (2/7) Epoch 26, batch 5450, loss[loss=0.1227, simple_loss=0.2014, pruned_loss=0.02203, over 4692.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2378, pruned_loss=0.04627, over 954804.28 frames. ], batch size: 23, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:20:12,391 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148643.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:20:17,270 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148651.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:20:31,291 INFO [optim.py:369] (2/7) 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 00:20:45,353 INFO [finetune.py:976] (2/7) Epoch 26, batch 5500, loss[loss=0.1151, simple_loss=0.185, pruned_loss=0.02258, over 4753.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2349, pruned_loss=0.04546, over 954671.19 frames. ], batch size: 27, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:20:57,703 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148712.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:00,142 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148716.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:17,047 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3927, 1.1639, 4.1390, 3.8680, 3.5970, 3.9267, 3.8432, 3.6042], + device='cuda:2'), covar=tensor([0.7561, 0.6052, 0.1132, 0.1974, 0.1260, 0.1671, 0.2039, 0.1724], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0311, 0.0409, 0.0410, 0.0348, 0.0414, 0.0319, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:21:18,755 INFO [finetune.py:976] (2/7) Epoch 26, batch 5550, loss[loss=0.1882, simple_loss=0.2755, pruned_loss=0.05039, over 4812.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2366, pruned_loss=0.04528, over 955738.15 frames. ], batch size: 41, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:21:38,044 INFO [optim.py:369] (2/7) 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 00:21:41,039 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148777.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:45,569 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148785.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:21:49,593 INFO [finetune.py:976] (2/7) Epoch 26, batch 5600, loss[loss=0.1305, simple_loss=0.1975, pruned_loss=0.03181, over 4196.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2412, pruned_loss=0.04709, over 954807.62 frames. ], batch size: 18, lr: 2.94e-03, grad_scale: 64.0 +2023-04-28 00:21:55,050 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5565, 1.6263, 0.6524, 1.2782, 1.7260, 1.3840, 1.3053, 1.3228], + device='cuda:2'), covar=tensor([0.0631, 0.0363, 0.0392, 0.0635, 0.0293, 0.0690, 0.0672, 0.0705], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0050, 0.0052], + device='cuda:2') +2023-04-28 00:21:58,152 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8154, 1.3109, 1.4843, 1.4257, 1.8986, 1.5722, 1.3445, 1.4195], + device='cuda:2'), covar=tensor([0.1769, 0.1474, 0.1925, 0.1385, 0.0966, 0.1450, 0.1973, 0.2452], + device='cuda:2'), in_proj_covar=tensor([0.0318, 0.0312, 0.0355, 0.0290, 0.0330, 0.0308, 0.0302, 0.0378], + device='cuda:2'), out_proj_covar=tensor([6.4942e-05, 6.4157e-05, 7.4522e-05, 5.8044e-05, 6.7482e-05, 6.4356e-05, + 6.2551e-05, 8.0038e-05], device='cuda:2') +2023-04-28 00:22:14,334 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7490, 1.3738, 1.8222, 2.2031, 1.8106, 1.7186, 1.7792, 1.7597], + device='cuda:2'), covar=tensor([0.4333, 0.6672, 0.6022, 0.5284, 0.5791, 0.7515, 0.7092, 0.8452], + device='cuda:2'), in_proj_covar=tensor([0.0441, 0.0421, 0.0515, 0.0504, 0.0468, 0.0506, 0.0506, 0.0518], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:22:20,663 INFO [finetune.py:976] (2/7) Epoch 26, batch 5650, loss[loss=0.1782, simple_loss=0.2633, pruned_loss=0.04653, over 4814.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2437, pruned_loss=0.04769, over 955103.91 frames. ], batch size: 45, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:22:27,777 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148854.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:22:39,032 INFO [optim.py:369] (2/7) 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 00:22:50,267 INFO [finetune.py:976] (2/7) Epoch 26, batch 5700, loss[loss=0.1428, simple_loss=0.1962, pruned_loss=0.04474, over 4198.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2397, pruned_loss=0.04759, over 933295.53 frames. ], batch size: 18, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:22:55,767 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0239, 0.8535, 0.7883, 0.8925, 1.1665, 0.9416, 0.9382, 0.8159], + device='cuda:2'), covar=tensor([0.2030, 0.1759, 0.2237, 0.1647, 0.1230, 0.1552, 0.1597, 0.2674], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0311, 0.0354, 0.0289, 0.0329, 0.0308, 0.0302, 0.0377], + device='cuda:2'), out_proj_covar=tensor([6.4791e-05, 6.3888e-05, 7.4393e-05, 5.7959e-05, 6.7213e-05, 6.4240e-05, + 6.2484e-05, 7.9881e-05], device='cuda:2') +2023-04-28 00:23:20,154 INFO [finetune.py:976] (2/7) Epoch 27, batch 0, loss[loss=0.1917, simple_loss=0.2654, pruned_loss=0.05895, over 4907.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2654, pruned_loss=0.05895, over 4907.00 frames. ], batch size: 46, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:23:20,155 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 00:23:31,703 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2358, 2.5376, 1.1596, 1.5002, 2.0988, 1.3771, 3.0711, 1.8179], + device='cuda:2'), covar=tensor([0.0616, 0.0749, 0.0692, 0.1160, 0.0358, 0.0844, 0.0252, 0.0516], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0050, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 00:23:32,178 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4940, 1.3779, 1.6706, 1.7118, 1.4040, 1.3207, 1.4782, 0.8828], + device='cuda:2'), covar=tensor([0.0494, 0.0582, 0.0435, 0.0425, 0.0689, 0.1079, 0.0402, 0.0511], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0068, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:23:41,724 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6400MB +2023-04-28 00:24:11,142 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148943.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:24:41,546 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-28 00:24:43,144 INFO [finetune.py:976] (2/7) Epoch 27, batch 50, loss[loss=0.1848, simple_loss=0.2531, pruned_loss=0.05828, over 4786.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2477, pruned_loss=0.04953, over 217349.50 frames. ], batch size: 29, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:24:45,491 INFO [optim.py:369] (2/7) 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 00:24:53,726 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0632, 1.3670, 4.9916, 4.7539, 4.3025, 4.8063, 4.4594, 4.3718], + device='cuda:2'), covar=tensor([0.7091, 0.5995, 0.1041, 0.1648, 0.1140, 0.0963, 0.1348, 0.1644], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0310, 0.0410, 0.0411, 0.0349, 0.0415, 0.0320, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:25:07,487 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148991.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:25:28,980 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 100, loss[loss=0.1349, simple_loss=0.2139, pruned_loss=0.02795, over 4772.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2406, pruned_loss=0.04718, over 381850.44 frames. ], batch size: 26, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:25:48,518 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4789, 2.3194, 2.6599, 2.7728, 2.9255, 2.3293, 2.1841, 2.7603], + device='cuda:2'), covar=tensor([0.0851, 0.0923, 0.0590, 0.0616, 0.0568, 0.0864, 0.0713, 0.0496], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0200, 0.0182, 0.0169, 0.0176, 0.0175, 0.0150, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:25:58,872 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.1997, 4.1482, 2.9754, 4.8384, 4.1155, 4.1977, 2.1252, 4.1737], + device='cuda:2'), covar=tensor([0.1812, 0.0940, 0.3110, 0.1444, 0.2880, 0.1780, 0.5146, 0.2156], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0222, 0.0255, 0.0306, 0.0301, 0.0250, 0.0276, 0.0276], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:26:00,967 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-04-28 00:26:11,901 INFO [finetune.py:976] (2/7) Epoch 27, batch 150, loss[loss=0.1347, simple_loss=0.2029, pruned_loss=0.03328, over 4725.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2356, pruned_loss=0.04588, over 509513.93 frames. ], batch size: 59, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:26:13,186 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149072.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:26:13,726 INFO [optim.py:369] (2/7) 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 00:26:18,476 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.11 vs. limit=5.0 +2023-04-28 00:26:22,647 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149085.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:26:45,180 INFO [finetune.py:976] (2/7) Epoch 27, batch 200, loss[loss=0.1428, simple_loss=0.2239, pruned_loss=0.0309, over 4848.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2325, pruned_loss=0.04472, over 609013.62 frames. ], batch size: 49, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:26:55,608 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149133.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:27:08,547 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149154.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:27:18,794 INFO [finetune.py:976] (2/7) Epoch 27, batch 250, loss[loss=0.2408, simple_loss=0.3033, pruned_loss=0.08911, over 4147.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2361, pruned_loss=0.04618, over 685867.82 frames. ], batch size: 65, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:27:21,648 INFO [optim.py:369] (2/7) 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 00:27:29,594 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7534, 1.7240, 2.2398, 2.3894, 1.5711, 1.4310, 1.8495, 1.0231], + device='cuda:2'), covar=tensor([0.0629, 0.0711, 0.0378, 0.0577, 0.0726, 0.1147, 0.0584, 0.0688], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0075, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:27:34,352 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-28 00:27:35,662 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-04-28 00:27:41,049 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149202.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:27:50,690 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-28 00:27:52,339 INFO [finetune.py:976] (2/7) Epoch 27, batch 300, loss[loss=0.1824, simple_loss=0.2667, pruned_loss=0.04902, over 4827.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2405, pruned_loss=0.04719, over 746099.85 frames. ], batch size: 38, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:27:55,001 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 00:28:22,268 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1126, 1.9333, 4.7354, 4.5049, 4.2427, 4.5361, 4.3752, 4.2752], + device='cuda:2'), covar=tensor([0.6313, 0.5081, 0.0978, 0.1525, 0.0862, 0.1556, 0.1080, 0.1353], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0312, 0.0413, 0.0413, 0.0350, 0.0416, 0.0321, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:28:25,826 INFO [finetune.py:976] (2/7) Epoch 27, batch 350, loss[loss=0.1563, simple_loss=0.2351, pruned_loss=0.03873, over 4871.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2425, pruned_loss=0.04792, over 791915.78 frames. ], batch size: 34, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:28:28,122 INFO [optim.py:369] (2/7) 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 00:28:41,704 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2355, 1.6845, 2.1568, 2.6333, 2.1312, 1.7501, 1.4366, 1.9548], + device='cuda:2'), covar=tensor([0.2997, 0.2867, 0.1484, 0.1800, 0.2558, 0.2435, 0.3814, 0.1813], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0227, 0.0313, 0.0220, 0.0233, 0.0227, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 00:28:51,971 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149307.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:28:59,851 INFO [finetune.py:976] (2/7) Epoch 27, batch 400, loss[loss=0.145, simple_loss=0.2298, pruned_loss=0.03014, over 4777.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2433, pruned_loss=0.04746, over 830737.85 frames. ], batch size: 29, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:29:07,754 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3917, 1.1171, 1.2039, 1.1840, 1.5661, 1.2733, 1.1020, 1.1250], + device='cuda:2'), covar=tensor([0.1727, 0.1236, 0.1605, 0.1226, 0.0703, 0.1401, 0.1729, 0.2152], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0308, 0.0351, 0.0287, 0.0326, 0.0306, 0.0299, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.4305e-05, 6.3354e-05, 7.3556e-05, 5.7557e-05, 6.6794e-05, 6.3813e-05, + 6.1891e-05, 7.9155e-05], device='cuda:2') +2023-04-28 00:29:08,524 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-04-28 00:29:39,546 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149355.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:29:42,716 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1962, 1.4618, 1.6908, 1.7774, 1.7393, 1.8387, 1.7189, 1.7416], + device='cuda:2'), covar=tensor([0.3831, 0.5080, 0.4337, 0.4205, 0.5297, 0.6328, 0.4861, 0.4457], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0376, 0.0331, 0.0340, 0.0352, 0.0396, 0.0360, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:29:48,685 INFO [finetune.py:976] (2/7) Epoch 27, batch 450, loss[loss=0.1595, simple_loss=0.2338, pruned_loss=0.04255, over 4762.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2424, pruned_loss=0.0472, over 858080.59 frames. ], batch size: 27, lr: 2.94e-03, grad_scale: 32.0 +2023-04-28 00:29:50,015 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 00:30:29,030 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-28 00:30:41,613 INFO [finetune.py:976] (2/7) Epoch 27, batch 500, loss[loss=0.1351, simple_loss=0.2054, pruned_loss=0.03234, over 4737.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2391, pruned_loss=0.04668, over 877289.06 frames. ], batch size: 59, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:30:41,677 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149420.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:31:34,474 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-04-28 00:31:43,468 INFO [finetune.py:976] (2/7) Epoch 27, batch 550, loss[loss=0.1504, simple_loss=0.2174, pruned_loss=0.04173, over 4869.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2369, pruned_loss=0.04638, over 895741.06 frames. ], batch size: 34, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:31:45,296 INFO [optim.py:369] (2/7) 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 00:32:08,204 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2879, 1.7706, 2.1368, 2.5708, 2.1337, 1.7065, 1.5028, 1.9918], + device='cuda:2'), covar=tensor([0.3159, 0.3104, 0.1821, 0.2149, 0.2564, 0.2599, 0.3857, 0.1867], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0246, 0.0228, 0.0314, 0.0221, 0.0233, 0.0227, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 00:32:13,122 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 600, loss[loss=0.1156, simple_loss=0.193, pruned_loss=0.01905, over 4766.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2374, pruned_loss=0.0468, over 908789.98 frames. ], batch size: 23, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:32:30,341 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1503, 1.3575, 1.2562, 1.6762, 1.4706, 1.5314, 1.3569, 2.4415], + device='cuda:2'), covar=tensor([0.0617, 0.0883, 0.0879, 0.1272, 0.0701, 0.0473, 0.0774, 0.0247], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 00:32:36,322 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-28 00:32:47,533 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 650, loss[loss=0.1814, simple_loss=0.2567, pruned_loss=0.05307, over 4900.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2417, pruned_loss=0.04785, over 918353.36 frames. ], batch size: 43, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:32:52,255 INFO [optim.py:369] (2/7) 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 00:32:53,619 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149575.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:33:23,677 INFO [finetune.py:976] (2/7) Epoch 27, batch 700, loss[loss=0.1904, simple_loss=0.2704, pruned_loss=0.05519, over 4833.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2429, pruned_loss=0.04766, over 925562.94 frames. ], batch size: 47, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:33:27,503 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149626.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:33:35,926 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-04-28 00:33:53,502 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4500, 1.7144, 1.8763, 1.9403, 1.8288, 1.8588, 1.9147, 1.8994], + device='cuda:2'), covar=tensor([0.4025, 0.5487, 0.4190, 0.4258, 0.5428, 0.7154, 0.4867, 0.4836], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0375, 0.0329, 0.0339, 0.0350, 0.0394, 0.0360, 0.0332], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:33:56,990 INFO [finetune.py:976] (2/7) Epoch 27, batch 750, loss[loss=0.2181, simple_loss=0.2825, pruned_loss=0.07684, over 4819.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2444, pruned_loss=0.0483, over 931609.63 frames. ], batch size: 38, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:33:58,784 INFO [optim.py:369] (2/7) 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 00:34:42,993 INFO [finetune.py:976] (2/7) Epoch 27, batch 800, loss[loss=0.2005, simple_loss=0.2756, pruned_loss=0.06268, over 4884.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2434, pruned_loss=0.04746, over 936743.00 frames. ], batch size: 36, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:35:14,862 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149746.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:35:48,631 INFO [finetune.py:976] (2/7) Epoch 27, batch 850, loss[loss=0.1234, simple_loss=0.2004, pruned_loss=0.02318, over 4753.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2416, pruned_loss=0.04722, over 939834.44 frames. ], batch size: 27, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:35:55,544 INFO [optim.py:369] (2/7) 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 00:36:07,914 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-28 00:36:18,340 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3785, 1.2723, 1.6158, 1.6276, 1.2878, 1.1955, 1.3657, 0.7463], + device='cuda:2'), covar=tensor([0.0537, 0.0754, 0.0439, 0.0668, 0.0750, 0.1165, 0.0625, 0.0624], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0066, 0.0069, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:36:24,427 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 900, loss[loss=0.1961, simple_loss=0.2583, pruned_loss=0.067, over 4793.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2376, pruned_loss=0.046, over 942204.01 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:36:46,736 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7869, 1.2657, 1.8313, 2.2158, 1.8423, 1.7580, 1.7942, 1.7620], + device='cuda:2'), covar=tensor([0.4080, 0.6256, 0.5739, 0.5080, 0.5342, 0.6937, 0.7418, 0.8696], + device='cuda:2'), in_proj_covar=tensor([0.0443, 0.0422, 0.0517, 0.0506, 0.0470, 0.0507, 0.0509, 0.0522], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:37:03,443 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-28 00:37:06,210 INFO [finetune.py:976] (2/7) Epoch 27, batch 950, loss[loss=0.1851, simple_loss=0.247, pruned_loss=0.06161, over 4796.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2363, pruned_loss=0.04603, over 945022.04 frames. ], batch size: 29, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:37:06,275 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149870.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:37:08,538 INFO [optim.py:369] (2/7) 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 00:37:14,070 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9912, 2.2740, 1.8959, 2.2611, 1.6644, 1.9386, 2.0232, 1.6260], + device='cuda:2'), covar=tensor([0.1662, 0.1120, 0.0760, 0.0921, 0.2905, 0.1045, 0.1596, 0.2135], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0301, 0.0215, 0.0277, 0.0315, 0.0254, 0.0249, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1309e-04, 1.1831e-04, 8.4660e-05, 1.0902e-04, 1.2677e-04, 9.9890e-05, + 1.0054e-04, 1.0428e-04], device='cuda:2') +2023-04-28 00:37:17,204 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 00:37:40,039 INFO [finetune.py:976] (2/7) Epoch 27, batch 1000, loss[loss=0.2023, simple_loss=0.267, pruned_loss=0.06879, over 4175.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2393, pruned_loss=0.04739, over 946294.26 frames. ], batch size: 65, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:37:40,686 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149921.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:38:07,557 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2330, 1.7652, 2.0714, 2.2262, 2.0813, 1.6630, 1.1889, 1.8107], + device='cuda:2'), covar=tensor([0.3334, 0.3215, 0.1715, 0.2188, 0.2786, 0.2835, 0.4151, 0.1983], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0244, 0.0227, 0.0312, 0.0219, 0.0232, 0.0226, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 00:38:13,689 INFO [finetune.py:976] (2/7) Epoch 27, batch 1050, loss[loss=0.1782, simple_loss=0.2519, pruned_loss=0.05226, over 4926.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2408, pruned_loss=0.04729, over 946217.27 frames. ], batch size: 33, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:38:15,492 INFO [optim.py:369] (2/7) 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 00:38:48,335 INFO [finetune.py:976] (2/7) Epoch 27, batch 1100, loss[loss=0.1575, simple_loss=0.2341, pruned_loss=0.04047, over 4763.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2413, pruned_loss=0.04693, over 949990.68 frames. ], batch size: 28, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:39:22,077 INFO [finetune.py:976] (2/7) Epoch 27, batch 1150, loss[loss=0.2178, simple_loss=0.2948, pruned_loss=0.07037, over 4835.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2425, pruned_loss=0.04731, over 950381.23 frames. ], batch size: 47, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:39:23,886 INFO [optim.py:369] (2/7) 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 00:39:28,095 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150079.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:39:40,846 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.96 vs. limit=5.0 +2023-04-28 00:39:43,691 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150102.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 00:39:56,125 INFO [finetune.py:976] (2/7) Epoch 27, batch 1200, loss[loss=0.18, simple_loss=0.2549, pruned_loss=0.05252, over 4899.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2417, pruned_loss=0.0474, over 950813.23 frames. ], batch size: 36, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:40:10,318 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150140.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:40:35,481 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-04-28 00:40:45,077 INFO [finetune.py:976] (2/7) Epoch 27, batch 1250, loss[loss=0.1492, simple_loss=0.2281, pruned_loss=0.03517, over 4898.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2396, pruned_loss=0.04693, over 951892.39 frames. ], batch size: 35, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:40:45,198 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150170.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:40:47,935 INFO [optim.py:369] (2/7) 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 00:41:49,587 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 1300, loss[loss=0.1762, simple_loss=0.238, pruned_loss=0.05724, over 4894.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2365, pruned_loss=0.04594, over 953041.22 frames. ], batch size: 43, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:41:51,511 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150221.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:42:01,188 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6307, 1.4030, 0.7473, 1.2915, 1.5991, 1.4835, 1.3968, 1.4444], + device='cuda:2'), covar=tensor([0.0472, 0.0363, 0.0326, 0.0513, 0.0271, 0.0451, 0.0447, 0.0545], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 00:42:25,951 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9784, 1.5593, 1.5325, 1.6826, 2.1876, 1.7952, 1.6308, 1.4765], + device='cuda:2'), covar=tensor([0.1558, 0.1504, 0.1746, 0.1323, 0.0829, 0.1438, 0.1826, 0.2259], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0307, 0.0349, 0.0286, 0.0326, 0.0305, 0.0298, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.4064e-05, 6.3046e-05, 7.3304e-05, 5.7200e-05, 6.6736e-05, 6.3701e-05, + 6.1693e-05, 7.8974e-05], device='cuda:2') +2023-04-28 00:42:37,618 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8037, 1.3541, 1.8705, 2.3335, 1.9130, 1.7594, 1.8183, 1.7319], + device='cuda:2'), covar=tensor([0.4212, 0.6264, 0.5905, 0.4706, 0.5376, 0.7333, 0.7119, 0.8447], + device='cuda:2'), in_proj_covar=tensor([0.0441, 0.0422, 0.0515, 0.0505, 0.0470, 0.0506, 0.0508, 0.0521], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:42:44,771 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-28 00:42:56,733 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150269.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:42:57,287 INFO [finetune.py:976] (2/7) Epoch 27, batch 1350, loss[loss=0.1599, simple_loss=0.2276, pruned_loss=0.04607, over 4906.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2371, pruned_loss=0.04652, over 952464.94 frames. ], batch size: 35, lr: 2.93e-03, grad_scale: 32.0 +2023-04-28 00:42:59,129 INFO [optim.py:369] (2/7) 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 00:43:05,194 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150274.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:44:01,099 INFO [finetune.py:976] (2/7) Epoch 27, batch 1400, loss[loss=0.1601, simple_loss=0.2343, pruned_loss=0.04291, over 4862.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2415, pruned_loss=0.04765, over 954107.72 frames. ], batch size: 44, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:44:22,633 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 1450, loss[loss=0.2101, simple_loss=0.2803, pruned_loss=0.06996, over 4813.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.243, pruned_loss=0.04774, over 954586.36 frames. ], batch size: 38, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:45:03,057 INFO [optim.py:369] (2/7) 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 00:45:22,885 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150402.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 00:45:27,761 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2485, 1.3985, 1.3092, 1.6557, 1.4900, 1.8553, 1.3451, 3.3198], + device='cuda:2'), covar=tensor([0.0599, 0.0813, 0.0775, 0.1180, 0.0654, 0.0490, 0.0734, 0.0134], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 00:45:29,565 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6219, 3.5868, 2.5587, 4.2018, 3.6412, 3.5503, 1.4284, 3.5873], + device='cuda:2'), covar=tensor([0.1679, 0.1313, 0.3329, 0.1769, 0.2681, 0.1987, 0.6277, 0.2630], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0219, 0.0252, 0.0302, 0.0298, 0.0247, 0.0274, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:45:33,782 INFO [finetune.py:976] (2/7) Epoch 27, batch 1500, loss[loss=0.1569, simple_loss=0.2342, pruned_loss=0.03976, over 4837.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2422, pruned_loss=0.04752, over 953072.85 frames. ], batch size: 30, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:45:44,395 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 27, batch 1550, loss[loss=0.1441, simple_loss=0.2273, pruned_loss=0.03045, over 4671.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2419, pruned_loss=0.04716, over 954095.81 frames. ], batch size: 23, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:46:09,360 INFO [optim.py:369] (2/7) 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 00:46:40,767 INFO [finetune.py:976] (2/7) Epoch 27, batch 1600, loss[loss=0.197, simple_loss=0.2632, pruned_loss=0.06543, over 4875.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2392, pruned_loss=0.04644, over 953186.76 frames. ], batch size: 31, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:47:08,147 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150537.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:47:40,813 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8535, 2.2161, 2.2369, 2.3247, 2.2036, 2.2489, 2.3656, 2.2461], + device='cuda:2'), covar=tensor([0.3237, 0.4597, 0.4114, 0.4199, 0.5115, 0.6434, 0.4318, 0.4453], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0373, 0.0329, 0.0339, 0.0349, 0.0392, 0.0358, 0.0331], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:47:41,250 INFO [finetune.py:976] (2/7) Epoch 27, batch 1650, loss[loss=0.1489, simple_loss=0.22, pruned_loss=0.03887, over 4153.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2374, pruned_loss=0.04593, over 955077.14 frames. ], batch size: 18, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:47:43,700 INFO [optim.py:369] (2/7) 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 00:48:00,531 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150598.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:48:14,792 INFO [finetune.py:976] (2/7) Epoch 27, batch 1700, loss[loss=0.1659, simple_loss=0.2364, pruned_loss=0.04774, over 4759.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2363, pruned_loss=0.04595, over 954162.31 frames. ], batch size: 26, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:48:20,946 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150630.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:48:45,819 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6127, 3.4754, 0.8093, 1.9185, 1.8172, 2.4560, 1.9537, 1.0265], + device='cuda:2'), covar=tensor([0.1340, 0.0891, 0.2096, 0.1241, 0.1111, 0.1061, 0.1564, 0.1974], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0236, 0.0134, 0.0120, 0.0130, 0.0151, 0.0117, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 00:48:48,161 INFO [finetune.py:976] (2/7) Epoch 27, batch 1750, loss[loss=0.1794, simple_loss=0.2422, pruned_loss=0.0583, over 4789.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.239, pruned_loss=0.04708, over 953125.76 frames. ], batch size: 29, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:48:50,601 INFO [optim.py:369] (2/7) 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 00:49:19,219 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5769, 0.7815, 1.4970, 1.9075, 1.6904, 1.5606, 1.5380, 1.5285], + device='cuda:2'), covar=tensor([0.3839, 0.5351, 0.5120, 0.5136, 0.5065, 0.6006, 0.6076, 0.6203], + device='cuda:2'), in_proj_covar=tensor([0.0442, 0.0422, 0.0518, 0.0507, 0.0470, 0.0507, 0.0508, 0.0522], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:49:21,910 INFO [finetune.py:976] (2/7) Epoch 27, batch 1800, loss[loss=0.2105, simple_loss=0.2726, pruned_loss=0.07425, over 4819.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2418, pruned_loss=0.04769, over 953350.21 frames. ], batch size: 30, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:49:29,212 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150732.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:49:30,998 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150735.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:49:54,541 INFO [finetune.py:976] (2/7) Epoch 27, batch 1850, loss[loss=0.1975, simple_loss=0.2692, pruned_loss=0.06289, over 4837.00 frames. ], tot_loss[loss=0.172, simple_loss=0.245, pruned_loss=0.04952, over 951621.25 frames. ], batch size: 47, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:50:02,278 INFO [optim.py:369] (2/7) 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 00:50:13,366 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150783.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:50:25,719 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150793.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 00:50:56,511 INFO [finetune.py:976] (2/7) Epoch 27, batch 1900, loss[loss=0.1796, simple_loss=0.2534, pruned_loss=0.05291, over 4841.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2444, pruned_loss=0.04898, over 951157.71 frames. ], batch size: 47, lr: 2.93e-03, grad_scale: 16.0 +2023-04-28 00:51:09,967 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9195, 1.4652, 1.4713, 1.5835, 2.0381, 1.6495, 1.4718, 1.4253], + device='cuda:2'), covar=tensor([0.1624, 0.1548, 0.1640, 0.1379, 0.0862, 0.1812, 0.1963, 0.2392], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0305, 0.0349, 0.0284, 0.0325, 0.0304, 0.0297, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.3645e-05, 6.2752e-05, 7.3190e-05, 5.6948e-05, 6.6484e-05, 6.3340e-05, + 6.1577e-05, 7.8812e-05], device='cuda:2') +2023-04-28 00:51:18,118 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8333, 1.6418, 4.4459, 4.2083, 3.9114, 4.1990, 4.2341, 3.9447], + device='cuda:2'), covar=tensor([0.7032, 0.5787, 0.1149, 0.1783, 0.1084, 0.2141, 0.1060, 0.1566], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0310, 0.0412, 0.0411, 0.0349, 0.0417, 0.0321, 0.0365], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:51:29,792 INFO [finetune.py:976] (2/7) Epoch 27, batch 1950, loss[loss=0.1384, simple_loss=0.21, pruned_loss=0.03341, over 4812.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.243, pruned_loss=0.048, over 951214.33 frames. ], batch size: 25, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:51:32,193 INFO [optim.py:369] (2/7) 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 00:51:44,421 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150893.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:51:46,881 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8380, 2.9069, 2.2717, 3.3006, 2.9017, 2.8647, 1.2296, 2.7634], + device='cuda:2'), covar=tensor([0.2172, 0.1574, 0.3078, 0.2681, 0.3617, 0.2213, 0.5716, 0.2891], + device='cuda:2'), in_proj_covar=tensor([0.0242, 0.0217, 0.0249, 0.0299, 0.0295, 0.0245, 0.0271, 0.0270], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 00:51:49,215 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150900.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:51:54,960 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-28 00:52:03,055 INFO [finetune.py:976] (2/7) Epoch 27, batch 2000, loss[loss=0.1371, simple_loss=0.2144, pruned_loss=0.02993, over 4900.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2398, pruned_loss=0.04689, over 952758.81 frames. ], batch size: 32, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:52:09,165 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150930.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:52:51,907 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150961.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:53:02,955 INFO [finetune.py:976] (2/7) Epoch 27, batch 2050, loss[loss=0.1723, simple_loss=0.2381, pruned_loss=0.05325, over 4937.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2368, pruned_loss=0.04576, over 951425.70 frames. ], batch size: 33, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:53:05,396 INFO [optim.py:369] (2/7) 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 00:53:12,616 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150978.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:53:27,353 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.71 vs. limit=5.0 +2023-04-28 00:54:07,738 INFO [finetune.py:976] (2/7) Epoch 27, batch 2100, loss[loss=0.156, simple_loss=0.2334, pruned_loss=0.03932, over 4836.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2366, pruned_loss=0.04586, over 952181.10 frames. ], batch size: 33, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:55:11,731 INFO [finetune.py:976] (2/7) Epoch 27, batch 2150, loss[loss=0.2208, simple_loss=0.2954, pruned_loss=0.07307, over 4920.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2396, pruned_loss=0.04666, over 955255.48 frames. ], batch size: 42, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:55:19,340 INFO [optim.py:369] (2/7) 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 00:55:21,952 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6953, 1.3368, 1.2999, 1.5203, 1.8889, 1.5437, 1.3995, 1.2505], + device='cuda:2'), covar=tensor([0.1726, 0.1528, 0.1528, 0.1424, 0.0899, 0.1667, 0.2016, 0.2133], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0308, 0.0350, 0.0286, 0.0327, 0.0306, 0.0299, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.4140e-05, 6.3184e-05, 7.3519e-05, 5.7263e-05, 6.6923e-05, 6.3785e-05, + 6.1971e-05, 7.9292e-05], device='cuda:2') +2023-04-28 00:55:33,247 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151088.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 00:56:16,162 INFO [finetune.py:976] (2/7) Epoch 27, batch 2200, loss[loss=0.2007, simple_loss=0.2744, pruned_loss=0.06347, over 4857.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2411, pruned_loss=0.04745, over 955277.29 frames. ], batch size: 44, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:57:19,653 INFO [finetune.py:976] (2/7) Epoch 27, batch 2250, loss[loss=0.1496, simple_loss=0.2291, pruned_loss=0.03503, over 4818.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2428, pruned_loss=0.04805, over 954707.15 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:57:20,381 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2848, 2.2950, 2.0912, 2.0489, 2.4487, 2.0282, 3.0705, 1.8255], + device='cuda:2'), covar=tensor([0.3523, 0.2113, 0.4045, 0.2970, 0.1636, 0.2400, 0.1256, 0.4335], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0354, 0.0424, 0.0350, 0.0380, 0.0375, 0.0366, 0.0420], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 00:57:27,225 INFO [optim.py:369] (2/7) 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 00:57:38,577 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151184.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:57:49,167 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151193.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:58:22,083 INFO [finetune.py:976] (2/7) Epoch 27, batch 2300, loss[loss=0.1806, simple_loss=0.244, pruned_loss=0.05857, over 4849.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2427, pruned_loss=0.04776, over 953257.90 frames. ], batch size: 44, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:58:42,722 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151241.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:58:45,189 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151245.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:58:52,343 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151256.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:59:01,205 INFO [finetune.py:976] (2/7) Epoch 27, batch 2350, loss[loss=0.1528, simple_loss=0.2296, pruned_loss=0.03796, over 4825.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2413, pruned_loss=0.04771, over 953912.66 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:59:03,628 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1168, 2.1893, 2.0612, 1.8658, 2.1948, 1.9104, 2.8313, 1.6217], + device='cuda:2'), covar=tensor([0.3336, 0.1876, 0.4212, 0.2735, 0.1573, 0.2259, 0.1171, 0.4484], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0354, 0.0424, 0.0351, 0.0380, 0.0375, 0.0366, 0.0419], + device='cuda:2'), out_proj_covar=tensor([9.9938e-05, 1.0536e-04, 1.2831e-04, 1.0499e-04, 1.1263e-04, 1.1128e-04, + 1.0695e-04, 1.2611e-04], device='cuda:2') +2023-04-28 00:59:04,082 INFO [optim.py:369] (2/7) 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 00:59:10,173 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151282.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 00:59:20,476 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 2400, loss[loss=0.1202, simple_loss=0.1952, pruned_loss=0.02264, over 4782.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2385, pruned_loss=0.04668, over 953318.40 frames. ], batch size: 29, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 00:59:50,883 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151343.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:00:01,129 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151359.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:00:07,712 INFO [finetune.py:976] (2/7) Epoch 27, batch 2450, loss[loss=0.1549, simple_loss=0.2414, pruned_loss=0.0342, over 4909.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2359, pruned_loss=0.04556, over 955115.12 frames. ], batch size: 37, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:00:10,100 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7718, 2.0404, 1.1377, 1.4494, 2.2169, 1.6618, 1.5642, 1.6707], + device='cuda:2'), covar=tensor([0.0466, 0.0333, 0.0277, 0.0523, 0.0234, 0.0489, 0.0458, 0.0549], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0050, 0.0052], + device='cuda:2') +2023-04-28 01:00:10,590 INFO [optim.py:369] (2/7) 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 01:00:21,183 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151388.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:00:31,719 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-28 01:00:41,810 INFO [finetune.py:976] (2/7) Epoch 27, batch 2500, loss[loss=0.1509, simple_loss=0.2285, pruned_loss=0.03668, over 4829.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2368, pruned_loss=0.04596, over 955606.44 frames. ], batch size: 39, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:00:45,411 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8790, 1.6860, 1.6217, 1.3426, 1.8010, 1.5454, 2.2053, 1.4472], + device='cuda:2'), covar=tensor([0.3335, 0.1908, 0.4700, 0.2787, 0.1618, 0.2103, 0.1511, 0.4589], + device='cuda:2'), in_proj_covar=tensor([0.0336, 0.0352, 0.0421, 0.0348, 0.0377, 0.0372, 0.0364, 0.0417], + device='cuda:2'), 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:2') +2023-04-28 01:00:53,953 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 2550, loss[loss=0.1787, simple_loss=0.2592, pruned_loss=0.04908, over 4810.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2392, pruned_loss=0.04601, over 956028.29 frames. ], batch size: 40, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:01:17,942 INFO [optim.py:369] (2/7) 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 01:01:37,791 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-04-28 01:01:48,925 INFO [finetune.py:976] (2/7) Epoch 27, batch 2600, loss[loss=0.148, simple_loss=0.2341, pruned_loss=0.03092, over 4887.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2412, pruned_loss=0.04681, over 955295.66 frames. ], batch size: 35, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:02:02,558 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151540.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:02:19,376 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151556.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:02:33,623 INFO [finetune.py:976] (2/7) Epoch 27, batch 2650, loss[loss=0.1963, simple_loss=0.2767, pruned_loss=0.05792, over 4821.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2418, pruned_loss=0.04667, over 955215.68 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:02:41,497 INFO [optim.py:369] (2/7) 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 01:03:23,448 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151604.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:03:39,056 INFO [finetune.py:976] (2/7) Epoch 27, batch 2700, loss[loss=0.1611, simple_loss=0.2353, pruned_loss=0.04339, over 4843.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2409, pruned_loss=0.04625, over 954145.18 frames. ], batch size: 44, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:04:05,032 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151638.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:04:05,661 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.2893, 4.0887, 3.0068, 4.8900, 4.2268, 4.2314, 1.7784, 4.1345], + device='cuda:2'), covar=tensor([0.1404, 0.1171, 0.3430, 0.0962, 0.2851, 0.1596, 0.5406, 0.2083], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0219, 0.0251, 0.0302, 0.0297, 0.0247, 0.0274, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:04:08,046 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6185, 3.3371, 0.9003, 1.9423, 1.8460, 2.3892, 1.8908, 1.0158], + device='cuda:2'), covar=tensor([0.1302, 0.1045, 0.2020, 0.1219, 0.1100, 0.1020, 0.1612, 0.2022], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0237, 0.0134, 0.0120, 0.0130, 0.0151, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:04:21,254 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151654.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:04:31,300 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-04-28 01:04:40,569 INFO [finetune.py:976] (2/7) Epoch 27, batch 2750, loss[loss=0.1663, simple_loss=0.2427, pruned_loss=0.04495, over 4860.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2379, pruned_loss=0.04541, over 955584.64 frames. ], batch size: 49, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:04:48,521 INFO [optim.py:369] (2/7) 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 01:05:07,212 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151688.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:05:44,740 INFO [finetune.py:976] (2/7) Epoch 27, batch 2800, loss[loss=0.1432, simple_loss=0.2185, pruned_loss=0.03393, over 4763.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2351, pruned_loss=0.0447, over 954481.07 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:06:24,882 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151749.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:06:53,983 INFO [finetune.py:976] (2/7) Epoch 27, batch 2850, loss[loss=0.1391, simple_loss=0.2156, pruned_loss=0.03125, over 4865.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.235, pruned_loss=0.04491, over 955141.74 frames. ], batch size: 31, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:06:56,484 INFO [optim.py:369] (2/7) 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 01:07:05,531 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6730, 3.3314, 2.6539, 2.8867, 2.0467, 2.0850, 2.8456, 2.1401], + device='cuda:2'), covar=tensor([0.1484, 0.1187, 0.1173, 0.1252, 0.2011, 0.1672, 0.0800, 0.1652], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0208, 0.0169, 0.0204, 0.0200, 0.0186, 0.0155, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 01:07:07,842 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.6655, 3.6516, 2.5821, 4.1725, 3.6520, 3.6509, 1.3616, 3.5496], + device='cuda:2'), covar=tensor([0.1793, 0.1341, 0.3312, 0.1917, 0.3799, 0.1852, 0.6516, 0.2525], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0218, 0.0251, 0.0302, 0.0298, 0.0247, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:07:29,458 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151801.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:07:58,282 INFO [finetune.py:976] (2/7) Epoch 27, batch 2900, loss[loss=0.1839, simple_loss=0.2672, pruned_loss=0.05026, over 4744.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2379, pruned_loss=0.04583, over 954918.29 frames. ], batch size: 54, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:08:11,749 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151840.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:08:32,213 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 2950, loss[loss=0.177, simple_loss=0.2586, pruned_loss=0.04774, over 4850.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2408, pruned_loss=0.04677, over 955300.12 frames. ], batch size: 44, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:08:50,555 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151888.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:09:34,745 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5806, 1.5814, 0.8503, 1.2839, 1.7788, 1.4663, 1.3787, 1.4665], + device='cuda:2'), covar=tensor([0.0479, 0.0363, 0.0318, 0.0523, 0.0263, 0.0481, 0.0455, 0.0544], + device='cuda:2'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0038, 0.0053, 0.0039, 0.0051, 0.0050, 0.0052], + device='cuda:2') +2023-04-28 01:09:44,036 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-28 01:09:46,546 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-04-28 01:09:47,998 INFO [finetune.py:976] (2/7) Epoch 27, batch 3000, loss[loss=0.1695, simple_loss=0.2503, pruned_loss=0.04433, over 4811.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2428, pruned_loss=0.04759, over 955808.51 frames. ], batch size: 25, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:09:47,998 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 01:10:08,676 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6400MB +2023-04-28 01:10:09,429 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4407, 1.2870, 1.6111, 1.6308, 1.3505, 1.2826, 1.3603, 0.8064], + device='cuda:2'), covar=tensor([0.0440, 0.0587, 0.0401, 0.0485, 0.0637, 0.1067, 0.0494, 0.0583], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0065, 0.0068, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:10:24,777 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151938.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:10:27,989 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-04-28 01:10:35,071 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151954.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:10:45,718 INFO [finetune.py:976] (2/7) Epoch 27, batch 3050, loss[loss=0.1736, simple_loss=0.2526, pruned_loss=0.04725, over 4879.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2421, pruned_loss=0.04695, over 955098.24 frames. ], batch size: 34, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:10:48,112 INFO [optim.py:369] (2/7) 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 01:10:52,359 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-04-28 01:10:57,068 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151986.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:10:59,052 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-04-28 01:11:08,926 INFO [zipformer.py:1188] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.7702, 3.9407, 0.7833, 2.1083, 2.0591, 2.5637, 2.3171, 0.9679], + device='cuda:2'), covar=tensor([0.1285, 0.0819, 0.2003, 0.1181, 0.1073, 0.1122, 0.1348, 0.2216], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0135, 0.0121, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:11:20,459 INFO [finetune.py:976] (2/7) Epoch 27, batch 3100, loss[loss=0.1676, simple_loss=0.2531, pruned_loss=0.04106, over 4814.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2405, pruned_loss=0.04638, over 955270.09 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:11:37,591 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152044.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:11:44,950 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8306, 1.2799, 1.4939, 1.4464, 1.9394, 1.5903, 1.3389, 1.4377], + device='cuda:2'), covar=tensor([0.1677, 0.1391, 0.2068, 0.1493, 0.0936, 0.1469, 0.1750, 0.2366], + device='cuda:2'), in_proj_covar=tensor([0.0319, 0.0311, 0.0356, 0.0290, 0.0331, 0.0310, 0.0303, 0.0380], + device='cuda:2'), out_proj_covar=tensor([6.5050e-05, 6.3873e-05, 7.4743e-05, 5.8074e-05, 6.7726e-05, 6.4659e-05, + 6.2780e-05, 8.0311e-05], device='cuda:2') +2023-04-28 01:11:54,337 INFO [finetune.py:976] (2/7) Epoch 27, batch 3150, loss[loss=0.1548, simple_loss=0.2201, pruned_loss=0.04474, over 4852.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.238, pruned_loss=0.04611, over 955070.75 frames. ], batch size: 49, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:11:56,743 INFO [optim.py:369] (2/7) 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 01:11:59,445 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-28 01:12:15,295 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-04-28 01:12:27,068 INFO [finetune.py:976] (2/7) Epoch 27, batch 3200, loss[loss=0.1644, simple_loss=0.2382, pruned_loss=0.04528, over 4872.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2346, pruned_loss=0.04477, over 956980.17 frames. ], batch size: 31, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:12:30,695 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2205, 2.6870, 1.0732, 1.5311, 2.1683, 1.2465, 3.7108, 1.9470], + device='cuda:2'), covar=tensor([0.0643, 0.0607, 0.0730, 0.1180, 0.0453, 0.0951, 0.0195, 0.0536], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 01:12:52,024 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152157.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:13:00,433 INFO [finetune.py:976] (2/7) Epoch 27, batch 3250, loss[loss=0.1785, simple_loss=0.2468, pruned_loss=0.05514, over 4097.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2347, pruned_loss=0.04476, over 955801.85 frames. ], batch size: 65, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:13:02,811 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 27, batch 3300, loss[loss=0.1385, simple_loss=0.2115, pruned_loss=0.03271, over 4781.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2374, pruned_loss=0.04537, over 956000.01 frames. ], batch size: 26, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:13:36,732 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152225.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:14:13,058 INFO [finetune.py:976] (2/7) Epoch 27, batch 3350, loss[loss=0.1829, simple_loss=0.2582, pruned_loss=0.05376, over 4916.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2403, pruned_loss=0.0466, over 955105.16 frames. ], batch size: 42, lr: 2.92e-03, grad_scale: 16.0 +2023-04-28 01:14:14,903 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152273.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:14:15,408 INFO [optim.py:369] (2/7) 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 01:14:34,638 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152286.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:14:45,504 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-04-28 01:14:57,212 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-04-28 01:15:17,722 INFO [finetune.py:976] (2/7) Epoch 27, batch 3400, loss[loss=0.1629, simple_loss=0.2325, pruned_loss=0.04662, over 4835.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.242, pruned_loss=0.04765, over 956857.80 frames. ], batch size: 30, lr: 2.92e-03, grad_scale: 32.0 +2023-04-28 01:15:36,891 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152334.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:15:50,208 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152344.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:16:01,851 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9724, 1.8394, 2.4062, 2.5374, 1.7091, 1.5275, 1.9736, 1.1552], + device='cuda:2'), covar=tensor([0.0557, 0.0680, 0.0359, 0.0593, 0.0646, 0.1138, 0.0592, 0.0659], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0069, 0.0075, 0.0095, 0.0073, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:16:22,884 INFO [finetune.py:976] (2/7) Epoch 27, batch 3450, loss[loss=0.1417, simple_loss=0.2195, pruned_loss=0.03197, over 4897.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2419, pruned_loss=0.04714, over 955803.72 frames. ], batch size: 32, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:16:31,451 INFO [optim.py:369] (2/7) 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 01:16:53,821 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152392.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:17:28,482 INFO [finetune.py:976] (2/7) Epoch 27, batch 3500, loss[loss=0.1514, simple_loss=0.2258, pruned_loss=0.03855, over 4735.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2407, pruned_loss=0.04707, over 955719.31 frames. ], batch size: 54, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:18:00,033 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152445.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:18:18,960 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152457.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:18:20,862 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6610, 1.0681, 1.5441, 1.6431, 1.5880, 1.6293, 1.5492, 1.5569], + device='cuda:2'), covar=tensor([0.3113, 0.3989, 0.3468, 0.3400, 0.4476, 0.5768, 0.3536, 0.3481], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0376, 0.0332, 0.0341, 0.0350, 0.0394, 0.0362, 0.0335], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:18:29,524 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-28 01:18:32,262 INFO [finetune.py:976] (2/7) Epoch 27, batch 3550, loss[loss=0.1463, simple_loss=0.2121, pruned_loss=0.04024, over 4903.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2376, pruned_loss=0.04595, over 956613.15 frames. ], batch size: 43, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:18:32,374 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1101, 2.5115, 2.2842, 2.5235, 1.8111, 2.1525, 2.1886, 1.6518], + device='cuda:2'), covar=tensor([0.1966, 0.1071, 0.0666, 0.0868, 0.3087, 0.0990, 0.1783, 0.2522], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0301, 0.0215, 0.0276, 0.0315, 0.0254, 0.0247, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1272e-04, 1.1832e-04, 8.4625e-05, 1.0857e-04, 1.2697e-04, 9.9682e-05, + 9.9605e-05, 1.0403e-04], device='cuda:2') +2023-04-28 01:18:34,714 INFO [optim.py:369] (2/7) 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 01:18:42,726 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6824, 1.2695, 4.2559, 4.0179, 3.6905, 3.9690, 3.9313, 3.7610], + device='cuda:2'), covar=tensor([0.7264, 0.5838, 0.1098, 0.1548, 0.1098, 0.1712, 0.1988, 0.1581], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0310, 0.0412, 0.0410, 0.0351, 0.0418, 0.0321, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:18:52,636 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-04-28 01:19:25,033 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152505.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:19:25,738 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152506.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:19:39,436 INFO [finetune.py:976] (2/7) Epoch 27, batch 3600, loss[loss=0.1565, simple_loss=0.2413, pruned_loss=0.03581, over 4902.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.236, pruned_loss=0.04589, over 956446.08 frames. ], batch size: 36, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:20:44,730 INFO [finetune.py:976] (2/7) Epoch 27, batch 3650, loss[loss=0.1352, simple_loss=0.2033, pruned_loss=0.03351, over 3874.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.238, pruned_loss=0.04641, over 955282.50 frames. ], batch size: 17, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:20:51,634 INFO [optim.py:369] (2/7) 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 01:20:56,029 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152581.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:21:02,407 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152582.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:21:12,803 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8468, 2.1649, 2.0001, 2.1965, 1.5133, 1.8770, 1.8201, 1.4822], + device='cuda:2'), covar=tensor([0.1793, 0.1342, 0.0767, 0.0935, 0.3552, 0.1128, 0.1981, 0.2473], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0300, 0.0215, 0.0275, 0.0314, 0.0253, 0.0246, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1253e-04, 1.1811e-04, 8.4584e-05, 1.0838e-04, 1.2645e-04, 9.9637e-05, + 9.9260e-05, 1.0373e-04], device='cuda:2') +2023-04-28 01:21:12,810 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152589.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:21:47,843 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3368, 2.1467, 2.6016, 2.9508, 2.8577, 2.3603, 2.0190, 2.6807], + device='cuda:2'), covar=tensor([0.0833, 0.1019, 0.0566, 0.0528, 0.0546, 0.0798, 0.0698, 0.0490], + device='cuda:2'), in_proj_covar=tensor([0.0182, 0.0198, 0.0181, 0.0168, 0.0175, 0.0175, 0.0149, 0.0175], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:21:49,538 INFO [finetune.py:976] (2/7) Epoch 27, batch 3700, loss[loss=0.1481, simple_loss=0.2386, pruned_loss=0.0288, over 4908.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2401, pruned_loss=0.04661, over 952556.86 frames. ], batch size: 37, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:22:00,775 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152629.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:22:21,246 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152643.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:22:31,110 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152650.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:22:52,120 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6194, 1.9669, 1.8197, 1.9607, 1.5328, 1.6557, 1.6200, 1.3203], + device='cuda:2'), covar=tensor([0.1746, 0.1310, 0.0747, 0.1043, 0.3182, 0.1245, 0.1738, 0.2347], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0301, 0.0215, 0.0276, 0.0314, 0.0253, 0.0246, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1266e-04, 1.1830e-04, 8.4753e-05, 1.0846e-04, 1.2658e-04, 9.9584e-05, + 9.9302e-05, 1.0388e-04], device='cuda:2') +2023-04-28 01:22:56,764 INFO [finetune.py:976] (2/7) Epoch 27, batch 3750, loss[loss=0.1862, simple_loss=0.2585, pruned_loss=0.05694, over 4890.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2426, pruned_loss=0.04745, over 953086.31 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:23:03,984 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7369, 1.1969, 1.7948, 2.2062, 1.7863, 1.6986, 1.7563, 1.7069], + device='cuda:2'), covar=tensor([0.4682, 0.6844, 0.6089, 0.5557, 0.5788, 0.7811, 0.8083, 0.9186], + device='cuda:2'), in_proj_covar=tensor([0.0442, 0.0422, 0.0518, 0.0506, 0.0469, 0.0506, 0.0505, 0.0521], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:23:04,995 INFO [optim.py:369] (2/7) 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 01:23:56,894 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8995, 2.2611, 2.0530, 2.2426, 1.6882, 1.8668, 1.8188, 1.4291], + device='cuda:2'), covar=tensor([0.1654, 0.1228, 0.0757, 0.0928, 0.2975, 0.1023, 0.1809, 0.2516], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0300, 0.0215, 0.0275, 0.0314, 0.0253, 0.0245, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1235e-04, 1.1802e-04, 8.4640e-05, 1.0836e-04, 1.2633e-04, 9.9455e-05, + 9.9022e-05, 1.0384e-04], device='cuda:2') +2023-04-28 01:24:08,009 INFO [finetune.py:976] (2/7) Epoch 27, batch 3800, loss[loss=0.1603, simple_loss=0.231, pruned_loss=0.04481, over 4749.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.244, pruned_loss=0.04784, over 953148.95 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:24:11,808 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152726.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:25:13,179 INFO [finetune.py:976] (2/7) Epoch 27, batch 3850, loss[loss=0.1901, simple_loss=0.2555, pruned_loss=0.06237, over 4202.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2428, pruned_loss=0.04712, over 953582.87 frames. ], batch size: 65, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:25:15,589 INFO [optim.py:369] (2/7) 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 01:25:33,572 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152787.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:25:34,925 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-28 01:25:46,556 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9791, 4.0379, 1.0035, 2.2417, 2.2154, 2.7077, 2.3900, 1.0059], + device='cuda:2'), covar=tensor([0.1254, 0.0829, 0.1886, 0.1171, 0.1007, 0.1038, 0.1409, 0.2113], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0134, 0.0120, 0.0131, 0.0152, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:25:47,900 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152801.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:26:07,728 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2934, 1.5964, 1.4666, 1.7976, 1.7641, 1.9908, 1.4512, 3.7655], + device='cuda:2'), covar=tensor([0.0588, 0.0795, 0.0776, 0.1232, 0.0621, 0.0475, 0.0750, 0.0101], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 01:26:17,504 INFO [finetune.py:976] (2/7) Epoch 27, batch 3900, loss[loss=0.1206, simple_loss=0.2057, pruned_loss=0.01775, over 4872.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2394, pruned_loss=0.04593, over 954467.37 frames. ], batch size: 31, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:26:28,055 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4663, 1.4045, 1.8268, 1.8752, 1.3330, 1.2688, 1.5371, 0.9728], + device='cuda:2'), covar=tensor([0.0554, 0.0637, 0.0332, 0.0469, 0.0731, 0.0967, 0.0581, 0.0574], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0066, 0.0069, 0.0075, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:26:28,593 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3446, 3.2904, 2.4935, 3.8755, 3.3729, 3.3952, 1.5769, 3.2796], + device='cuda:2'), covar=tensor([0.1767, 0.1373, 0.3419, 0.2223, 0.2949, 0.1858, 0.5651, 0.2663], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0219, 0.0252, 0.0303, 0.0300, 0.0249, 0.0274, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:26:43,348 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4536, 1.3363, 4.0804, 3.7907, 3.5688, 3.8618, 3.7882, 3.5662], + device='cuda:2'), covar=tensor([0.6953, 0.6081, 0.1131, 0.1806, 0.1183, 0.1898, 0.1772, 0.1656], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0310, 0.0411, 0.0409, 0.0351, 0.0419, 0.0321, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:27:03,885 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8404, 3.7003, 0.8119, 2.0345, 1.9642, 2.5209, 2.0795, 0.9941], + device='cuda:2'), covar=tensor([0.1251, 0.0867, 0.2013, 0.1198, 0.1054, 0.1050, 0.1461, 0.1902], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0134, 0.0120, 0.0131, 0.0152, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:27:22,909 INFO [finetune.py:976] (2/7) Epoch 27, batch 3950, loss[loss=0.1703, simple_loss=0.2387, pruned_loss=0.05092, over 4905.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2364, pruned_loss=0.04524, over 956308.04 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:27:26,330 INFO [optim.py:369] (2/7) 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 01:27:36,690 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152881.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:28:19,599 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6703, 0.6937, 1.5842, 1.9989, 1.7415, 1.5681, 1.5812, 1.5771], + device='cuda:2'), covar=tensor([0.4017, 0.6378, 0.5494, 0.5177, 0.5423, 0.7212, 0.6993, 0.7843], + device='cuda:2'), in_proj_covar=tensor([0.0442, 0.0423, 0.0519, 0.0508, 0.0470, 0.0507, 0.0507, 0.0523], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:28:27,783 INFO [finetune.py:976] (2/7) Epoch 27, batch 4000, loss[loss=0.139, simple_loss=0.2123, pruned_loss=0.03284, over 4764.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2361, pruned_loss=0.04535, over 958572.59 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:28:39,700 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:28:39,728 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:28:50,589 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152938.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:28:50,815 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-04-28 01:29:00,136 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152945.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:29:32,518 INFO [finetune.py:976] (2/7) Epoch 27, batch 4050, loss[loss=0.1786, simple_loss=0.2602, pruned_loss=0.04853, over 4830.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2403, pruned_loss=0.0471, over 957889.48 frames. ], batch size: 51, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:29:35,452 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152977.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:29:55,054 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2771, 1.4390, 1.2319, 1.6203, 1.6001, 1.9423, 1.3247, 3.4651], + device='cuda:2'), covar=tensor([0.0608, 0.0874, 0.0863, 0.1325, 0.0670, 0.0546, 0.0828, 0.0138], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 01:29:55,860 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 01:30:36,405 INFO [finetune.py:976] (2/7) Epoch 27, batch 4100, loss[loss=0.1657, simple_loss=0.2375, pruned_loss=0.04692, over 4857.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2421, pruned_loss=0.04746, over 956636.14 frames. ], batch size: 44, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:30:57,450 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2872, 1.5841, 1.3246, 1.7970, 1.7250, 2.0308, 1.4076, 3.7257], + device='cuda:2'), covar=tensor([0.0566, 0.0779, 0.0834, 0.1233, 0.0638, 0.0486, 0.0770, 0.0137], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 01:31:12,227 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-04-28 01:31:40,930 INFO [finetune.py:976] (2/7) Epoch 27, batch 4150, loss[loss=0.2007, simple_loss=0.271, pruned_loss=0.06516, over 4820.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2435, pruned_loss=0.04766, over 958176.39 frames. ], batch size: 33, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:31:43,380 INFO [optim.py:369] (2/7) 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 01:31:54,614 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153082.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:32:22,861 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153101.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:32:43,322 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 4200, loss[loss=0.1444, simple_loss=0.2213, pruned_loss=0.03372, over 4762.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.244, pruned_loss=0.04734, over 957535.35 frames. ], batch size: 28, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:33:26,254 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 4250, loss[loss=0.1596, simple_loss=0.2257, pruned_loss=0.04671, over 4787.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2407, pruned_loss=0.04611, over 955988.72 frames. ], batch size: 29, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:33:57,257 INFO [optim.py:369] (2/7) 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 01:33:59,835 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 4300, loss[loss=0.1638, simple_loss=0.24, pruned_loss=0.04384, over 4903.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2396, pruned_loss=0.04641, over 957211.69 frames. ], batch size: 43, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:35:03,524 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8317, 1.3312, 1.9007, 2.2874, 1.9230, 1.7764, 1.8224, 1.7702], + device='cuda:2'), covar=tensor([0.4249, 0.6796, 0.5812, 0.5299, 0.5450, 0.7372, 0.7200, 0.9566], + device='cuda:2'), in_proj_covar=tensor([0.0442, 0.0422, 0.0517, 0.0507, 0.0469, 0.0506, 0.0507, 0.0523], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:35:17,369 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153238.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:35:23,024 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3536, 1.4721, 1.7852, 1.8971, 1.7786, 1.8719, 1.8001, 1.8613], + device='cuda:2'), covar=tensor([0.3290, 0.5107, 0.4544, 0.4519, 0.5252, 0.6776, 0.4759, 0.4600], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0373, 0.0329, 0.0340, 0.0349, 0.0393, 0.0360, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:35:23,568 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 4350, loss[loss=0.16, simple_loss=0.2317, pruned_loss=0.04418, over 4899.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2359, pruned_loss=0.04509, over 956330.44 frames. ], batch size: 36, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:35:38,835 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4770, 1.3923, 1.6635, 1.7476, 1.3279, 1.2528, 1.4642, 0.9731], + device='cuda:2'), covar=tensor([0.0515, 0.0611, 0.0375, 0.0491, 0.0757, 0.1006, 0.0521, 0.0515], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0074, 0.0094, 0.0072, 0.0063], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:35:41,097 INFO [optim.py:369] (2/7) 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 01:35:44,148 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153278.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:35:48,969 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153286.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:35:53,609 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.9139, 4.8230, 3.2958, 5.6616, 4.9794, 4.8869, 2.3156, 4.8225], + device='cuda:2'), covar=tensor([0.1621, 0.1026, 0.2738, 0.0894, 0.2636, 0.1539, 0.5514, 0.2031], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0218, 0.0251, 0.0303, 0.0299, 0.0247, 0.0274, 0.0273], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:35:54,195 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153293.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:36:12,554 INFO [finetune.py:976] (2/7) Epoch 27, batch 4400, loss[loss=0.1775, simple_loss=0.2486, pruned_loss=0.05322, over 4845.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.236, pruned_loss=0.04514, over 954902.71 frames. ], batch size: 44, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:36:15,819 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-28 01:36:33,326 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 4450, loss[loss=0.1652, simple_loss=0.2316, pruned_loss=0.04938, over 4897.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2388, pruned_loss=0.04588, over 954647.54 frames. ], batch size: 32, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:37:18,372 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2679, 4.3819, 0.8519, 2.2525, 2.4272, 2.7081, 2.4757, 1.0028], + device='cuda:2'), covar=tensor([0.1271, 0.1228, 0.2222, 0.1306, 0.1118, 0.1242, 0.1537, 0.2093], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0136, 0.0121, 0.0132, 0.0153, 0.0118, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:37:18,875 INFO [optim.py:369] (2/7) 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 01:37:30,040 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153382.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:38:10,737 INFO [finetune.py:976] (2/7) Epoch 27, batch 4500, loss[loss=0.2053, simple_loss=0.2705, pruned_loss=0.06999, over 4921.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2411, pruned_loss=0.04685, over 953833.27 frames. ], batch size: 42, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:38:11,463 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153421.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:38:16,906 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153430.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:38:43,733 INFO [finetune.py:976] (2/7) Epoch 27, batch 4550, loss[loss=0.1538, simple_loss=0.2188, pruned_loss=0.04438, over 4373.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2427, pruned_loss=0.04768, over 953936.55 frames. ], batch size: 66, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:38:45,570 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 01:38:51,072 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153482.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:39:15,745 INFO [finetune.py:976] (2/7) Epoch 27, batch 4600, loss[loss=0.1749, simple_loss=0.251, pruned_loss=0.04938, over 4849.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2423, pruned_loss=0.04742, over 953945.73 frames. ], batch size: 44, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:39:17,141 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7291, 0.9385, 1.6484, 2.1609, 1.8233, 1.6239, 1.6539, 1.6318], + device='cuda:2'), covar=tensor([0.4361, 0.7269, 0.6152, 0.5380, 0.5758, 0.7529, 0.7554, 0.9382], + device='cuda:2'), in_proj_covar=tensor([0.0444, 0.0424, 0.0520, 0.0509, 0.0471, 0.0509, 0.0510, 0.0525], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:39:48,979 INFO [finetune.py:976] (2/7) Epoch 27, batch 4650, loss[loss=0.1456, simple_loss=0.2273, pruned_loss=0.03191, over 4819.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.239, pruned_loss=0.04613, over 955739.49 frames. ], batch size: 41, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:39:51,379 INFO [optim.py:369] (2/7) 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 01:39:54,510 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6296, 2.0013, 1.8041, 1.9684, 1.5134, 1.7192, 1.7648, 1.3749], + device='cuda:2'), covar=tensor([0.1708, 0.1262, 0.0802, 0.1095, 0.3349, 0.1084, 0.1636, 0.2306], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0300, 0.0216, 0.0276, 0.0315, 0.0254, 0.0248, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1315e-04, 1.1792e-04, 8.4781e-05, 1.0867e-04, 1.2705e-04, 9.9619e-05, + 9.9874e-05, 1.0408e-04], device='cuda:2') +2023-04-28 01:39:59,505 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-04-28 01:40:09,753 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-28 01:40:27,427 INFO [finetune.py:976] (2/7) Epoch 27, batch 4700, loss[loss=0.1478, simple_loss=0.2146, pruned_loss=0.0405, over 4893.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2372, pruned_loss=0.04589, over 955414.59 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:40:28,933 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-28 01:40:41,152 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153634.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:41:06,298 INFO [finetune.py:976] (2/7) Epoch 27, batch 4750, loss[loss=0.1943, simple_loss=0.2515, pruned_loss=0.06851, over 4090.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2352, pruned_loss=0.04563, over 954702.40 frames. ], batch size: 65, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:41:08,703 INFO [optim.py:369] (2/7) 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 01:41:39,702 INFO [finetune.py:976] (2/7) Epoch 27, batch 4800, loss[loss=0.2011, simple_loss=0.2796, pruned_loss=0.0613, over 4795.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2357, pruned_loss=0.04506, over 952903.39 frames. ], batch size: 45, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:41:47,005 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9259, 2.3865, 1.8730, 1.8730, 1.4132, 1.4469, 1.9708, 1.3057], + device='cuda:2'), covar=tensor([0.1576, 0.1284, 0.1417, 0.1510, 0.2192, 0.1821, 0.0934, 0.1936], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0209, 0.0169, 0.0204, 0.0200, 0.0186, 0.0155, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 01:42:13,210 INFO [finetune.py:976] (2/7) Epoch 27, batch 4850, loss[loss=0.2305, simple_loss=0.2954, pruned_loss=0.08279, over 4900.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2381, pruned_loss=0.04549, over 953931.46 frames. ], batch size: 43, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:42:15,638 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153773.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:42:16,120 INFO [optim.py:369] (2/7) 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 01:42:17,993 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153777.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:42:44,723 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2495, 2.1303, 1.8307, 1.7448, 2.2348, 1.8028, 2.8476, 1.6393], + device='cuda:2'), covar=tensor([0.3847, 0.2124, 0.5101, 0.3451, 0.1887, 0.2695, 0.1526, 0.4680], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0355, 0.0424, 0.0351, 0.0380, 0.0374, 0.0369, 0.0422], + device='cuda:2'), 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:2') +2023-04-28 01:43:04,184 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5659, 2.7797, 2.1033, 2.3327, 2.4899, 2.2248, 3.6838, 1.8113], + device='cuda:2'), covar=tensor([0.3729, 0.2072, 0.4628, 0.3506, 0.2222, 0.2724, 0.1453, 0.4669], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0355, 0.0424, 0.0351, 0.0379, 0.0374, 0.0369, 0.0422], + device='cuda:2'), out_proj_covar=tensor([9.9795e-05, 1.0563e-04, 1.2840e-04, 1.0506e-04, 1.1227e-04, 1.1111e-04, + 1.0791e-04, 1.2680e-04], device='cuda:2') +2023-04-28 01:43:05,707 INFO [finetune.py:976] (2/7) Epoch 27, batch 4900, loss[loss=0.2186, simple_loss=0.2864, pruned_loss=0.07539, over 4851.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2388, pruned_loss=0.04595, over 952639.61 frames. ], batch size: 44, lr: 2.91e-03, grad_scale: 32.0 +2023-04-28 01:43:05,827 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6247, 1.3340, 1.3137, 1.0847, 1.3589, 1.2503, 1.6259, 1.2448], + device='cuda:2'), covar=tensor([0.3328, 0.1692, 0.4674, 0.2525, 0.1492, 0.1782, 0.1763, 0.4298], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0355, 0.0424, 0.0351, 0.0379, 0.0374, 0.0369, 0.0422], + device='cuda:2'), 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:2') +2023-04-28 01:43:06,908 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153821.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:44:09,302 INFO [finetune.py:976] (2/7) Epoch 27, batch 4950, loss[loss=0.1936, simple_loss=0.275, pruned_loss=0.05609, over 4838.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2405, pruned_loss=0.04635, over 953469.14 frames. ], batch size: 47, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:44:10,720 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-28 01:44:18,028 INFO [optim.py:369] (2/7) 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 01:45:13,176 INFO [finetune.py:976] (2/7) Epoch 27, batch 5000, loss[loss=0.1436, simple_loss=0.2176, pruned_loss=0.03483, over 4925.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2393, pruned_loss=0.04629, over 952641.69 frames. ], batch size: 38, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:45:21,165 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153923.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:45:25,218 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7776, 2.4331, 1.8245, 1.7674, 1.2834, 1.3083, 1.9632, 1.2446], + device='cuda:2'), covar=tensor([0.1588, 0.1304, 0.1327, 0.1642, 0.2256, 0.1811, 0.0919, 0.2011], + device='cuda:2'), in_proj_covar=tensor([0.0197, 0.0209, 0.0169, 0.0204, 0.0200, 0.0186, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 01:45:34,545 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153934.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:45:44,679 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0756, 1.8103, 2.1678, 2.4366, 2.4579, 1.9342, 1.7886, 2.1631], + device='cuda:2'), covar=tensor([0.0853, 0.1144, 0.0620, 0.0566, 0.0636, 0.0851, 0.0723, 0.0586], + device='cuda:2'), in_proj_covar=tensor([0.0186, 0.0203, 0.0185, 0.0172, 0.0179, 0.0180, 0.0153, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:45:44,842 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-04-28 01:46:07,451 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-28 01:46:17,421 INFO [finetune.py:976] (2/7) Epoch 27, batch 5050, loss[loss=0.1162, simple_loss=0.1946, pruned_loss=0.01891, over 4832.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2371, pruned_loss=0.04569, over 951746.15 frames. ], batch size: 33, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:46:25,377 INFO [optim.py:369] (2/7) 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 01:46:32,336 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153982.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:46:34,100 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 27, batch 5100, loss[loss=0.1399, simple_loss=0.2219, pruned_loss=0.02891, over 4808.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.2337, pruned_loss=0.04449, over 952586.97 frames. ], batch size: 51, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:47:18,516 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154049.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 01:47:28,931 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.4345, 1.2983, 1.4149, 1.0293, 1.3333, 1.1047, 1.7477, 1.3088], + device='cuda:2'), covar=tensor([0.3347, 0.1800, 0.4879, 0.2692, 0.1440, 0.2129, 0.1630, 0.4491], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0353, 0.0422, 0.0350, 0.0379, 0.0374, 0.0368, 0.0420], + device='cuda:2'), 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:2') +2023-04-28 01:47:31,249 INFO [finetune.py:976] (2/7) Epoch 27, batch 5150, loss[loss=0.1721, simple_loss=0.2475, pruned_loss=0.04834, over 4859.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2342, pruned_loss=0.04502, over 950612.37 frames. ], batch size: 34, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:47:33,690 INFO [optim.py:369] (2/7) 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 01:47:35,039 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154076.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:47:35,621 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154077.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:47:40,844 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154084.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:47:40,864 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0345, 0.8024, 0.7648, 0.8250, 1.1231, 0.9214, 0.8606, 0.8472], + device='cuda:2'), covar=tensor([0.1491, 0.1159, 0.1980, 0.1403, 0.0942, 0.1294, 0.1367, 0.2220], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0305, 0.0349, 0.0284, 0.0325, 0.0303, 0.0299, 0.0374], + device='cuda:2'), 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:2') +2023-04-28 01:48:15,439 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154110.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 01:48:26,894 INFO [finetune.py:976] (2/7) Epoch 27, batch 5200, loss[loss=0.1671, simple_loss=0.2463, pruned_loss=0.04394, over 4923.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2388, pruned_loss=0.04677, over 949855.89 frames. ], batch size: 38, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:48:35,421 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154125.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:48:36,815 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-28 01:48:49,924 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154137.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 01:49:00,228 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154145.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:49:20,753 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.5061, 4.4067, 3.1438, 5.1046, 4.5192, 4.4403, 1.9807, 4.4099], + device='cuda:2'), covar=tensor([0.1512, 0.0944, 0.3205, 0.1018, 0.2828, 0.1576, 0.5318, 0.2143], + device='cuda:2'), in_proj_covar=tensor([0.0244, 0.0218, 0.0250, 0.0303, 0.0298, 0.0247, 0.0272, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:49:21,567 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-04-28 01:49:31,678 INFO [finetune.py:976] (2/7) Epoch 27, batch 5250, loss[loss=0.1942, simple_loss=0.2715, pruned_loss=0.05841, over 4904.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2417, pruned_loss=0.04778, over 951401.16 frames. ], batch size: 37, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:49:34,127 INFO [optim.py:369] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.0783, 2.0553, 1.8195, 1.6875, 2.1360, 1.9172, 2.6637, 1.6254], + device='cuda:2'), covar=tensor([0.3586, 0.1936, 0.4502, 0.3050, 0.1601, 0.2065, 0.1320, 0.4277], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0353, 0.0423, 0.0349, 0.0378, 0.0374, 0.0368, 0.0421], + device='cuda:2'), 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:2') +2023-04-28 01:50:35,789 INFO [finetune.py:976] (2/7) Epoch 27, batch 5300, loss[loss=0.1694, simple_loss=0.2514, pruned_loss=0.04364, over 4817.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.242, pruned_loss=0.04736, over 953371.81 frames. ], batch size: 39, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:50:36,521 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154221.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:51:15,584 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 01:51:16,107 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5846, 2.0478, 1.7816, 1.9941, 1.5325, 1.7707, 1.6265, 1.3710], + device='cuda:2'), covar=tensor([0.1915, 0.1234, 0.0802, 0.1146, 0.3389, 0.1179, 0.1835, 0.2290], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0300, 0.0214, 0.0275, 0.0312, 0.0253, 0.0246, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1236e-04, 1.1792e-04, 8.4356e-05, 1.0801e-04, 1.2585e-04, 9.9299e-05, + 9.9104e-05, 1.0288e-04], device='cuda:2') +2023-04-28 01:51:41,568 INFO [finetune.py:976] (2/7) Epoch 27, batch 5350, loss[loss=0.1878, simple_loss=0.2576, pruned_loss=0.059, over 4858.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2421, pruned_loss=0.04671, over 955280.08 frames. ], batch size: 34, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:51:49,112 INFO [optim.py:369] (2/7) 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 01:51:52,261 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154279.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:51:54,101 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154282.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 01:51:56,599 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154286.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:00,108 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154291.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:19,624 INFO [finetune.py:976] (2/7) Epoch 27, batch 5400, loss[loss=0.1847, simple_loss=0.2504, pruned_loss=0.05949, over 4904.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2405, pruned_loss=0.04658, over 955209.96 frames. ], batch size: 43, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:52:30,054 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154337.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 01:52:37,683 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154347.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:41,220 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154352.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:52:44,940 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7903, 1.3185, 1.8867, 2.0902, 1.8103, 1.7712, 1.8021, 1.8239], + device='cuda:2'), covar=tensor([0.5429, 0.7834, 0.7381, 0.8324, 0.6691, 0.9477, 0.9497, 0.9965], + device='cuda:2'), in_proj_covar=tensor([0.0441, 0.0423, 0.0517, 0.0505, 0.0471, 0.0506, 0.0507, 0.0523], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:52:52,200 INFO [finetune.py:976] (2/7) Epoch 27, batch 5450, loss[loss=0.1363, simple_loss=0.2063, pruned_loss=0.03311, over 4849.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2367, pruned_loss=0.04522, over 956097.16 frames. ], batch size: 47, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:52:54,632 INFO [optim.py:369] (2/7) 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 01:53:09,450 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154398.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 01:53:16,766 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154405.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 01:53:20,460 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2520, 1.5322, 1.3792, 1.5194, 1.3079, 1.2966, 1.3943, 1.0889], + device='cuda:2'), covar=tensor([0.1701, 0.1370, 0.0883, 0.1139, 0.3402, 0.1185, 0.1600, 0.2069], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0302, 0.0216, 0.0277, 0.0314, 0.0254, 0.0248, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1299e-04, 1.1871e-04, 8.4936e-05, 1.0889e-04, 1.2676e-04, 1.0001e-04, + 9.9976e-05, 1.0340e-04], device='cuda:2') +2023-04-28 01:53:31,458 INFO [finetune.py:976] (2/7) Epoch 27, batch 5500, loss[loss=0.1573, simple_loss=0.2242, pruned_loss=0.04518, over 4820.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2342, pruned_loss=0.04526, over 953761.03 frames. ], batch size: 30, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:53:44,351 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154432.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 01:53:52,971 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-28 01:53:54,046 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154440.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:54:35,187 INFO [finetune.py:976] (2/7) Epoch 27, batch 5550, loss[loss=0.1469, simple_loss=0.2184, pruned_loss=0.03768, over 4343.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2364, pruned_loss=0.04584, over 952347.14 frames. ], batch size: 19, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:54:37,630 INFO [optim.py:369] (2/7) 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 01:54:37,744 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154474.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:55:07,208 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9178, 1.6972, 2.4148, 2.5542, 1.7422, 1.6397, 1.8783, 0.9396], + device='cuda:2'), covar=tensor([0.0649, 0.0792, 0.0356, 0.0734, 0.0822, 0.1148, 0.0822, 0.0811], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:55:31,959 INFO [finetune.py:976] (2/7) Epoch 27, batch 5600, loss[loss=0.1594, simple_loss=0.2367, pruned_loss=0.04099, over 4883.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2402, pruned_loss=0.04679, over 953346.57 frames. ], batch size: 32, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:55:50,541 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154535.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:56:00,978 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-04-28 01:56:32,738 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4482, 1.0189, 0.4890, 1.1780, 1.0491, 1.3489, 1.2654, 1.2712], + device='cuda:2'), covar=tensor([0.0486, 0.0395, 0.0368, 0.0561, 0.0289, 0.0500, 0.0480, 0.0550], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 01:56:35,596 INFO [finetune.py:976] (2/7) Epoch 27, batch 5650, loss[loss=0.1554, simple_loss=0.2324, pruned_loss=0.03917, over 4871.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2429, pruned_loss=0.04697, over 954053.08 frames. ], batch size: 34, lr: 2.90e-03, grad_scale: 64.0 +2023-04-28 01:56:42,718 INFO [optim.py:369] (2/7) 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 01:56:44,486 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154577.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 01:56:45,659 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154579.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:57:19,038 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-28 01:57:35,025 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3579, 2.8297, 2.6370, 2.7103, 2.5015, 2.6070, 2.7095, 2.7183], + device='cuda:2'), covar=tensor([0.3420, 0.5476, 0.4680, 0.4486, 0.5292, 0.6128, 0.5311, 0.4956], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0375, 0.0330, 0.0341, 0.0351, 0.0394, 0.0361, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 01:57:36,099 INFO [finetune.py:976] (2/7) Epoch 27, batch 5700, loss[loss=0.1741, simple_loss=0.2251, pruned_loss=0.06152, over 3960.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2389, pruned_loss=0.0468, over 933747.40 frames. ], batch size: 17, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:57:37,950 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1697, 3.6962, 3.2633, 3.6080, 2.8624, 3.3440, 3.4150, 2.8099], + device='cuda:2'), covar=tensor([0.1412, 0.0819, 0.0600, 0.0842, 0.2406, 0.0850, 0.1397, 0.2080], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0304, 0.0217, 0.0277, 0.0315, 0.0256, 0.0248, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1349e-04, 1.1954e-04, 8.5389e-05, 1.0901e-04, 1.2716e-04, 1.0042e-04, + 1.0019e-04, 1.0383e-04], device='cuda:2') +2023-04-28 01:57:45,551 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154627.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:57:46,204 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154628.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:57:59,107 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154642.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:58:12,013 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 0, loss[loss=0.1624, simple_loss=0.2395, pruned_loss=0.04267, over 4820.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2395, pruned_loss=0.04267, over 4820.00 frames. ], batch size: 30, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:58:12,561 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 01:58:17,132 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4739, 1.2176, 1.2843, 1.2285, 1.5816, 1.3588, 1.1851, 1.2638], + device='cuda:2'), covar=tensor([0.1640, 0.1247, 0.1550, 0.1316, 0.0785, 0.1338, 0.1629, 0.2194], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0306, 0.0350, 0.0286, 0.0325, 0.0304, 0.0300, 0.0375], + device='cuda:2'), 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:2') +2023-04-28 01:58:32,530 INFO [finetune.py:1010] (2/7) Epoch 28, validation: loss=0.1549, simple_loss=0.224, pruned_loss=0.04297, over 2265189.00 frames. +2023-04-28 01:58:32,532 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6415MB +2023-04-28 01:58:33,234 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0212, 2.5531, 1.0533, 1.4040, 1.8880, 1.1586, 3.3567, 1.8286], + device='cuda:2'), covar=tensor([0.0723, 0.0657, 0.0808, 0.1210, 0.0538, 0.1058, 0.0221, 0.0620], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0050, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 01:58:43,444 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2805, 2.4056, 0.8230, 1.4257, 1.5116, 1.8126, 1.5144, 0.8906], + device='cuda:2'), covar=tensor([0.1769, 0.1370, 0.2231, 0.1714, 0.1391, 0.1184, 0.2033, 0.1981], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0136, 0.0121, 0.0132, 0.0152, 0.0118, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 01:58:55,769 INFO [optim.py:369] (2/7) 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 01:59:03,274 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6547, 1.4892, 1.7031, 1.9947, 2.0424, 1.6001, 1.4542, 1.7771], + device='cuda:2'), covar=tensor([0.0818, 0.1371, 0.0885, 0.0619, 0.0657, 0.0816, 0.0703, 0.0595], + device='cuda:2'), in_proj_covar=tensor([0.0186, 0.0204, 0.0186, 0.0173, 0.0180, 0.0179, 0.0153, 0.0180], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:59:04,499 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154689.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:06,857 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154693.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 01:59:09,839 INFO [finetune.py:976] (2/7) Epoch 28, batch 50, loss[loss=0.1947, simple_loss=0.2611, pruned_loss=0.06412, over 4858.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2451, pruned_loss=0.04957, over 216357.37 frames. ], batch size: 31, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:59:17,146 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154705.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 01:59:26,294 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154720.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:33,613 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154732.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:38,581 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154740.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 01:59:40,632 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-28 01:59:43,314 INFO [finetune.py:976] (2/7) Epoch 28, batch 100, loss[loss=0.1375, simple_loss=0.217, pruned_loss=0.02895, over 4778.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2392, pruned_loss=0.04732, over 379452.22 frames. ], batch size: 28, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 01:59:43,400 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1357, 1.2764, 5.0392, 4.7557, 4.3602, 4.7848, 4.4999, 4.4342], + device='cuda:2'), covar=tensor([0.6520, 0.6152, 0.0939, 0.1667, 0.1068, 0.1114, 0.1433, 0.1569], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0311, 0.0409, 0.0411, 0.0351, 0.0419, 0.0320, 0.0367], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 01:59:48,391 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154753.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:00:02,205 INFO [optim.py:369] (2/7) 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 02:00:05,350 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154780.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:00:06,018 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154781.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:00:07,840 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4432, 1.3665, 1.6791, 1.7159, 1.3061, 1.2509, 1.3889, 0.8637], + device='cuda:2'), covar=tensor([0.0470, 0.0504, 0.0331, 0.0485, 0.0727, 0.0951, 0.0514, 0.0514], + device='cuda:2'), in_proj_covar=tensor([0.0069, 0.0067, 0.0065, 0.0068, 0.0074, 0.0093, 0.0072, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 02:00:10,229 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 150, loss[loss=0.1498, simple_loss=0.218, pruned_loss=0.04077, over 4911.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.2324, pruned_loss=0.0452, over 504836.73 frames. ], batch size: 43, lr: 2.90e-03, grad_scale: 32.0 +2023-04-28 02:00:29,330 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1189, 2.7518, 2.1358, 2.2194, 1.5986, 1.5880, 2.1987, 1.5477], + device='cuda:2'), covar=tensor([0.1609, 0.1249, 0.1332, 0.1440, 0.2136, 0.1865, 0.0918, 0.1947], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0210, 0.0171, 0.0205, 0.0201, 0.0187, 0.0157, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 02:00:38,342 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154830.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:00:49,253 INFO [finetune.py:976] (2/7) Epoch 28, batch 200, loss[loss=0.1249, simple_loss=0.2057, pruned_loss=0.02201, over 4820.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.2297, pruned_loss=0.04384, over 605771.04 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:00:50,622 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6812, 2.4687, 2.3364, 2.2863, 2.6712, 2.2607, 3.2341, 2.1332], + device='cuda:2'), covar=tensor([0.3292, 0.1790, 0.3753, 0.2678, 0.1593, 0.2542, 0.1355, 0.3778], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0356, 0.0425, 0.0352, 0.0382, 0.0378, 0.0370, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:01:08,233 INFO [optim.py:369] (2/7) 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 02:01:09,541 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154877.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 02:01:22,150 INFO [finetune.py:976] (2/7) Epoch 28, batch 250, loss[loss=0.1307, simple_loss=0.2092, pruned_loss=0.02615, over 4758.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2348, pruned_loss=0.04505, over 683147.53 frames. ], batch size: 28, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:01:35,621 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.16 vs. limit=5.0 +2023-04-28 02:01:41,469 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154925.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:01:51,916 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154942.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:01:54,966 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154947.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:00,647 INFO [finetune.py:976] (2/7) Epoch 28, batch 300, loss[loss=0.1448, simple_loss=0.2157, pruned_loss=0.03696, over 4871.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2388, pruned_loss=0.04583, over 743777.86 frames. ], batch size: 31, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:02:11,865 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2925, 1.8818, 1.6893, 2.4129, 2.5811, 2.0528, 1.9814, 1.7207], + device='cuda:2'), covar=tensor([0.2470, 0.1878, 0.2024, 0.1585, 0.1317, 0.2064, 0.2498, 0.2704], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0308, 0.0351, 0.0286, 0.0326, 0.0305, 0.0301, 0.0376], + device='cuda:2'), out_proj_covar=tensor([6.4256e-05, 6.3268e-05, 7.3557e-05, 5.7292e-05, 6.6496e-05, 6.3549e-05, + 6.2309e-05, 7.9693e-05], device='cuda:2') +2023-04-28 02:02:36,611 INFO [optim.py:369] (2/7) 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 02:02:37,346 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154976.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:47,498 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154984.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:56,112 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154990.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:02:57,993 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154993.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 02:02:59,162 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 350, loss[loss=0.1483, simple_loss=0.2263, pruned_loss=0.03511, over 4844.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2414, pruned_loss=0.04691, over 791770.20 frames. ], batch size: 47, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:03:28,775 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3840, 2.0442, 2.2487, 2.9817, 2.4133, 1.9137, 2.1572, 2.3476], + device='cuda:2'), covar=tensor([0.2411, 0.2648, 0.1494, 0.1678, 0.2285, 0.2281, 0.3143, 0.1736], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0248, 0.0230, 0.0316, 0.0224, 0.0237, 0.0230, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 02:03:33,790 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5936, 1.3011, 4.1052, 3.8665, 3.5370, 3.9013, 3.8230, 3.6401], + device='cuda:2'), covar=tensor([0.6921, 0.5699, 0.1117, 0.1665, 0.1113, 0.1516, 0.2278, 0.1348], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0310, 0.0408, 0.0410, 0.0350, 0.0418, 0.0319, 0.0366], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:03:45,311 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155037.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:03:47,613 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155041.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:03:51,695 INFO [finetune.py:976] (2/7) Epoch 28, batch 400, loss[loss=0.2358, simple_loss=0.2978, pruned_loss=0.08691, over 4894.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2433, pruned_loss=0.0481, over 828990.10 frames. ], batch size: 36, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:04:06,190 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155069.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:11,280 INFO [optim.py:369] (2/7) 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 02:04:11,988 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155076.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:13,372 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.79 vs. limit=5.0 +2023-04-28 02:04:25,388 INFO [finetune.py:976] (2/7) Epoch 28, batch 450, loss[loss=0.1552, simple_loss=0.2249, pruned_loss=0.04281, over 4896.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2415, pruned_loss=0.04717, over 856288.34 frames. ], batch size: 37, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:04:48,040 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:48,069 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:04:58,929 INFO [finetune.py:976] (2/7) Epoch 28, batch 500, loss[loss=0.1853, simple_loss=0.2533, pruned_loss=0.05861, over 4832.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2379, pruned_loss=0.04603, over 878938.81 frames. ], batch size: 33, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:05:00,292 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2912, 3.0438, 2.5925, 2.9134, 2.1445, 2.4812, 2.6506, 2.0471], + device='cuda:2'), covar=tensor([0.2142, 0.1040, 0.0776, 0.1095, 0.3177, 0.1159, 0.2085, 0.2620], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0303, 0.0216, 0.0275, 0.0312, 0.0254, 0.0248, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1274e-04, 1.1921e-04, 8.4992e-05, 1.0835e-04, 1.2595e-04, 9.9897e-05, + 1.0003e-04, 1.0341e-04], device='cuda:2') +2023-04-28 02:05:17,835 INFO [optim.py:369] (2/7) 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 02:05:20,242 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155178.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:05:25,168 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-04-28 02:05:32,277 INFO [finetune.py:976] (2/7) Epoch 28, batch 550, loss[loss=0.1938, simple_loss=0.2591, pruned_loss=0.06422, over 4816.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2361, pruned_loss=0.04587, over 894645.25 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:05:44,717 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0133, 1.0224, 1.2001, 1.1530, 0.9913, 0.8881, 0.9399, 0.4657], + device='cuda:2'), covar=tensor([0.0549, 0.0534, 0.0410, 0.0577, 0.0630, 0.1224, 0.0486, 0.0647], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0068, 0.0066, 0.0069, 0.0075, 0.0095, 0.0073, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 02:06:05,934 INFO [finetune.py:976] (2/7) Epoch 28, batch 600, loss[loss=0.1525, simple_loss=0.2368, pruned_loss=0.0341, over 4834.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2376, pruned_loss=0.04644, over 910063.57 frames. ], batch size: 47, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:06:23,749 INFO [optim.py:369] (2/7) 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 02:06:30,723 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155284.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:06:39,097 INFO [finetune.py:976] (2/7) Epoch 28, batch 650, loss[loss=0.1992, simple_loss=0.2739, pruned_loss=0.0622, over 4758.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2416, pruned_loss=0.04784, over 921131.99 frames. ], batch size: 54, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:06:41,051 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3033, 1.8556, 2.0944, 2.6505, 2.6420, 2.2036, 2.0021, 2.4980], + device='cuda:2'), covar=tensor([0.0760, 0.1263, 0.0799, 0.0572, 0.0571, 0.0822, 0.0678, 0.0487], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0201, 0.0183, 0.0170, 0.0177, 0.0176, 0.0150, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:07:02,744 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155332.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:07:12,880 INFO [finetune.py:976] (2/7) Epoch 28, batch 700, loss[loss=0.1525, simple_loss=0.2463, pruned_loss=0.02937, over 4898.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2422, pruned_loss=0.04751, over 928596.36 frames. ], batch size: 43, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:07:18,024 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-04-28 02:07:34,633 INFO [optim.py:369] (2/7) 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 02:07:40,201 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155376.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:08:09,455 INFO [finetune.py:976] (2/7) Epoch 28, batch 750, loss[loss=0.1484, simple_loss=0.2326, pruned_loss=0.03216, over 4905.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2426, pruned_loss=0.04753, over 933454.82 frames. ], batch size: 37, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:08:41,160 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155425.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:09:12,225 INFO [finetune.py:976] (2/7) Epoch 28, batch 800, loss[loss=0.1362, simple_loss=0.2157, pruned_loss=0.02832, over 4918.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2419, pruned_loss=0.04668, over 937478.04 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:09:22,989 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155457.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:09:42,706 INFO [optim.py:369] (2/7) 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 02:09:58,621 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.7913, 3.8143, 2.9089, 4.4871, 3.9313, 3.9023, 1.6380, 3.8125], + device='cuda:2'), covar=tensor([0.1819, 0.1158, 0.3351, 0.1421, 0.3405, 0.1726, 0.6161, 0.2374], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0220, 0.0252, 0.0305, 0.0299, 0.0250, 0.0275, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 02:09:59,777 INFO [finetune.py:976] (2/7) Epoch 28, batch 850, loss[loss=0.16, simple_loss=0.2312, pruned_loss=0.0444, over 4856.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2403, pruned_loss=0.04642, over 938961.91 frames. ], batch size: 49, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:10:12,004 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155518.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:10:28,869 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4259, 2.9979, 0.9412, 1.7185, 1.8078, 2.0581, 1.8439, 1.0184], + device='cuda:2'), covar=tensor([0.1483, 0.1417, 0.2044, 0.1474, 0.1151, 0.1334, 0.1650, 0.2061], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0239, 0.0135, 0.0121, 0.0131, 0.0153, 0.0117, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 02:10:32,819 INFO [finetune.py:976] (2/7) Epoch 28, batch 900, loss[loss=0.1985, simple_loss=0.2584, pruned_loss=0.06934, over 4862.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2374, pruned_loss=0.04533, over 941712.68 frames. ], batch size: 49, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:10:49,546 INFO [optim.py:369] (2/7) 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 02:11:05,999 INFO [finetune.py:976] (2/7) Epoch 28, batch 950, loss[loss=0.1485, simple_loss=0.2185, pruned_loss=0.03927, over 4903.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2359, pruned_loss=0.04481, over 944855.53 frames. ], batch size: 32, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:11:10,952 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155605.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:11:27,337 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155632.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:11:39,453 INFO [finetune.py:976] (2/7) Epoch 28, batch 1000, loss[loss=0.1759, simple_loss=0.2654, pruned_loss=0.04324, over 4815.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2362, pruned_loss=0.04487, over 948811.84 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:11:45,541 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155657.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:11:49,201 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2381, 2.5000, 2.1476, 2.4310, 1.5650, 1.9942, 2.2014, 1.7322], + device='cuda:2'), covar=tensor([0.1674, 0.1110, 0.0889, 0.1220, 0.3871, 0.1189, 0.1768, 0.2310], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0303, 0.0216, 0.0276, 0.0314, 0.0254, 0.0248, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1237e-04, 1.1923e-04, 8.4970e-05, 1.0842e-04, 1.2640e-04, 9.9779e-05, + 9.9882e-05, 1.0335e-04], device='cuda:2') +2023-04-28 02:11:51,046 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155680.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:12:10,798 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-28 02:12:12,448 INFO [finetune.py:976] (2/7) Epoch 28, batch 1050, loss[loss=0.1618, simple_loss=0.2521, pruned_loss=0.03571, over 4802.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2412, pruned_loss=0.04691, over 949584.82 frames. ], batch size: 41, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:12:25,700 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155718.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:12:29,976 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155725.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:12:45,200 INFO [finetune.py:976] (2/7) Epoch 28, batch 1100, loss[loss=0.1761, simple_loss=0.2598, pruned_loss=0.04621, over 4810.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2421, pruned_loss=0.04689, over 950313.83 frames. ], batch size: 51, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:13:13,292 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155773.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:13:19,856 INFO [optim.py:369] (2/7) 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 02:13:30,944 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155784.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:13:46,059 INFO [finetune.py:976] (2/7) Epoch 28, batch 1150, loss[loss=0.1644, simple_loss=0.2367, pruned_loss=0.046, over 4882.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.242, pruned_loss=0.0468, over 952183.49 frames. ], batch size: 32, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:13:58,506 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155813.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:14:18,110 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155845.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:14:19,787 INFO [finetune.py:976] (2/7) Epoch 28, batch 1200, loss[loss=0.1435, simple_loss=0.215, pruned_loss=0.03603, over 4795.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2409, pruned_loss=0.04666, over 953196.10 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:14:31,094 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6726, 1.7068, 0.6348, 1.3744, 1.9029, 1.5268, 1.4288, 1.4962], + device='cuda:2'), covar=tensor([0.0482, 0.0373, 0.0350, 0.0556, 0.0269, 0.0523, 0.0523, 0.0572], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 02:14:51,964 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-04-28 02:14:54,753 INFO [optim.py:369] (2/7) 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 02:15:25,827 INFO [finetune.py:976] (2/7) Epoch 28, batch 1250, loss[loss=0.1288, simple_loss=0.2121, pruned_loss=0.02276, over 4766.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2391, pruned_loss=0.04693, over 953780.39 frames. ], batch size: 27, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:16:33,390 INFO [finetune.py:976] (2/7) Epoch 28, batch 1300, loss[loss=0.1752, simple_loss=0.2485, pruned_loss=0.05096, over 4856.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.237, pruned_loss=0.04637, over 954224.29 frames. ], batch size: 44, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:16:53,069 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:17:37,317 INFO [finetune.py:976] (2/7) Epoch 28, batch 1350, loss[loss=0.2168, simple_loss=0.2817, pruned_loss=0.07593, over 4222.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2362, pruned_loss=0.04592, over 954218.68 frames. ], batch size: 65, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:17:56,770 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156013.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:18:17,965 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-04-28 02:18:40,867 INFO [finetune.py:976] (2/7) Epoch 28, batch 1400, loss[loss=0.1595, simple_loss=0.2455, pruned_loss=0.03677, over 4853.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2405, pruned_loss=0.04747, over 955347.65 frames. ], batch size: 44, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:19:12,184 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156070.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:19:20,859 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 28, batch 1450, loss[loss=0.1288, simple_loss=0.2161, pruned_loss=0.02078, over 4758.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2411, pruned_loss=0.04659, over 955938.21 frames. ], batch size: 28, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:19:39,080 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-28 02:19:46,155 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156113.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:19:53,767 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9339, 1.5580, 2.0805, 2.3449, 2.0486, 1.9344, 2.0187, 1.9422], + device='cuda:2'), covar=tensor([0.4291, 0.6686, 0.5883, 0.5595, 0.5473, 0.8082, 0.7673, 0.9377], + device='cuda:2'), in_proj_covar=tensor([0.0443, 0.0425, 0.0520, 0.0507, 0.0472, 0.0509, 0.0510, 0.0524], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:19:59,759 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156140.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:20:09,928 INFO [finetune.py:976] (2/7) Epoch 28, batch 1500, loss[loss=0.1647, simple_loss=0.2404, pruned_loss=0.04452, over 4150.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2421, pruned_loss=0.04684, over 954590.52 frames. ], batch size: 65, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:20:23,241 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:21:08,783 INFO [finetune.py:976] (2/7) Epoch 28, batch 1550, loss[loss=0.1893, simple_loss=0.2519, pruned_loss=0.06334, over 4811.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2419, pruned_loss=0.04661, over 954256.52 frames. ], batch size: 40, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:21:10,372 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 02:21:31,138 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156216.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:04,807 INFO [finetune.py:976] (2/7) Epoch 28, batch 1600, loss[loss=0.1624, simple_loss=0.2314, pruned_loss=0.04668, over 4865.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2393, pruned_loss=0.04599, over 954367.05 frames. ], batch size: 34, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:22:12,818 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156261.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:24,232 INFO [optim.py:369] (2/7) 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 02:22:25,027 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4533, 1.9278, 2.3366, 2.6740, 2.3534, 1.7743, 1.5888, 2.1745], + device='cuda:2'), covar=tensor([0.2897, 0.2732, 0.1482, 0.2082, 0.2355, 0.2588, 0.3904, 0.1758], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0228, 0.0313, 0.0221, 0.0233, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 02:22:25,631 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156277.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:34,033 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156290.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:38,825 INFO [finetune.py:976] (2/7) Epoch 28, batch 1650, loss[loss=0.1352, simple_loss=0.2074, pruned_loss=0.0315, over 4757.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.237, pruned_loss=0.04534, over 957704.89 frames. ], batch size: 27, lr: 2.89e-03, grad_scale: 32.0 +2023-04-28 02:22:45,574 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156309.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:22:46,296 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4267, 1.8398, 2.3155, 2.6514, 2.3351, 1.8239, 1.4640, 2.1196], + device='cuda:2'), covar=tensor([0.3148, 0.3048, 0.1625, 0.2188, 0.2348, 0.2529, 0.3921, 0.1745], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0227, 0.0313, 0.0221, 0.0233, 0.0228, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 02:22:48,048 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 1700, loss[loss=0.1738, simple_loss=0.2405, pruned_loss=0.05352, over 4843.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2359, pruned_loss=0.04541, over 959168.34 frames. ], batch size: 47, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:23:14,220 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156361.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:23:28,668 INFO [optim.py:369] (2/7) 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 02:23:45,148 INFO [finetune.py:976] (2/7) Epoch 28, batch 1750, loss[loss=0.1389, simple_loss=0.2212, pruned_loss=0.02834, over 4774.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2379, pruned_loss=0.04607, over 958260.62 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:23:50,880 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5065, 1.0917, 0.4268, 1.2584, 1.0536, 1.3833, 1.3221, 1.3269], + device='cuda:2'), covar=tensor([0.0510, 0.0390, 0.0422, 0.0537, 0.0326, 0.0504, 0.0499, 0.0552], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 02:24:08,541 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156426.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 02:24:29,107 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156440.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 02:24:39,110 INFO [finetune.py:976] (2/7) Epoch 28, batch 1800, loss[loss=0.1908, simple_loss=0.2739, pruned_loss=0.05385, over 4766.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2398, pruned_loss=0.04634, over 957745.64 frames. ], batch size: 54, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:25:11,110 INFO [optim.py:369] (2/7) 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 02:25:25,172 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5363, 1.1104, 1.2500, 1.1785, 1.6925, 1.3221, 1.2148, 1.1818], + device='cuda:2'), covar=tensor([0.1602, 0.1397, 0.1830, 0.1344, 0.0916, 0.1501, 0.1608, 0.2184], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0306, 0.0351, 0.0285, 0.0324, 0.0305, 0.0299, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.4093e-05, 6.2776e-05, 7.3551e-05, 5.7001e-05, 6.6151e-05, 6.3545e-05, + 6.1855e-05, 7.9281e-05], device='cuda:2') +2023-04-28 02:25:31,032 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 1850, loss[loss=0.1537, simple_loss=0.231, pruned_loss=0.03818, over 4752.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2414, pruned_loss=0.04722, over 957118.91 frames. ], batch size: 27, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:26:39,956 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0214, 1.8760, 1.9389, 1.6352, 2.1481, 1.7306, 2.6370, 1.7243], + device='cuda:2'), covar=tensor([0.3353, 0.1809, 0.4217, 0.2797, 0.1340, 0.2258, 0.1135, 0.4038], + device='cuda:2'), in_proj_covar=tensor([0.0341, 0.0357, 0.0427, 0.0352, 0.0383, 0.0379, 0.0372, 0.0425], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:26:48,597 INFO [finetune.py:976] (2/7) Epoch 28, batch 1900, loss[loss=0.1626, simple_loss=0.2372, pruned_loss=0.04404, over 4819.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2423, pruned_loss=0.04764, over 956775.13 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:27:20,593 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156572.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:27:22,347 INFO [optim.py:369] (2/7) 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 02:27:54,785 INFO [finetune.py:976] (2/7) Epoch 28, batch 1950, loss[loss=0.2287, simple_loss=0.2785, pruned_loss=0.08941, over 4762.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2413, pruned_loss=0.04783, over 955684.16 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:28:02,312 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9745, 4.3652, 0.7443, 2.3118, 2.4819, 2.9009, 2.3975, 1.0129], + device='cuda:2'), covar=tensor([0.1286, 0.0823, 0.2039, 0.1180, 0.1023, 0.1104, 0.1462, 0.2118], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0241, 0.0136, 0.0121, 0.0132, 0.0153, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 02:28:14,638 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3106, 1.5200, 1.8895, 1.9788, 1.9076, 1.9303, 1.8939, 1.9277], + device='cuda:2'), covar=tensor([0.3636, 0.4813, 0.4073, 0.4070, 0.4999, 0.6539, 0.4626, 0.4229], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0374, 0.0330, 0.0340, 0.0349, 0.0391, 0.0359, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 02:28:50,761 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5947, 1.9076, 2.0444, 2.0973, 1.9625, 2.0048, 2.0791, 2.0640], + device='cuda:2'), covar=tensor([0.3626, 0.5138, 0.4269, 0.4152, 0.5142, 0.6630, 0.4909, 0.4722], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0373, 0.0329, 0.0340, 0.0349, 0.0391, 0.0359, 0.0333], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 02:29:00,503 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156646.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:29:01,641 INFO [finetune.py:976] (2/7) Epoch 28, batch 2000, loss[loss=0.1424, simple_loss=0.2151, pruned_loss=0.03481, over 4908.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2387, pruned_loss=0.04672, over 957010.67 frames. ], batch size: 43, lr: 2.88e-03, grad_scale: 64.0 +2023-04-28 02:29:23,947 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2492, 1.4919, 1.4131, 1.5009, 1.2863, 1.3096, 1.4083, 1.0406], + device='cuda:2'), covar=tensor([0.1716, 0.1246, 0.0868, 0.1120, 0.3429, 0.1136, 0.1642, 0.2074], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0301, 0.0217, 0.0275, 0.0314, 0.0253, 0.0247, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1231e-04, 1.1854e-04, 8.5122e-05, 1.0835e-04, 1.2649e-04, 9.9489e-05, + 9.9542e-05, 1.0350e-04], device='cuda:2') +2023-04-28 02:29:34,745 INFO [optim.py:369] (2/7) 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 02:30:06,048 INFO [finetune.py:976] (2/7) Epoch 28, batch 2050, loss[loss=0.1963, simple_loss=0.2664, pruned_loss=0.06313, over 4924.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2358, pruned_loss=0.04577, over 958382.81 frames. ], batch size: 43, lr: 2.88e-03, grad_scale: 64.0 +2023-04-28 02:30:43,829 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156726.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:30:45,676 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7687, 2.1086, 1.7893, 1.4641, 1.2613, 1.3399, 1.8371, 1.2510], + device='cuda:2'), covar=tensor([0.1710, 0.1238, 0.1583, 0.1847, 0.2470, 0.2144, 0.1045, 0.2251], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0209, 0.0169, 0.0205, 0.0200, 0.0186, 0.0156, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 02:30:56,464 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2986, 1.6021, 1.4669, 1.8622, 1.6707, 2.0530, 1.4289, 3.7157], + device='cuda:2'), covar=tensor([0.0638, 0.0779, 0.0850, 0.1207, 0.0658, 0.0511, 0.0749, 0.0144], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0042, 0.0040, 0.0038, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 02:31:05,851 INFO [finetune.py:976] (2/7) Epoch 28, batch 2100, loss[loss=0.1366, simple_loss=0.2064, pruned_loss=0.03343, over 4789.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2358, pruned_loss=0.04584, over 957174.63 frames. ], batch size: 29, lr: 2.88e-03, grad_scale: 64.0 +2023-04-28 02:31:07,076 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156749.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:31:39,407 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156774.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:31:39,961 INFO [optim.py:369] (2/7) 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 02:31:40,177 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.49 vs. limit=2.0 +2023-04-28 02:32:10,051 INFO [finetune.py:976] (2/7) Epoch 28, batch 2150, loss[loss=0.1914, simple_loss=0.2658, pruned_loss=0.05847, over 4748.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2399, pruned_loss=0.04692, over 955237.70 frames. ], batch size: 54, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:32:22,733 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156810.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:32:57,257 INFO [finetune.py:976] (2/7) Epoch 28, batch 2200, loss[loss=0.1772, simple_loss=0.2615, pruned_loss=0.04643, over 4724.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2413, pruned_loss=0.04682, over 957362.44 frames. ], batch size: 54, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:33:09,563 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2872, 1.8110, 2.1374, 2.5052, 2.1246, 1.7098, 1.4257, 1.9970], + device='cuda:2'), covar=tensor([0.2751, 0.2784, 0.1528, 0.1984, 0.2345, 0.2400, 0.3909, 0.1731], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0246, 0.0230, 0.0315, 0.0222, 0.0235, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 02:33:15,054 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:33:23,512 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 2250, loss[loss=0.1349, simple_loss=0.2161, pruned_loss=0.02684, over 4800.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2429, pruned_loss=0.04744, over 954248.84 frames. ], batch size: 40, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:33:40,658 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2322, 1.8733, 2.1161, 2.5721, 2.5456, 2.0381, 1.9373, 2.2760], + device='cuda:2'), covar=tensor([0.0789, 0.1257, 0.0787, 0.0562, 0.0605, 0.0839, 0.0683, 0.0568], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0203, 0.0186, 0.0171, 0.0179, 0.0179, 0.0152, 0.0179], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:33:47,977 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156920.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:04,146 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:04,191 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:05,298 INFO [finetune.py:976] (2/7) Epoch 28, batch 2300, loss[loss=0.1471, simple_loss=0.2227, pruned_loss=0.03571, over 4921.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2428, pruned_loss=0.04692, over 956594.89 frames. ], batch size: 33, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:34:25,202 INFO [optim.py:369] (2/7) 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 02:34:28,282 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-28 02:34:36,691 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156994.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:39,082 INFO [finetune.py:976] (2/7) Epoch 28, batch 2350, loss[loss=0.1374, simple_loss=0.2108, pruned_loss=0.03197, over 4851.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.241, pruned_loss=0.047, over 956201.40 frames. ], batch size: 49, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:34:39,844 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-04-28 02:34:47,985 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157010.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:34:53,077 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9163, 1.7619, 2.5388, 2.5493, 1.8216, 1.6186, 2.0038, 1.1812], + device='cuda:2'), covar=tensor([0.0566, 0.0653, 0.0331, 0.0560, 0.0684, 0.1073, 0.0558, 0.0666], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0068, 0.0065, 0.0069, 0.0075, 0.0095, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 02:34:58,743 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 02:35:23,601 INFO [finetune.py:976] (2/7) Epoch 28, batch 2400, loss[loss=0.1775, simple_loss=0.2647, pruned_loss=0.04517, over 4829.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2379, pruned_loss=0.04606, over 957219.17 frames. ], batch size: 39, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:35:50,212 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157071.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:35:53,124 INFO [optim.py:369] (2/7) 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 02:36:07,059 INFO [finetune.py:976] (2/7) Epoch 28, batch 2450, loss[loss=0.1512, simple_loss=0.221, pruned_loss=0.04066, over 4839.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2368, pruned_loss=0.0461, over 957241.34 frames. ], batch size: 49, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:36:11,410 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157105.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:36:34,457 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0906, 2.4861, 2.1778, 2.5220, 1.8032, 2.1400, 2.1461, 1.5818], + device='cuda:2'), covar=tensor([0.1662, 0.1134, 0.0783, 0.0911, 0.3081, 0.1099, 0.1707, 0.2498], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0303, 0.0218, 0.0277, 0.0316, 0.0254, 0.0249, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1310e-04, 1.1903e-04, 8.5565e-05, 1.0898e-04, 1.2739e-04, 9.9954e-05, + 1.0014e-04, 1.0411e-04], device='cuda:2') +2023-04-28 02:36:36,268 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8102, 2.8529, 2.1653, 3.2735, 2.8554, 2.8507, 1.2068, 2.8121], + device='cuda:2'), covar=tensor([0.2349, 0.1780, 0.3390, 0.2878, 0.4478, 0.2297, 0.6234, 0.2816], + device='cuda:2'), in_proj_covar=tensor([0.0245, 0.0220, 0.0251, 0.0303, 0.0298, 0.0248, 0.0274, 0.0271], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 02:36:45,809 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-04-28 02:36:58,724 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 2500, loss[loss=0.1215, simple_loss=0.2037, pruned_loss=0.01963, over 4793.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2374, pruned_loss=0.04594, over 955345.53 frames. ], batch size: 29, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:37:44,330 INFO [optim.py:369] (2/7) 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 02:37:51,578 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.77 vs. limit=5.0 +2023-04-28 02:38:03,438 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-04-28 02:38:14,042 INFO [finetune.py:976] (2/7) Epoch 28, batch 2550, loss[loss=0.1582, simple_loss=0.2391, pruned_loss=0.03869, over 4901.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2403, pruned_loss=0.04609, over 954580.32 frames. ], batch size: 36, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:38:22,917 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157203.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:38:54,086 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6419, 1.4199, 0.6116, 1.2899, 1.4286, 1.4771, 1.3420, 1.3948], + device='cuda:2'), covar=tensor([0.0473, 0.0375, 0.0378, 0.0540, 0.0302, 0.0518, 0.0501, 0.0547], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0054, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 02:39:05,865 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1162, 1.4815, 1.4797, 1.7000, 2.0702, 1.7296, 1.5484, 1.4433], + device='cuda:2'), covar=tensor([0.1617, 0.1900, 0.2377, 0.1516, 0.1134, 0.1839, 0.2006, 0.2523], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0306, 0.0349, 0.0284, 0.0323, 0.0305, 0.0299, 0.0374], + device='cuda:2'), out_proj_covar=tensor([6.3899e-05, 6.2676e-05, 7.3209e-05, 5.6875e-05, 6.5981e-05, 6.3446e-05, + 6.1735e-05, 7.9211e-05], device='cuda:2') +2023-04-28 02:39:11,686 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157241.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:39:16,444 INFO [finetune.py:976] (2/7) Epoch 28, batch 2600, loss[loss=0.153, simple_loss=0.2398, pruned_loss=0.03311, over 4900.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2428, pruned_loss=0.04749, over 954973.71 frames. ], batch size: 43, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:39:50,654 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 28, batch 2650, loss[loss=0.112, simple_loss=0.1841, pruned_loss=0.01989, over 4693.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2443, pruned_loss=0.04831, over 956020.14 frames. ], batch size: 23, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:40:53,194 INFO [finetune.py:976] (2/7) Epoch 28, batch 2700, loss[loss=0.182, simple_loss=0.2454, pruned_loss=0.05932, over 4180.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2421, pruned_loss=0.04729, over 955069.08 frames. ], batch size: 18, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:40:58,694 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157356.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:04,713 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157366.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:10,644 INFO [optim.py:369] (2/7) 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 02:41:15,937 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157382.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:15,954 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9682, 2.5868, 1.8981, 1.8454, 1.4463, 1.4834, 2.1068, 1.4167], + device='cuda:2'), covar=tensor([0.1576, 0.1132, 0.1301, 0.1552, 0.2055, 0.1832, 0.0839, 0.1865], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0209, 0.0169, 0.0204, 0.0200, 0.0186, 0.0156, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 02:41:23,776 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1871, 2.8941, 2.5254, 2.7918, 2.0791, 2.5338, 2.5824, 1.9175], + device='cuda:2'), covar=tensor([0.2277, 0.1198, 0.0769, 0.1187, 0.3164, 0.1017, 0.1992, 0.2870], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0302, 0.0217, 0.0276, 0.0315, 0.0254, 0.0248, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1298e-04, 1.1886e-04, 8.5544e-05, 1.0877e-04, 1.2709e-04, 9.9764e-05, + 9.9997e-05, 1.0366e-04], device='cuda:2') +2023-04-28 02:41:26,099 INFO [finetune.py:976] (2/7) Epoch 28, batch 2750, loss[loss=0.2075, simple_loss=0.2647, pruned_loss=0.07519, over 4921.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2398, pruned_loss=0.04689, over 956069.50 frames. ], batch size: 37, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:41:30,003 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157404.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:41:31,085 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157405.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:41:35,311 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0922, 2.4024, 2.3166, 2.4285, 2.1991, 2.3150, 2.3587, 2.2736], + device='cuda:2'), covar=tensor([0.3672, 0.5633, 0.4583, 0.4212, 0.5774, 0.6772, 0.5726, 0.5554], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0376, 0.0329, 0.0342, 0.0351, 0.0393, 0.0360, 0.0335], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 02:41:36,435 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4872, 1.3275, 4.1760, 3.9820, 3.6282, 4.0134, 3.9789, 3.6820], + device='cuda:2'), covar=tensor([0.7186, 0.5563, 0.1068, 0.1526, 0.1063, 0.1519, 0.1119, 0.1679], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0310, 0.0408, 0.0408, 0.0349, 0.0417, 0.0319, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:41:38,341 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157417.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:42:02,291 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 2800, loss[loss=0.2062, simple_loss=0.2795, pruned_loss=0.0664, over 4829.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2358, pruned_loss=0.04564, over 956727.90 frames. ], batch size: 40, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:42:13,772 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157453.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:42:26,567 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157465.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 02:42:38,119 INFO [optim.py:369] (2/7) 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 02:42:46,873 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7843, 3.9554, 0.9673, 2.2486, 2.3108, 2.7372, 2.3782, 1.0440], + device='cuda:2'), covar=tensor([0.1312, 0.0736, 0.2043, 0.1097, 0.1017, 0.0990, 0.1266, 0.2110], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0238, 0.0135, 0.0120, 0.0131, 0.0152, 0.0116, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 02:42:49,355 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-04-28 02:43:05,501 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9632, 1.2188, 4.8215, 4.5041, 4.1963, 4.6013, 4.2914, 4.3113], + device='cuda:2'), covar=tensor([0.7015, 0.6377, 0.0978, 0.2020, 0.1118, 0.1976, 0.1989, 0.1626], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0309, 0.0406, 0.0406, 0.0348, 0.0415, 0.0318, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:43:06,201 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7666, 1.4802, 1.3185, 1.5591, 1.9477, 1.6347, 1.4462, 1.2719], + device='cuda:2'), covar=tensor([0.1489, 0.1506, 0.1865, 0.1339, 0.1000, 0.1547, 0.2113, 0.2208], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0306, 0.0351, 0.0285, 0.0324, 0.0306, 0.0299, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.4296e-05, 6.2790e-05, 7.3489e-05, 5.7007e-05, 6.6206e-05, 6.3588e-05, + 6.1805e-05, 7.9327e-05], device='cuda:2') +2023-04-28 02:43:07,286 INFO [finetune.py:976] (2/7) Epoch 28, batch 2850, loss[loss=0.1525, simple_loss=0.2362, pruned_loss=0.03439, over 4938.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2349, pruned_loss=0.04519, over 954955.34 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:43:07,365 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157498.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:43:17,205 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157506.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:43:40,383 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3902, 3.3770, 2.4810, 3.9454, 3.4607, 3.4318, 1.4543, 3.3442], + device='cuda:2'), covar=tensor([0.1842, 0.1416, 0.3097, 0.2102, 0.3465, 0.1865, 0.5971, 0.2670], + device='cuda:2'), in_proj_covar=tensor([0.0246, 0.0220, 0.0252, 0.0304, 0.0299, 0.0248, 0.0275, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 02:43:47,281 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9315, 2.7502, 1.0283, 1.2818, 1.9632, 1.1366, 3.5294, 1.6737], + device='cuda:2'), covar=tensor([0.0947, 0.0736, 0.0998, 0.1737, 0.0652, 0.1408, 0.0551, 0.0922], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 02:44:00,022 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 2900, loss[loss=0.179, simple_loss=0.2586, pruned_loss=0.04969, over 4913.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2391, pruned_loss=0.04637, over 952385.79 frames. ], batch size: 37, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:44:23,216 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8416, 2.1602, 1.7484, 1.4436, 1.3357, 1.3793, 1.7372, 1.2840], + device='cuda:2'), covar=tensor([0.1662, 0.1371, 0.1409, 0.1694, 0.2287, 0.1951, 0.1057, 0.2036], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0210, 0.0170, 0.0205, 0.0201, 0.0187, 0.0157, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 02:44:33,682 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157567.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:44:34,312 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:44:52,340 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.30 vs. limit=5.0 +2023-04-28 02:45:03,638 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 2950, loss[loss=0.1723, simple_loss=0.2522, pruned_loss=0.04616, over 4822.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2405, pruned_loss=0.04629, over 952213.36 frames. ], batch size: 51, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:45:36,881 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-28 02:45:50,079 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157629.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:46:18,330 INFO [finetune.py:976] (2/7) Epoch 28, batch 3000, loss[loss=0.1678, simple_loss=0.2397, pruned_loss=0.04794, over 4898.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2419, pruned_loss=0.04688, over 953959.36 frames. ], batch size: 32, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:46:18,330 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 02:46:21,917 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7816, 1.0784, 1.7353, 2.2548, 1.8575, 1.7096, 1.7148, 1.6833], + device='cuda:2'), covar=tensor([0.4510, 0.7207, 0.6526, 0.5559, 0.6064, 0.8352, 0.8320, 0.8768], + device='cuda:2'), in_proj_covar=tensor([0.0447, 0.0425, 0.0522, 0.0509, 0.0475, 0.0512, 0.0513, 0.0527], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 02:46:34,885 INFO [finetune.py:1010] (2/7) Epoch 28, validation: loss=0.153, simple_loss=0.2217, pruned_loss=0.04213, over 2265189.00 frames. +2023-04-28 02:46:34,885 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6415MB +2023-04-28 02:46:49,099 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157666.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:46:55,059 INFO [optim.py:369] (2/7) 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 02:47:09,008 INFO [finetune.py:976] (2/7) Epoch 28, batch 3050, loss[loss=0.1567, simple_loss=0.2408, pruned_loss=0.03624, over 4747.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2438, pruned_loss=0.04776, over 955189.68 frames. ], batch size: 27, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:47:27,241 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157714.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:47:41,862 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2999, 1.5205, 1.9899, 2.3250, 2.1061, 1.6632, 1.2979, 1.7048], + device='cuda:2'), covar=tensor([0.2888, 0.3405, 0.1707, 0.2133, 0.2501, 0.2577, 0.3983, 0.1935], + device='cuda:2'), in_proj_covar=tensor([0.0289, 0.0244, 0.0227, 0.0313, 0.0221, 0.0234, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 02:47:58,439 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 3100, loss[loss=0.1676, simple_loss=0.2397, pruned_loss=0.04777, over 4874.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2416, pruned_loss=0.04751, over 954319.01 frames. ], batch size: 34, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:48:22,605 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157760.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:48:32,272 INFO [optim.py:369] (2/7) 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 02:48:40,380 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-28 02:48:46,288 INFO [finetune.py:976] (2/7) Epoch 28, batch 3150, loss[loss=0.1319, simple_loss=0.2152, pruned_loss=0.02424, over 4816.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2385, pruned_loss=0.04636, over 953099.00 frames. ], batch size: 25, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:48:46,381 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157798.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:49:18,160 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 3200, loss[loss=0.1222, simple_loss=0.1922, pruned_loss=0.02606, over 4460.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.236, pruned_loss=0.04571, over 954416.36 frames. ], batch size: 19, lr: 2.88e-03, grad_scale: 32.0 +2023-04-28 02:49:22,550 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157853.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:49:22,560 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1559, 0.7910, 0.9185, 0.8227, 1.2620, 1.0266, 0.9608, 0.9815], + device='cuda:2'), covar=tensor([0.1653, 0.1546, 0.2359, 0.1566, 0.1143, 0.1486, 0.1509, 0.2380], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0304, 0.0348, 0.0284, 0.0322, 0.0304, 0.0298, 0.0373], + device='cuda:2'), out_proj_covar=tensor([6.3845e-05, 6.2379e-05, 7.3045e-05, 5.6829e-05, 6.5642e-05, 6.3225e-05, + 6.1422e-05, 7.8825e-05], device='cuda:2') +2023-04-28 02:49:27,981 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157862.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:49:28,631 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4594, 3.0282, 0.9521, 1.7476, 1.7899, 2.2287, 1.7440, 1.0273], + device='cuda:2'), covar=tensor([0.1321, 0.1022, 0.1831, 0.1224, 0.1007, 0.0906, 0.1547, 0.1803], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0238, 0.0134, 0.0120, 0.0131, 0.0152, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 02:49:38,913 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 28, batch 3250, loss[loss=0.1992, simple_loss=0.278, pruned_loss=0.06023, over 4854.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2364, pruned_loss=0.04576, over 952357.61 frames. ], batch size: 44, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:50:29,717 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157914.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:50:43,190 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157924.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:51:14,742 INFO [finetune.py:976] (2/7) Epoch 28, batch 3300, loss[loss=0.1914, simple_loss=0.2772, pruned_loss=0.05282, over 4821.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2384, pruned_loss=0.04593, over 951993.65 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:51:51,135 INFO [optim.py:369] (2/7) 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 02:52:20,397 INFO [finetune.py:976] (2/7) Epoch 28, batch 3350, loss[loss=0.2144, simple_loss=0.2737, pruned_loss=0.07761, over 4761.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.241, pruned_loss=0.0466, over 952934.88 frames. ], batch size: 28, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:52:41,453 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158012.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:53:12,156 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158038.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:53:21,807 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-04-28 02:53:23,377 INFO [finetune.py:976] (2/7) Epoch 28, batch 3400, loss[loss=0.16, simple_loss=0.2476, pruned_loss=0.03623, over 4838.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2415, pruned_loss=0.04672, over 948916.91 frames. ], batch size: 49, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:53:41,205 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158060.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:53:41,230 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158060.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:53:56,199 INFO [optim.py:369] (2/7) 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 02:54:10,631 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158086.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:54:23,034 INFO [finetune.py:976] (2/7) Epoch 28, batch 3450, loss[loss=0.1803, simple_loss=0.2583, pruned_loss=0.05116, over 4897.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2422, pruned_loss=0.04686, over 950842.60 frames. ], batch size: 37, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:54:34,534 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158108.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 02:54:54,852 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 3500, loss[loss=0.143, simple_loss=0.2119, pruned_loss=0.03706, over 4826.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2392, pruned_loss=0.04629, over 951923.18 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:55:48,557 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158162.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:56:07,655 INFO [optim.py:369] (2/7) 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 02:56:12,592 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158183.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:56:23,134 INFO [finetune.py:976] (2/7) Epoch 28, batch 3550, loss[loss=0.1341, simple_loss=0.2065, pruned_loss=0.03083, over 4729.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2365, pruned_loss=0.04581, over 953957.38 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:56:27,780 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4885, 1.1627, 1.3108, 1.1829, 1.6275, 1.3305, 1.1000, 1.2782], + device='cuda:2'), covar=tensor([0.1604, 0.1228, 0.1621, 0.1365, 0.0749, 0.1414, 0.1803, 0.2118], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0307, 0.0350, 0.0286, 0.0324, 0.0306, 0.0300, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.4214e-05, 6.2984e-05, 7.3369e-05, 5.7160e-05, 6.6151e-05, 6.3634e-05, + 6.2003e-05, 7.9269e-05], device='cuda:2') +2023-04-28 02:56:30,195 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158210.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 02:56:38,624 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5034, 3.1229, 2.6106, 2.9642, 2.1505, 2.7580, 2.7056, 2.2271], + device='cuda:2'), covar=tensor([0.1759, 0.1292, 0.0715, 0.0945, 0.3010, 0.0892, 0.1637, 0.2312], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0302, 0.0217, 0.0276, 0.0314, 0.0253, 0.0247, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1222e-04, 1.1892e-04, 8.5473e-05, 1.0869e-04, 1.2667e-04, 9.9594e-05, + 9.9622e-05, 1.0344e-04], device='cuda:2') +2023-04-28 02:56:39,838 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158224.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:56:42,238 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1295, 2.7857, 1.0806, 1.4693, 2.2957, 1.3154, 3.7093, 2.1313], + device='cuda:2'), covar=tensor([0.0662, 0.0549, 0.0800, 0.1266, 0.0457, 0.0966, 0.0234, 0.0503], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0045, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 02:56:56,634 INFO [finetune.py:976] (2/7) Epoch 28, batch 3600, loss[loss=0.1172, simple_loss=0.1926, pruned_loss=0.02091, over 4748.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2337, pruned_loss=0.04463, over 955319.07 frames. ], batch size: 27, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:57:11,757 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 02:57:20,498 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-28 02:57:29,419 INFO [finetune.py:976] (2/7) Epoch 28, batch 3650, loss[loss=0.2035, simple_loss=0.2565, pruned_loss=0.07522, over 4740.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2356, pruned_loss=0.04562, over 955580.96 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:57:37,486 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8672, 2.4046, 1.0672, 1.3181, 1.8762, 1.1517, 2.9725, 1.5970], + device='cuda:2'), covar=tensor([0.0714, 0.0528, 0.0675, 0.1215, 0.0476, 0.1033, 0.0306, 0.0588], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 02:58:02,904 INFO [finetune.py:976] (2/7) Epoch 28, batch 3700, loss[loss=0.1939, simple_loss=0.2644, pruned_loss=0.06176, over 4894.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2383, pruned_loss=0.04582, over 955887.01 frames. ], batch size: 35, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:58:19,868 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158384.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:58:31,280 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1634, 2.7800, 1.2524, 1.4540, 2.2825, 1.3687, 3.7619, 2.0203], + device='cuda:2'), covar=tensor([0.0646, 0.0640, 0.0692, 0.1307, 0.0446, 0.0967, 0.0260, 0.0546], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 02:58:35,209 INFO [finetune.py:976] (2/7) Epoch 28, batch 3750, loss[loss=0.1729, simple_loss=0.2542, pruned_loss=0.0458, over 4846.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2395, pruned_loss=0.046, over 955285.71 frames. ], batch size: 44, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:59:05,536 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158445.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 02:59:07,664 INFO [finetune.py:976] (2/7) Epoch 28, batch 3800, loss[loss=0.1691, simple_loss=0.2557, pruned_loss=0.0413, over 4922.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2417, pruned_loss=0.04661, over 955280.30 frames. ], batch size: 41, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 02:59:42,092 INFO [optim.py:369] (2/7) 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 02:59:43,422 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158478.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:00:07,895 INFO [finetune.py:976] (2/7) Epoch 28, batch 3850, loss[loss=0.1489, simple_loss=0.2173, pruned_loss=0.04022, over 4778.00 frames. ], tot_loss[loss=0.167, simple_loss=0.241, pruned_loss=0.04649, over 954365.64 frames. ], batch size: 26, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:00:25,642 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4663, 1.7097, 1.6054, 1.9380, 1.8498, 1.9865, 1.5350, 3.7009], + device='cuda:2'), covar=tensor([0.0580, 0.0859, 0.0820, 0.1260, 0.0648, 0.0517, 0.0787, 0.0196], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 03:00:25,765 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-04-28 03:00:26,827 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158509.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:01:10,983 INFO [finetune.py:976] (2/7) Epoch 28, batch 3900, loss[loss=0.1564, simple_loss=0.2291, pruned_loss=0.04182, over 4824.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2372, pruned_loss=0.04531, over 956172.10 frames. ], batch size: 40, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:01:28,079 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158557.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:01:51,821 INFO [optim.py:369] (2/7) 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 03:02:23,133 INFO [finetune.py:976] (2/7) Epoch 28, batch 3950, loss[loss=0.1548, simple_loss=0.21, pruned_loss=0.04984, over 4173.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2343, pruned_loss=0.04445, over 954004.64 frames. ], batch size: 18, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:02:36,357 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.0084, 2.3889, 2.2504, 2.3751, 2.1873, 2.3286, 2.3722, 2.2986], + device='cuda:2'), covar=tensor([0.3699, 0.5776, 0.4666, 0.4493, 0.5713, 0.6830, 0.5225, 0.5046], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0376, 0.0330, 0.0342, 0.0350, 0.0392, 0.0361, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 03:02:48,082 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9515, 2.7127, 1.2382, 1.4094, 2.0934, 1.1863, 3.5126, 1.8528], + device='cuda:2'), covar=tensor([0.0745, 0.0545, 0.0746, 0.1256, 0.0505, 0.1096, 0.0278, 0.0605], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 03:02:58,282 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8892, 1.1478, 5.0745, 4.7564, 4.3841, 4.9110, 4.5084, 4.4466], + device='cuda:2'), covar=tensor([0.7112, 0.6609, 0.0929, 0.1758, 0.1144, 0.1315, 0.1552, 0.1551], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0310, 0.0406, 0.0407, 0.0350, 0.0416, 0.0318, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:03:31,309 INFO [finetune.py:976] (2/7) Epoch 28, batch 4000, loss[loss=0.151, simple_loss=0.2205, pruned_loss=0.04075, over 4914.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2337, pruned_loss=0.04429, over 955438.64 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:04:13,586 INFO [optim.py:369] (2/7) 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 03:04:28,415 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.09 vs. limit=5.0 +2023-04-28 03:04:38,150 INFO [finetune.py:976] (2/7) Epoch 28, batch 4050, loss[loss=0.1794, simple_loss=0.2712, pruned_loss=0.04383, over 4843.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2378, pruned_loss=0.04573, over 954696.26 frames. ], batch size: 49, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:05:33,914 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158740.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:05:43,858 INFO [finetune.py:976] (2/7) Epoch 28, batch 4100, loss[loss=0.1595, simple_loss=0.2399, pruned_loss=0.03955, over 4840.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2405, pruned_loss=0.0465, over 954604.74 frames. ], batch size: 47, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:06:05,224 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.8527, 4.9438, 3.1807, 5.6096, 4.8919, 4.8465, 2.4336, 4.8465], + device='cuda:2'), covar=tensor([0.1638, 0.0906, 0.2977, 0.0795, 0.3072, 0.1542, 0.5448, 0.2039], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0221, 0.0253, 0.0305, 0.0302, 0.0251, 0.0278, 0.0276], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 03:06:25,665 INFO [optim.py:369] (2/7) 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 03:06:26,409 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158777.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:06:26,549 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-04-28 03:06:27,002 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158778.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:06:35,227 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158783.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:06:37,067 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5573, 1.1628, 1.2895, 1.2140, 1.7006, 1.3764, 1.1498, 1.2388], + device='cuda:2'), covar=tensor([0.1751, 0.1451, 0.2053, 0.1555, 0.0924, 0.1544, 0.1919, 0.2693], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0306, 0.0349, 0.0285, 0.0324, 0.0304, 0.0299, 0.0374], + device='cuda:2'), 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:2') +2023-04-28 03:06:49,758 INFO [finetune.py:976] (2/7) Epoch 28, batch 4150, loss[loss=0.1988, simple_loss=0.2724, pruned_loss=0.06257, over 4808.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2422, pruned_loss=0.04715, over 955113.27 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 64.0 +2023-04-28 03:07:30,858 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158826.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:07:41,990 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158835.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:07:43,883 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158838.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:07:52,633 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 4200, loss[loss=0.1476, simple_loss=0.2304, pruned_loss=0.03244, over 4753.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2426, pruned_loss=0.04722, over 955529.20 frames. ], batch size: 28, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:08:37,325 INFO [optim.py:369] (2/7) 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 03:08:54,581 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7046, 1.7468, 0.8528, 1.4154, 1.8792, 1.5960, 1.4662, 1.5397], + device='cuda:2'), covar=tensor([0.0465, 0.0352, 0.0330, 0.0546, 0.0272, 0.0480, 0.0484, 0.0549], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 03:08:56,476 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8437, 2.2766, 2.2752, 2.3874, 2.3411, 2.4226, 2.3655, 2.3332], + device='cuda:2'), covar=tensor([0.3086, 0.4563, 0.4382, 0.4038, 0.4562, 0.6016, 0.4710, 0.4600], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0376, 0.0329, 0.0342, 0.0351, 0.0392, 0.0361, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 03:09:04,984 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158896.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 03:09:06,081 INFO [finetune.py:976] (2/7) Epoch 28, batch 4250, loss[loss=0.1844, simple_loss=0.2551, pruned_loss=0.05678, over 4869.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2393, pruned_loss=0.04597, over 955571.25 frames. ], batch size: 34, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:09:15,890 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9672, 1.4902, 2.0458, 2.3810, 2.0216, 1.9181, 2.0070, 1.8794], + device='cuda:2'), covar=tensor([0.4541, 0.6785, 0.6333, 0.5414, 0.5636, 0.7732, 0.7637, 0.9474], + device='cuda:2'), in_proj_covar=tensor([0.0446, 0.0425, 0.0522, 0.0510, 0.0475, 0.0511, 0.0513, 0.0528], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:10:12,868 INFO [finetune.py:976] (2/7) Epoch 28, batch 4300, loss[loss=0.1537, simple_loss=0.2209, pruned_loss=0.04324, over 4902.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2358, pruned_loss=0.04477, over 956512.65 frames. ], batch size: 32, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:10:48,762 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 28, batch 4350, loss[loss=0.1878, simple_loss=0.2528, pruned_loss=0.0614, over 4853.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2343, pruned_loss=0.04443, over 955786.85 frames. ], batch size: 49, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:12:15,670 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 4400, loss[loss=0.1981, simple_loss=0.2774, pruned_loss=0.05942, over 4809.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2345, pruned_loss=0.04449, over 956927.97 frames. ], batch size: 51, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:12:54,442 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159069.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:13:05,082 INFO [optim.py:369] (2/7) 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 03:13:19,036 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159088.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:13:29,730 INFO [finetune.py:976] (2/7) Epoch 28, batch 4450, loss[loss=0.1793, simple_loss=0.2436, pruned_loss=0.05754, over 4825.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2383, pruned_loss=0.04534, over 956434.38 frames. ], batch size: 30, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:14:09,958 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159130.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:14:12,212 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159133.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:14:21,841 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159139.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:14:38,431 INFO [finetune.py:976] (2/7) Epoch 28, batch 4500, loss[loss=0.1171, simple_loss=0.1844, pruned_loss=0.02495, over 4261.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2403, pruned_loss=0.04608, over 954276.04 frames. ], batch size: 18, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:14:49,773 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159159.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:15:03,532 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-28 03:15:11,103 INFO [optim.py:369] (2/7) 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 03:15:31,027 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159191.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 03:15:41,715 INFO [finetune.py:976] (2/7) Epoch 28, batch 4550, loss[loss=0.1884, simple_loss=0.2599, pruned_loss=0.05846, over 4910.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2408, pruned_loss=0.04594, over 955978.31 frames. ], batch size: 37, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:16:05,613 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159220.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:16:25,495 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0987, 1.8239, 2.0930, 2.4180, 2.4740, 2.0488, 1.8007, 2.0472], + device='cuda:2'), covar=tensor([0.0832, 0.1167, 0.0691, 0.0583, 0.0570, 0.0835, 0.0684, 0.0551], + device='cuda:2'), in_proj_covar=tensor([0.0182, 0.0202, 0.0182, 0.0169, 0.0177, 0.0176, 0.0149, 0.0174], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:16:45,688 INFO [finetune.py:976] (2/7) Epoch 28, batch 4600, loss[loss=0.1812, simple_loss=0.2543, pruned_loss=0.05408, over 4909.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2395, pruned_loss=0.04502, over 954308.81 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:17:18,846 INFO [optim.py:369] (2/7) 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 03:17:29,640 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2716, 1.5017, 1.3691, 1.6875, 1.6299, 1.7253, 1.4318, 3.0865], + device='cuda:2'), covar=tensor([0.0625, 0.0812, 0.0819, 0.1247, 0.0658, 0.0481, 0.0742, 0.0180], + device='cuda:2'), in_proj_covar=tensor([0.0038, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 03:17:38,415 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159290.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:17:49,022 INFO [finetune.py:976] (2/7) Epoch 28, batch 4650, loss[loss=0.1696, simple_loss=0.2384, pruned_loss=0.05041, over 4250.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2378, pruned_loss=0.04568, over 951825.73 frames. ], batch size: 65, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:18:20,349 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0883, 1.3689, 1.2967, 1.6276, 1.5159, 1.5739, 1.3408, 2.4035], + device='cuda:2'), covar=tensor([0.0632, 0.0847, 0.0849, 0.1286, 0.0643, 0.0497, 0.0756, 0.0201], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0055], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 03:18:21,592 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5159, 3.1789, 2.6723, 3.0027, 2.2768, 2.6347, 2.8489, 2.1915], + device='cuda:2'), covar=tensor([0.1919, 0.1218, 0.0762, 0.1178, 0.3234, 0.1145, 0.2072, 0.2536], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0300, 0.0217, 0.0275, 0.0313, 0.0251, 0.0246, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1195e-04, 1.1818e-04, 8.5235e-05, 1.0808e-04, 1.2620e-04, 9.8737e-05, + 9.9064e-05, 1.0297e-04], device='cuda:2') +2023-04-28 03:18:30,832 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0187, 1.0512, 1.2353, 1.2048, 1.0401, 0.8930, 0.9254, 0.3840], + device='cuda:2'), covar=tensor([0.0602, 0.0527, 0.0495, 0.0508, 0.0699, 0.1412, 0.0552, 0.0726], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:18:54,507 INFO [finetune.py:976] (2/7) Epoch 28, batch 4700, loss[loss=0.1215, simple_loss=0.1949, pruned_loss=0.02405, over 4826.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.235, pruned_loss=0.04499, over 950556.60 frames. ], batch size: 39, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:19:03,228 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159351.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:19:05,580 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6644, 1.4374, 1.8791, 1.9657, 1.4363, 1.3876, 1.5800, 0.9076], + device='cuda:2'), covar=tensor([0.0635, 0.0651, 0.0336, 0.0596, 0.0813, 0.1098, 0.0531, 0.0618], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:19:18,092 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-28 03:19:36,242 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 28, batch 4750, loss[loss=0.174, simple_loss=0.2499, pruned_loss=0.04903, over 4815.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2342, pruned_loss=0.04463, over 952375.66 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-04-28 03:20:41,285 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159425.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:20:51,584 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159433.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:20:55,253 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 4800, loss[loss=0.1414, simple_loss=0.2151, pruned_loss=0.03383, over 4713.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2371, pruned_loss=0.04621, over 950775.85 frames. ], batch size: 23, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:21:25,691 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2033, 2.0561, 1.6710, 1.6804, 2.0848, 1.7000, 2.4716, 1.5094], + device='cuda:2'), covar=tensor([0.3716, 0.1922, 0.4861, 0.3026, 0.1805, 0.2390, 0.1589, 0.4521], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0355, 0.0422, 0.0353, 0.0383, 0.0376, 0.0372, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:21:48,628 INFO [optim.py:369] (2/7) 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 03:21:51,196 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159481.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:21:59,479 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159487.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:22:07,522 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159491.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 03:22:09,407 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0185, 0.9845, 1.2349, 1.1885, 0.9948, 0.9248, 0.9598, 0.5020], + device='cuda:2'), covar=tensor([0.0501, 0.0660, 0.0437, 0.0492, 0.0644, 0.1104, 0.0430, 0.0621], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:22:12,298 INFO [finetune.py:976] (2/7) Epoch 28, batch 4850, loss[loss=0.1675, simple_loss=0.2479, pruned_loss=0.04353, over 4764.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2409, pruned_loss=0.04712, over 949403.21 frames. ], batch size: 28, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:22:40,593 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159515.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:23:05,856 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 4900, loss[loss=0.2301, simple_loss=0.2937, pruned_loss=0.08326, over 4824.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2433, pruned_loss=0.04812, over 949189.13 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:23:56,948 INFO [optim.py:369] (2/7) 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 03:24:16,155 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-04-28 03:24:20,243 INFO [finetune.py:976] (2/7) Epoch 28, batch 4950, loss[loss=0.1875, simple_loss=0.2611, pruned_loss=0.05691, over 4745.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2432, pruned_loss=0.04755, over 950542.43 frames. ], batch size: 59, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:25:08,548 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159631.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:25:23,011 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159646.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:25:24,155 INFO [finetune.py:976] (2/7) Epoch 28, batch 5000, loss[loss=0.1546, simple_loss=0.2336, pruned_loss=0.03782, over 4869.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2417, pruned_loss=0.04726, over 950465.70 frames. ], batch size: 34, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:26:05,240 INFO [optim.py:369] (2/7) 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 03:26:24,449 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159692.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:26:32,513 INFO [finetune.py:976] (2/7) Epoch 28, batch 5050, loss[loss=0.1439, simple_loss=0.2123, pruned_loss=0.03769, over 4825.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.238, pruned_loss=0.04589, over 952195.13 frames. ], batch size: 30, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:26:37,053 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-04-28 03:27:06,795 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159725.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:27:18,063 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4443, 3.4353, 0.8116, 1.8919, 1.7780, 2.3838, 1.8773, 1.0477], + device='cuda:2'), covar=tensor([0.1370, 0.0849, 0.2087, 0.1149, 0.1057, 0.0973, 0.1470, 0.1901], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0240, 0.0137, 0.0121, 0.0133, 0.0154, 0.0118, 0.0120], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:27:35,600 INFO [finetune.py:976] (2/7) Epoch 28, batch 5100, loss[loss=0.1481, simple_loss=0.2222, pruned_loss=0.03699, over 4753.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2349, pruned_loss=0.04444, over 950521.97 frames. ], batch size: 27, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:27:36,966 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8920, 1.7081, 1.9546, 2.3021, 2.2542, 1.8893, 1.6501, 2.1174], + device='cuda:2'), covar=tensor([0.0871, 0.1116, 0.0791, 0.0545, 0.0625, 0.0846, 0.0707, 0.0516], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0203, 0.0183, 0.0170, 0.0178, 0.0177, 0.0150, 0.0175], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:28:09,662 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 03:28:41,782 INFO [finetune.py:976] (2/7) Epoch 28, batch 5150, loss[loss=0.1694, simple_loss=0.2489, pruned_loss=0.04496, over 4902.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2352, pruned_loss=0.0446, over 950578.86 frames. ], batch size: 35, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:29:03,525 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159815.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:29:16,440 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.7517, 1.7921, 1.5911, 1.3524, 1.7605, 1.4117, 2.1430, 1.3217], + device='cuda:2'), covar=tensor([0.3603, 0.1754, 0.4803, 0.2783, 0.1703, 0.2429, 0.1678, 0.4984], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0352, 0.0420, 0.0350, 0.0381, 0.0374, 0.0370, 0.0422], + device='cuda:2'), out_proj_covar=tensor([9.9492e-05, 1.0495e-04, 1.2707e-04, 1.0474e-04, 1.1260e-04, 1.1103e-04, + 1.0800e-04, 1.2676e-04], device='cuda:2') +2023-04-28 03:29:25,233 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1320, 1.6587, 1.4521, 2.0204, 2.2090, 1.8665, 1.7746, 1.5426], + device='cuda:2'), covar=tensor([0.1597, 0.1778, 0.1720, 0.1459, 0.1250, 0.1801, 0.2060, 0.2392], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0306, 0.0347, 0.0285, 0.0323, 0.0304, 0.0297, 0.0372], + device='cuda:2'), out_proj_covar=tensor([6.3942e-05, 6.2781e-05, 7.2621e-05, 5.7020e-05, 6.5770e-05, 6.3114e-05, + 6.1347e-05, 7.8575e-05], device='cuda:2') +2023-04-28 03:29:45,762 INFO [finetune.py:976] (2/7) Epoch 28, batch 5200, loss[loss=0.1394, simple_loss=0.2186, pruned_loss=0.03005, over 4356.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2379, pruned_loss=0.04552, over 951096.59 frames. ], batch size: 19, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:29:55,502 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159855.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:30:05,391 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159863.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:30:17,396 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7637, 1.4020, 1.8490, 2.3219, 1.8766, 1.7503, 1.8511, 1.7525], + device='cuda:2'), covar=tensor([0.4154, 0.6465, 0.6038, 0.5120, 0.5748, 0.7617, 0.7466, 0.9439], + device='cuda:2'), in_proj_covar=tensor([0.0447, 0.0426, 0.0522, 0.0510, 0.0475, 0.0512, 0.0514, 0.0529], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:30:26,727 INFO [optim.py:369] (2/7) 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 03:30:40,180 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.0053, 1.0340, 1.2282, 1.1600, 0.9881, 0.9789, 1.0214, 0.5178], + device='cuda:2'), covar=tensor([0.0562, 0.0543, 0.0393, 0.0514, 0.0756, 0.1049, 0.0462, 0.0586], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:30:51,959 INFO [finetune.py:976] (2/7) Epoch 28, batch 5250, loss[loss=0.1478, simple_loss=0.2185, pruned_loss=0.03851, over 4808.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2396, pruned_loss=0.04586, over 952927.29 frames. ], batch size: 25, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:31:09,955 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159910.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:31:19,006 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159916.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:31:55,434 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 28, batch 5300, loss[loss=0.1578, simple_loss=0.234, pruned_loss=0.0408, over 4705.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2412, pruned_loss=0.04679, over 952292.71 frames. ], batch size: 59, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:32:03,389 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3944, 1.6527, 5.4602, 5.1405, 4.7404, 5.2607, 4.8344, 4.8789], + device='cuda:2'), covar=tensor([0.5914, 0.5552, 0.0975, 0.1610, 0.0934, 0.1281, 0.0993, 0.1477], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0310, 0.0405, 0.0409, 0.0351, 0.0416, 0.0319, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:32:06,528 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 03:32:26,781 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159971.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:32:35,445 INFO [optim.py:369] (2/7) 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 03:32:46,960 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9643, 1.6797, 1.9016, 2.0083, 1.9212, 1.5762, 1.1901, 1.6843], + device='cuda:2'), covar=tensor([0.2783, 0.2449, 0.1498, 0.1830, 0.1974, 0.2256, 0.3406, 0.1611], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0245, 0.0227, 0.0313, 0.0222, 0.0234, 0.0228, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 03:32:48,729 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159987.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:32:48,880 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.94 vs. limit=5.0 +2023-04-28 03:32:58,790 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159994.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:33:06,690 INFO [finetune.py:976] (2/7) Epoch 28, batch 5350, loss[loss=0.1606, simple_loss=0.2374, pruned_loss=0.04188, over 4816.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.242, pruned_loss=0.04655, over 954231.09 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:33:11,415 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.02 vs. limit=5.0 +2023-04-28 03:34:00,196 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6265, 2.2174, 2.5408, 3.0270, 2.4123, 1.9552, 1.9373, 2.3891], + device='cuda:2'), covar=tensor([0.3104, 0.2800, 0.1540, 0.2112, 0.2617, 0.2566, 0.3508, 0.1921], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0244, 0.0226, 0.0312, 0.0222, 0.0234, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 03:34:13,400 INFO [finetune.py:976] (2/7) Epoch 28, batch 5400, loss[loss=0.1337, simple_loss=0.2048, pruned_loss=0.03133, over 4773.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2401, pruned_loss=0.04621, over 955752.36 frames. ], batch size: 28, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:34:46,552 INFO [optim.py:369] (2/7) 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 03:35:18,037 INFO [finetune.py:976] (2/7) Epoch 28, batch 5450, loss[loss=0.1527, simple_loss=0.2108, pruned_loss=0.04726, over 4285.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2371, pruned_loss=0.04551, over 956111.96 frames. ], batch size: 18, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:36:22,771 INFO [finetune.py:976] (2/7) Epoch 28, batch 5500, loss[loss=0.1886, simple_loss=0.2563, pruned_loss=0.06048, over 4910.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2348, pruned_loss=0.04485, over 957522.70 frames. ], batch size: 35, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:36:31,862 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.03 vs. limit=5.0 +2023-04-28 03:37:00,971 INFO [optim.py:369] (2/7) 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 03:37:04,144 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6489, 1.1402, 4.0786, 3.8302, 3.5072, 3.7776, 3.7304, 3.5762], + device='cuda:2'), covar=tensor([0.6968, 0.6069, 0.1029, 0.1667, 0.1207, 0.1835, 0.2459, 0.1595], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0310, 0.0406, 0.0408, 0.0351, 0.0416, 0.0319, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:37:27,649 INFO [finetune.py:976] (2/7) Epoch 28, batch 5550, loss[loss=0.1366, simple_loss=0.2073, pruned_loss=0.03293, over 4724.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2362, pruned_loss=0.04545, over 958135.56 frames. ], batch size: 23, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:37:41,112 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160211.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:38:03,241 INFO [finetune.py:976] (2/7) Epoch 28, batch 5600, loss[loss=0.1617, simple_loss=0.2488, pruned_loss=0.03733, over 4909.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2394, pruned_loss=0.04585, over 957930.73 frames. ], batch size: 43, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:38:07,420 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4002, 1.2695, 1.6232, 1.6122, 1.3030, 1.1599, 1.1551, 0.6444], + device='cuda:2'), covar=tensor([0.0488, 0.0592, 0.0355, 0.0489, 0.0677, 0.1284, 0.0598, 0.0629], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:38:13,710 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160266.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:38:14,360 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7366, 2.0371, 1.6844, 1.4409, 1.2960, 1.3065, 1.7060, 1.1927], + device='cuda:2'), covar=tensor([0.1662, 0.1307, 0.1497, 0.1691, 0.2260, 0.1917, 0.1021, 0.2104], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0209, 0.0170, 0.0205, 0.0201, 0.0186, 0.0157, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 03:38:20,147 INFO [optim.py:369] (2/7) 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 03:38:26,528 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160287.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:38:32,856 INFO [finetune.py:976] (2/7) Epoch 28, batch 5650, loss[loss=0.1757, simple_loss=0.2502, pruned_loss=0.05059, over 4936.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.242, pruned_loss=0.04642, over 957168.74 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:38:37,748 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-04-28 03:38:38,223 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160307.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:38:54,599 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-04-28 03:38:54,998 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160335.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:39:02,715 INFO [finetune.py:976] (2/7) Epoch 28, batch 5700, loss[loss=0.1655, simple_loss=0.2153, pruned_loss=0.05788, over 3838.00 frames. ], tot_loss[loss=0.165, simple_loss=0.238, pruned_loss=0.04604, over 936340.06 frames. ], batch size: 16, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:39:14,620 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160368.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 03:39:31,531 INFO [finetune.py:976] (2/7) Epoch 29, batch 0, loss[loss=0.1906, simple_loss=0.2606, pruned_loss=0.06029, over 4836.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2606, pruned_loss=0.06029, over 4836.00 frames. ], batch size: 49, lr: 2.86e-03, grad_scale: 32.0 +2023-04-28 03:39:31,531 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 03:39:42,779 INFO [finetune.py:1010] (2/7) Epoch 29, validation: loss=0.1546, simple_loss=0.2236, pruned_loss=0.04278, over 2265189.00 frames. +2023-04-28 03:39:42,779 INFO [finetune.py:1011] (2/7) Maximum memory allocated so far is 6415MB +2023-04-28 03:39:44,401 INFO [optim.py:369] (2/7) 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 03:39:45,706 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4702, 3.4092, 0.7973, 1.8568, 1.7966, 2.4667, 1.9757, 1.0519], + device='cuda:2'), covar=tensor([0.1428, 0.0780, 0.2021, 0.1182, 0.1055, 0.0941, 0.1342, 0.2055], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0120, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:40:16,112 INFO [finetune.py:976] (2/7) Epoch 29, batch 50, loss[loss=0.2191, simple_loss=0.2863, pruned_loss=0.07593, over 4253.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2469, pruned_loss=0.04977, over 216832.69 frames. ], batch size: 66, lr: 2.85e-03, grad_scale: 32.0 +2023-04-28 03:40:21,607 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 03:40:33,647 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-04-28 03:41:16,375 INFO [finetune.py:976] (2/7) Epoch 29, batch 100, loss[loss=0.1752, simple_loss=0.2473, pruned_loss=0.05158, over 4748.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2352, pruned_loss=0.04506, over 381435.31 frames. ], batch size: 54, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:41:24,293 INFO [optim.py:369] (2/7) 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 03:42:07,348 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 150, loss[loss=0.143, simple_loss=0.2179, pruned_loss=0.03401, over 4901.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2348, pruned_loss=0.04589, over 510976.18 frames. ], batch size: 37, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:42:43,462 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3243, 1.8085, 2.2022, 2.5860, 2.1737, 1.7305, 1.5021, 2.0140], + device='cuda:2'), covar=tensor([0.2772, 0.2908, 0.1488, 0.1969, 0.2393, 0.2458, 0.4090, 0.1808], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0247, 0.0228, 0.0315, 0.0223, 0.0236, 0.0229, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 03:43:05,314 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160566.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:43:27,102 INFO [finetune.py:976] (2/7) Epoch 29, batch 200, loss[loss=0.1474, simple_loss=0.2238, pruned_loss=0.03551, over 4899.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2311, pruned_loss=0.04437, over 610743.93 frames. ], batch size: 35, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:43:34,236 INFO [optim.py:369] (2/7) 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 03:43:37,297 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1588, 2.5323, 0.7931, 1.5342, 1.4764, 1.8582, 1.5700, 0.8668], + device='cuda:2'), covar=tensor([0.1447, 0.1073, 0.1815, 0.1245, 0.1153, 0.0966, 0.1566, 0.1813], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0135, 0.0119, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:44:17,862 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 250, loss[loss=0.2024, simple_loss=0.2722, pruned_loss=0.06634, over 4812.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2357, pruned_loss=0.04564, over 687830.97 frames. ], batch size: 39, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:45:20,093 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160663.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 03:45:32,745 INFO [finetune.py:976] (2/7) Epoch 29, batch 300, loss[loss=0.1411, simple_loss=0.2247, pruned_loss=0.02877, over 4873.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.24, pruned_loss=0.0468, over 747088.60 frames. ], batch size: 34, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:45:38,924 INFO [optim.py:369] (2/7) 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 03:46:01,445 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 03:46:15,263 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-04-28 03:46:37,392 INFO [finetune.py:976] (2/7) Epoch 29, batch 350, loss[loss=0.155, simple_loss=0.2267, pruned_loss=0.04166, over 4919.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2408, pruned_loss=0.04699, over 793581.21 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:47:08,176 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1094, 1.7054, 2.0035, 2.4769, 2.4808, 1.9610, 1.8500, 2.2683], + device='cuda:2'), covar=tensor([0.0776, 0.1227, 0.0715, 0.0529, 0.0546, 0.0828, 0.0651, 0.0510], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0205, 0.0184, 0.0171, 0.0178, 0.0177, 0.0151, 0.0176], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:47:35,585 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-28 03:47:45,907 INFO [finetune.py:976] (2/7) Epoch 29, batch 400, loss[loss=0.1269, simple_loss=0.2105, pruned_loss=0.02161, over 4776.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2421, pruned_loss=0.04691, over 829167.32 frames. ], batch size: 29, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:47:47,780 INFO [optim.py:369] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.9697, 1.6149, 1.5488, 1.8402, 2.1835, 1.8394, 1.5881, 1.4579], + device='cuda:2'), covar=tensor([0.1545, 0.1556, 0.2011, 0.1309, 0.0992, 0.1477, 0.1986, 0.2292], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0305, 0.0346, 0.0284, 0.0322, 0.0302, 0.0297, 0.0372], + device='cuda:2'), 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:2') +2023-04-28 03:48:36,948 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5892, 1.8045, 0.7857, 1.3301, 1.8102, 1.4908, 1.3924, 1.4584], + device='cuda:2'), covar=tensor([0.0484, 0.0351, 0.0328, 0.0552, 0.0263, 0.0509, 0.0477, 0.0557], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0028, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 03:48:48,103 INFO [finetune.py:976] (2/7) Epoch 29, batch 450, loss[loss=0.1893, simple_loss=0.2672, pruned_loss=0.05571, over 4927.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2404, pruned_loss=0.04622, over 855492.16 frames. ], batch size: 42, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:49:19,870 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.9072, 3.9429, 2.7660, 4.5716, 4.0410, 3.9702, 1.9589, 3.7643], + device='cuda:2'), covar=tensor([0.1559, 0.1101, 0.3058, 0.1336, 0.3088, 0.1605, 0.5598, 0.2483], + device='cuda:2'), in_proj_covar=tensor([0.0250, 0.0223, 0.0254, 0.0308, 0.0303, 0.0254, 0.0280, 0.0278], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 03:49:51,961 INFO [finetune.py:976] (2/7) Epoch 29, batch 500, loss[loss=0.1771, simple_loss=0.2475, pruned_loss=0.05335, over 4840.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2383, pruned_loss=0.04589, over 877880.85 frames. ], batch size: 44, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:49:53,843 INFO [optim.py:369] (2/7) 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 03:50:02,718 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-04-28 03:50:25,300 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160901.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:50:56,031 INFO [finetune.py:976] (2/7) Epoch 29, batch 550, loss[loss=0.1597, simple_loss=0.2373, pruned_loss=0.04108, over 4831.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2361, pruned_loss=0.04599, over 897844.30 frames. ], batch size: 40, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:51:48,583 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160962.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:51:49,189 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160963.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:51:50,433 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160965.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:51:51,661 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6928, 1.4843, 1.8311, 1.9007, 1.4876, 1.4420, 1.5036, 1.0311], + device='cuda:2'), covar=tensor([0.0475, 0.0583, 0.0412, 0.0456, 0.0634, 0.0963, 0.0494, 0.0574], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0066, 0.0065, 0.0068, 0.0074, 0.0093, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:52:01,358 INFO [finetune.py:976] (2/7) Epoch 29, batch 600, loss[loss=0.1703, simple_loss=0.2552, pruned_loss=0.04266, over 4901.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2363, pruned_loss=0.046, over 910694.27 frames. ], batch size: 43, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:52:03,146 INFO [optim.py:369] (2/7) 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 03:52:35,724 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161011.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 03:52:36,359 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1184, 2.8102, 2.3602, 2.6198, 2.0148, 2.3813, 2.3488, 1.8384], + device='cuda:2'), covar=tensor([0.2127, 0.1237, 0.0694, 0.1191, 0.3039, 0.0966, 0.1975, 0.2585], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0299, 0.0216, 0.0275, 0.0313, 0.0253, 0.0248, 0.0263], + device='cuda:2'), 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:2') +2023-04-28 03:52:38,801 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1198, 1.8711, 2.0262, 2.4176, 2.3946, 2.0585, 1.7696, 2.2465], + device='cuda:2'), covar=tensor([0.0706, 0.1114, 0.0762, 0.0561, 0.0551, 0.0773, 0.0710, 0.0491], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0203, 0.0183, 0.0171, 0.0177, 0.0177, 0.0150, 0.0176], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 03:52:44,112 INFO [finetune.py:976] (2/7) Epoch 29, batch 650, loss[loss=0.169, simple_loss=0.2383, pruned_loss=0.04988, over 4782.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2388, pruned_loss=0.04609, over 919527.79 frames. ], batch size: 25, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:52:44,870 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161026.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:52:51,769 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 03:53:17,107 INFO [finetune.py:976] (2/7) Epoch 29, batch 700, loss[loss=0.1776, simple_loss=0.2594, pruned_loss=0.04786, over 4891.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2403, pruned_loss=0.04682, over 928516.94 frames. ], batch size: 32, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:53:18,919 INFO [optim.py:369] (2/7) 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 03:53:50,550 INFO [finetune.py:976] (2/7) Epoch 29, batch 750, loss[loss=0.1655, simple_loss=0.2403, pruned_loss=0.04529, over 4827.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2426, pruned_loss=0.04798, over 935498.42 frames. ], batch size: 47, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:54:24,405 INFO [finetune.py:976] (2/7) Epoch 29, batch 800, loss[loss=0.1758, simple_loss=0.2565, pruned_loss=0.04752, over 4812.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2432, pruned_loss=0.0476, over 939312.15 frames. ], batch size: 41, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:54:26,204 INFO [optim.py:369] (2/7) 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 03:54:52,593 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161216.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:54:57,898 INFO [finetune.py:976] (2/7) Epoch 29, batch 850, loss[loss=0.2201, simple_loss=0.2777, pruned_loss=0.08125, over 4268.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.241, pruned_loss=0.04726, over 942084.76 frames. ], batch size: 65, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:54:58,674 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.6899, 2.3241, 2.5621, 3.2815, 2.5434, 2.0309, 2.2643, 2.6815], + device='cuda:2'), covar=tensor([0.3135, 0.2949, 0.1573, 0.2069, 0.2581, 0.2670, 0.3410, 0.1811], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0247, 0.0229, 0.0315, 0.0224, 0.0236, 0.0230, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 03:55:01,669 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161257.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:55:30,800 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-28 03:55:49,588 INFO [finetune.py:976] (2/7) Epoch 29, batch 900, loss[loss=0.1604, simple_loss=0.2235, pruned_loss=0.04869, over 4904.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.238, pruned_loss=0.04641, over 945527.67 frames. ], batch size: 32, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:55:50,927 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 03:56:00,185 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161292.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:56:21,349 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161321.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:56:23,750 INFO [finetune.py:976] (2/7) Epoch 29, batch 950, loss[loss=0.1259, simple_loss=0.197, pruned_loss=0.02735, over 4894.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2356, pruned_loss=0.04542, over 945208.34 frames. ], batch size: 32, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:56:35,381 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2946, 2.8957, 0.8436, 1.4814, 2.1667, 1.2336, 3.9101, 1.7791], + device='cuda:2'), covar=tensor([0.0692, 0.0921, 0.0908, 0.1245, 0.0518, 0.1013, 0.0190, 0.0624], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0065, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 03:56:38,096 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-04-28 03:57:07,673 INFO [finetune.py:976] (2/7) Epoch 29, batch 1000, loss[loss=0.1798, simple_loss=0.2597, pruned_loss=0.0499, over 4813.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2388, pruned_loss=0.04656, over 948029.83 frames. ], batch size: 45, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:57:09,492 INFO [optim.py:369] (2/7) 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] (2/7) attn_weights_entropy = tensor([1.7827, 1.2788, 1.4208, 1.4853, 1.9102, 1.5718, 1.3087, 1.4023], + device='cuda:2'), covar=tensor([0.1486, 0.1330, 0.1762, 0.1198, 0.0736, 0.1289, 0.1692, 0.2047], + device='cuda:2'), in_proj_covar=tensor([0.0316, 0.0309, 0.0351, 0.0288, 0.0328, 0.0307, 0.0301, 0.0377], + device='cuda:2'), 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:2') +2023-04-28 03:57:39,577 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161401.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:58:12,305 INFO [finetune.py:976] (2/7) Epoch 29, batch 1050, loss[loss=0.1935, simple_loss=0.2637, pruned_loss=0.06167, over 4806.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2419, pruned_loss=0.04708, over 950746.79 frames. ], batch size: 51, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:58:24,032 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8557, 1.7979, 2.3457, 2.4254, 1.6652, 1.5446, 1.8557, 1.1021], + device='cuda:2'), covar=tensor([0.0721, 0.0785, 0.0390, 0.0770, 0.0745, 0.1051, 0.0652, 0.0650], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0073, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 03:58:55,485 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161460.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 03:58:56,752 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 1100, loss[loss=0.1359, simple_loss=0.2161, pruned_loss=0.02781, over 4875.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2431, pruned_loss=0.04762, over 952545.06 frames. ], batch size: 35, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 03:59:20,410 INFO [optim.py:369] (2/7) 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 04:00:20,261 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 1150, loss[loss=0.1661, simple_loss=0.2309, pruned_loss=0.05064, over 4856.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2432, pruned_loss=0.04754, over 953553.75 frames. ], batch size: 31, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:00:23,848 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-04-28 04:00:47,321 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 04:01:04,609 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161557.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:01:25,593 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 1200, loss[loss=0.1322, simple_loss=0.2045, pruned_loss=0.03, over 4811.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.241, pruned_loss=0.04692, over 952327.46 frames. ], batch size: 51, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:01:29,679 INFO [optim.py:369] (2/7) 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 04:01:47,190 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161587.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:02:09,297 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161605.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:02:14,220 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1139, 2.1885, 2.2605, 2.7119, 2.7102, 2.2887, 1.9909, 2.4107], + device='cuda:2'), covar=tensor([0.0852, 0.0913, 0.0625, 0.0528, 0.0576, 0.0688, 0.0668, 0.0475], + device='cuda:2'), in_proj_covar=tensor([0.0183, 0.0202, 0.0183, 0.0170, 0.0177, 0.0176, 0.0150, 0.0175], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:02:23,987 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161621.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:02:33,464 INFO [finetune.py:976] (2/7) Epoch 29, batch 1250, loss[loss=0.1528, simple_loss=0.2155, pruned_loss=0.04507, over 4899.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2386, pruned_loss=0.04636, over 954500.17 frames. ], batch size: 32, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:03:05,219 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7846, 1.4893, 1.9031, 2.2978, 1.9231, 1.7539, 1.8186, 1.7903], + device='cuda:2'), covar=tensor([0.3881, 0.5968, 0.4975, 0.4800, 0.4907, 0.6920, 0.6999, 0.7794], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0427, 0.0522, 0.0511, 0.0475, 0.0514, 0.0515, 0.0530], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:03:28,495 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161669.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:03:28,559 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1025, 2.0099, 1.7339, 1.6599, 2.0824, 1.7898, 2.4306, 1.5507], + device='cuda:2'), covar=tensor([0.3169, 0.1529, 0.4145, 0.2449, 0.1435, 0.1856, 0.1316, 0.4049], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0356, 0.0424, 0.0353, 0.0386, 0.0377, 0.0373, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:03:28,570 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6286, 1.2290, 1.3057, 1.4264, 1.7826, 1.5285, 1.2505, 1.2537], + device='cuda:2'), covar=tensor([0.1459, 0.1352, 0.1481, 0.1095, 0.0765, 0.1212, 0.1522, 0.1901], + device='cuda:2'), in_proj_covar=tensor([0.0317, 0.0309, 0.0351, 0.0288, 0.0328, 0.0308, 0.0301, 0.0378], + device='cuda:2'), 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:2') +2023-04-28 04:03:35,474 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2912, 1.2454, 1.5260, 1.5125, 1.1812, 1.1382, 1.2600, 0.8141], + device='cuda:2'), covar=tensor([0.0547, 0.0475, 0.0375, 0.0422, 0.0719, 0.0848, 0.0455, 0.0483], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:03:38,164 INFO [finetune.py:976] (2/7) Epoch 29, batch 1300, loss[loss=0.1142, simple_loss=0.1952, pruned_loss=0.01655, over 4911.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2367, pruned_loss=0.04581, over 951293.13 frames. ], batch size: 36, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:03:46,267 INFO [optim.py:369] (2/7) 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 04:04:30,686 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-04-28 04:04:42,648 INFO [finetune.py:976] (2/7) Epoch 29, batch 1350, loss[loss=0.1885, simple_loss=0.2701, pruned_loss=0.05351, over 4926.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2358, pruned_loss=0.04518, over 953379.29 frames. ], batch size: 38, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:05:27,906 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161757.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:05:50,109 INFO [finetune.py:976] (2/7) Epoch 29, batch 1400, loss[loss=0.1279, simple_loss=0.1973, pruned_loss=0.02923, over 4783.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2393, pruned_loss=0.04623, over 953721.57 frames. ], batch size: 26, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:05:58,006 INFO [optim.py:369] (2/7) 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 04:06:51,812 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161816.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:07:03,297 INFO [finetune.py:976] (2/7) Epoch 29, batch 1450, loss[loss=0.1806, simple_loss=0.258, pruned_loss=0.05164, over 4836.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2402, pruned_loss=0.04593, over 954939.69 frames. ], batch size: 49, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:07:17,949 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.01 vs. limit=5.0 +2023-04-28 04:07:50,998 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2893, 2.0905, 2.2399, 2.7400, 2.7401, 2.1457, 1.9997, 2.3967], + device='cuda:2'), covar=tensor([0.0810, 0.1064, 0.0746, 0.0531, 0.0576, 0.0909, 0.0694, 0.0544], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0202, 0.0183, 0.0170, 0.0177, 0.0177, 0.0151, 0.0176], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:08:00,375 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161872.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:08:02,916 INFO [finetune.py:976] (2/7) Epoch 29, batch 1500, loss[loss=0.1579, simple_loss=0.2422, pruned_loss=0.03679, over 4902.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2427, pruned_loss=0.04721, over 955364.22 frames. ], batch size: 46, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:08:03,661 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161876.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:08:04,720 INFO [optim.py:369] (2/7) 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 04:08:15,991 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161887.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:08:39,422 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5403, 1.6785, 1.8235, 1.9466, 1.7566, 1.7848, 1.9413, 1.9532], + device='cuda:2'), covar=tensor([0.3844, 0.5452, 0.4579, 0.4262, 0.5667, 0.6956, 0.5184, 0.4517], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0374, 0.0331, 0.0341, 0.0349, 0.0394, 0.0362, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 04:09:01,028 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161920.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:09:09,619 INFO [finetune.py:976] (2/7) Epoch 29, batch 1550, loss[loss=0.1292, simple_loss=0.2118, pruned_loss=0.02323, over 4783.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2419, pruned_loss=0.04632, over 956420.49 frames. ], batch size: 29, lr: 2.85e-03, grad_scale: 16.0 +2023-04-28 04:09:20,178 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5575, 3.2736, 0.9914, 1.8977, 1.9298, 2.4418, 1.9712, 1.0144], + device='cuda:2'), covar=tensor([0.1283, 0.1043, 0.1883, 0.1139, 0.0995, 0.0918, 0.1402, 0.1939], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0236, 0.0135, 0.0119, 0.0131, 0.0152, 0.0116, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:09:21,320 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161935.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:10:05,051 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.7970, 2.3501, 2.1252, 2.2890, 2.0874, 2.1522, 2.2487, 2.2001], + device='cuda:2'), covar=tensor([0.4027, 0.5098, 0.4942, 0.3934, 0.5725, 0.6537, 0.5655, 0.4919], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0376, 0.0332, 0.0343, 0.0351, 0.0396, 0.0363, 0.0336], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 04:10:14,250 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8303, 1.3953, 1.6097, 1.6957, 1.6158, 1.3493, 0.9250, 1.4183], + device='cuda:2'), covar=tensor([0.2449, 0.2546, 0.1341, 0.1640, 0.2097, 0.2121, 0.4223, 0.1639], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0244, 0.0226, 0.0311, 0.0221, 0.0233, 0.0227, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 04:10:14,712 INFO [finetune.py:976] (2/7) Epoch 29, batch 1600, loss[loss=0.2289, simple_loss=0.2953, pruned_loss=0.08121, over 4341.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2408, pruned_loss=0.04648, over 954729.64 frames. ], batch size: 66, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:10:16,473 INFO [optim.py:369] (2/7) 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 04:10:57,116 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2122, 1.4748, 1.3830, 1.7489, 1.6819, 1.7280, 1.4349, 3.0926], + device='cuda:2'), covar=tensor([0.0652, 0.0842, 0.0801, 0.1215, 0.0612, 0.0520, 0.0707, 0.0182], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 04:11:22,117 INFO [finetune.py:976] (2/7) Epoch 29, batch 1650, loss[loss=0.1244, simple_loss=0.1982, pruned_loss=0.02536, over 4288.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2366, pruned_loss=0.04526, over 951539.31 frames. ], batch size: 19, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:11:23,615 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-28 04:12:07,061 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 1700, loss[loss=0.1371, simple_loss=0.213, pruned_loss=0.03064, over 4760.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2348, pruned_loss=0.04483, over 953152.23 frames. ], batch size: 28, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:12:36,099 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 04:12:59,487 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4250, 1.5732, 1.8157, 2.7856, 2.8045, 2.1947, 2.0775, 2.4622], + device='cuda:2'), covar=tensor([0.0833, 0.1768, 0.1210, 0.0537, 0.0577, 0.1019, 0.0786, 0.0581], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0203, 0.0185, 0.0171, 0.0178, 0.0178, 0.0151, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:13:11,488 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162105.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:13:24,520 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162116.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:13:34,111 INFO [finetune.py:976] (2/7) Epoch 29, batch 1750, loss[loss=0.2248, simple_loss=0.2996, pruned_loss=0.07502, over 4858.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2367, pruned_loss=0.04546, over 953927.71 frames. ], batch size: 49, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:13:34,220 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2488, 1.9124, 2.1337, 2.6502, 2.6822, 2.0809, 2.0001, 2.2626], + device='cuda:2'), covar=tensor([0.0880, 0.1296, 0.0864, 0.0613, 0.0655, 0.0941, 0.0750, 0.0647], + device='cuda:2'), in_proj_covar=tensor([0.0185, 0.0204, 0.0185, 0.0171, 0.0178, 0.0178, 0.0152, 0.0177], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:13:51,955 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162138.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:14:27,315 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162164.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:14:37,340 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162171.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:14:44,797 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.71 vs. limit=5.0 +2023-04-28 04:14:45,213 INFO [finetune.py:976] (2/7) Epoch 29, batch 1800, loss[loss=0.1527, simple_loss=0.2309, pruned_loss=0.03728, over 4883.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2381, pruned_loss=0.04498, over 953126.37 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:14:45,342 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7021, 2.1098, 1.5284, 1.4012, 1.2812, 1.2563, 1.4385, 1.1816], + device='cuda:2'), covar=tensor([0.1627, 0.1223, 0.1398, 0.1627, 0.2109, 0.1911, 0.1036, 0.1997], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0210, 0.0171, 0.0205, 0.0201, 0.0187, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 04:14:47,023 INFO [optim.py:369] (2/7) 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 04:15:29,735 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9170, 2.6233, 2.2390, 2.4954, 1.8255, 2.2194, 2.1686, 1.6567], + device='cuda:2'), covar=tensor([0.2265, 0.1181, 0.0784, 0.1196, 0.3126, 0.1010, 0.1929, 0.2589], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0300, 0.0216, 0.0276, 0.0313, 0.0252, 0.0248, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1244e-04, 1.1791e-04, 8.4745e-05, 1.0843e-04, 1.2622e-04, 9.9110e-05, + 9.9712e-05, 1.0364e-04], device='cuda:2') +2023-04-28 04:15:49,784 INFO [finetune.py:976] (2/7) Epoch 29, batch 1850, loss[loss=0.1325, simple_loss=0.2117, pruned_loss=0.02668, over 4739.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2392, pruned_loss=0.04543, over 952049.92 frames. ], batch size: 27, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:15:53,567 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162231.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:16:24,934 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8327, 1.6890, 1.6861, 1.4001, 1.8951, 1.5372, 2.3010, 1.4396], + device='cuda:2'), covar=tensor([0.2991, 0.1861, 0.4137, 0.2531, 0.1287, 0.2089, 0.1199, 0.4484], + device='cuda:2'), in_proj_covar=tensor([0.0338, 0.0353, 0.0421, 0.0352, 0.0384, 0.0374, 0.0370, 0.0423], + device='cuda:2'), out_proj_covar=tensor([9.9698e-05, 1.0488e-04, 1.2728e-04, 1.0529e-04, 1.1361e-04, 1.1111e-04, + 1.0795e-04, 1.2700e-04], device='cuda:2') +2023-04-28 04:16:53,967 INFO [finetune.py:976] (2/7) Epoch 29, batch 1900, loss[loss=0.1838, simple_loss=0.2689, pruned_loss=0.04934, over 4863.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2417, pruned_loss=0.0459, over 951276.15 frames. ], batch size: 31, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:16:55,788 INFO [optim.py:369] (2/7) 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 04:17:14,654 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162292.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:17:58,314 INFO [finetune.py:976] (2/7) Epoch 29, batch 1950, loss[loss=0.1936, simple_loss=0.2608, pruned_loss=0.06315, over 4821.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2415, pruned_loss=0.04582, over 953670.60 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:19:03,738 INFO [finetune.py:976] (2/7) Epoch 29, batch 2000, loss[loss=0.1639, simple_loss=0.2313, pruned_loss=0.04828, over 4906.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2381, pruned_loss=0.04485, over 954924.94 frames. ], batch size: 36, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:19:05,555 INFO [optim.py:369] (2/7) 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 04:19:43,204 INFO [finetune.py:976] (2/7) Epoch 29, batch 2050, loss[loss=0.1663, simple_loss=0.2257, pruned_loss=0.05346, over 4888.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2345, pruned_loss=0.04389, over 955230.74 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 16.0 +2023-04-28 04:19:48,151 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162433.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:19:58,571 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4274, 1.9132, 2.3549, 2.6659, 2.2612, 1.8123, 1.4784, 1.9918], + device='cuda:2'), covar=tensor([0.3427, 0.3076, 0.1686, 0.2226, 0.2713, 0.2722, 0.4192, 0.1936], + device='cuda:2'), in_proj_covar=tensor([0.0290, 0.0246, 0.0227, 0.0312, 0.0222, 0.0235, 0.0227, 0.0185], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 04:20:13,308 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162471.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:20:16,688 INFO [finetune.py:976] (2/7) Epoch 29, batch 2100, loss[loss=0.17, simple_loss=0.2347, pruned_loss=0.05268, over 4688.00 frames. ], tot_loss[loss=0.16, simple_loss=0.2333, pruned_loss=0.04338, over 956230.54 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:20:19,020 INFO [optim.py:369] (2/7) 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 04:20:39,348 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162511.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:20:45,708 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162519.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:20:50,372 INFO [finetune.py:976] (2/7) Epoch 29, batch 2150, loss[loss=0.154, simple_loss=0.2383, pruned_loss=0.0349, over 4779.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2372, pruned_loss=0.0447, over 957060.72 frames. ], batch size: 28, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:21:01,288 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-28 04:21:21,596 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162572.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 04:21:22,877 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.37 vs. limit=5.0 +2023-04-28 04:21:23,762 INFO [finetune.py:976] (2/7) Epoch 29, batch 2200, loss[loss=0.1656, simple_loss=0.2522, pruned_loss=0.03954, over 4891.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2387, pruned_loss=0.04494, over 956324.06 frames. ], batch size: 36, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:21:26,031 INFO [optim.py:369] (2/7) 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 04:21:37,951 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 2250, loss[loss=0.1609, simple_loss=0.2384, pruned_loss=0.04167, over 4806.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2401, pruned_loss=0.04582, over 955225.00 frames. ], batch size: 40, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:23:28,884 INFO [finetune.py:976] (2/7) Epoch 29, batch 2300, loss[loss=0.2129, simple_loss=0.2776, pruned_loss=0.0741, over 4883.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.241, pruned_loss=0.04571, over 955189.77 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:23:35,931 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4557, 3.3717, 1.0851, 1.7215, 1.9166, 2.3160, 2.0320, 1.0544], + device='cuda:2'), covar=tensor([0.1945, 0.1485, 0.2217, 0.1972, 0.1397, 0.1508, 0.1773, 0.2178], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0237, 0.0136, 0.0120, 0.0131, 0.0153, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:23:36,409 INFO [optim.py:369] (2/7) 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 04:24:08,637 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5001, 1.8749, 1.9629, 2.0747, 1.9184, 1.9499, 2.0299, 2.0512], + device='cuda:2'), covar=tensor([0.3969, 0.5416, 0.4132, 0.4152, 0.5108, 0.6444, 0.4828, 0.4284], + device='cuda:2'), in_proj_covar=tensor([0.0347, 0.0378, 0.0335, 0.0346, 0.0354, 0.0399, 0.0366, 0.0338], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 04:24:12,282 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1180, 1.7137, 2.2559, 2.4491, 2.1816, 2.0112, 2.1240, 2.0382], + device='cuda:2'), covar=tensor([0.4289, 0.7372, 0.6800, 0.5463, 0.5641, 0.7929, 0.8941, 1.1060], + device='cuda:2'), in_proj_covar=tensor([0.0449, 0.0426, 0.0522, 0.0511, 0.0477, 0.0513, 0.0516, 0.0530], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:24:31,808 INFO [finetune.py:976] (2/7) Epoch 29, batch 2350, loss[loss=0.1469, simple_loss=0.2298, pruned_loss=0.03198, over 4906.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2388, pruned_loss=0.04537, over 954005.00 frames. ], batch size: 43, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:24:41,990 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162733.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:24:50,130 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0283, 2.5065, 2.1847, 2.4534, 1.8311, 2.1003, 2.1703, 1.6263], + device='cuda:2'), covar=tensor([0.1869, 0.1198, 0.0754, 0.1052, 0.3135, 0.1003, 0.1824, 0.2206], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0300, 0.0217, 0.0277, 0.0314, 0.0254, 0.0249, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1327e-04, 1.1808e-04, 8.5307e-05, 1.0881e-04, 1.2657e-04, 9.9592e-05, + 1.0013e-04, 1.0411e-04], device='cuda:2') +2023-04-28 04:25:30,907 INFO [finetune.py:976] (2/7) Epoch 29, batch 2400, loss[loss=0.1331, simple_loss=0.1985, pruned_loss=0.03383, over 4761.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2353, pruned_loss=0.0444, over 953889.29 frames. ], batch size: 26, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:25:33,150 INFO [optim.py:369] (2/7) 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 04:25:36,110 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162781.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:26:04,588 INFO [finetune.py:976] (2/7) Epoch 29, batch 2450, loss[loss=0.1763, simple_loss=0.2412, pruned_loss=0.05565, over 4406.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.234, pruned_loss=0.04423, over 954946.07 frames. ], batch size: 19, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:26:05,936 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162827.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:26:17,721 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-28 04:26:33,770 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162867.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 04:26:36,424 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-28 04:26:38,577 INFO [finetune.py:976] (2/7) Epoch 29, batch 2500, loss[loss=0.1787, simple_loss=0.2488, pruned_loss=0.05431, over 4885.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2358, pruned_loss=0.04487, over 952866.37 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:26:40,362 INFO [optim.py:369] (2/7) 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 04:26:47,961 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162887.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:26:48,603 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162888.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:27:10,178 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-28 04:27:12,362 INFO [finetune.py:976] (2/7) Epoch 29, batch 2550, loss[loss=0.1686, simple_loss=0.2412, pruned_loss=0.04796, over 4927.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2407, pruned_loss=0.04623, over 952853.10 frames. ], batch size: 38, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:27:14,992 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9474, 1.4401, 2.0193, 2.4134, 2.0173, 1.9008, 1.9231, 1.9251], + device='cuda:2'), covar=tensor([0.4233, 0.6478, 0.6315, 0.5271, 0.5883, 0.7487, 0.8221, 0.8174], + device='cuda:2'), in_proj_covar=tensor([0.0447, 0.0424, 0.0518, 0.0506, 0.0473, 0.0509, 0.0513, 0.0527], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:27:19,030 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162935.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:27:40,128 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9440, 2.4941, 2.1207, 2.4149, 1.7698, 2.1053, 2.0340, 1.6323], + device='cuda:2'), covar=tensor([0.1799, 0.1087, 0.0766, 0.1036, 0.3045, 0.0986, 0.1836, 0.2474], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0299, 0.0216, 0.0275, 0.0313, 0.0252, 0.0248, 0.0263], + device='cuda:2'), out_proj_covar=tensor([1.1297e-04, 1.1751e-04, 8.4873e-05, 1.0828e-04, 1.2599e-04, 9.9074e-05, + 9.9688e-05, 1.0359e-04], device='cuda:2') +2023-04-28 04:27:46,026 INFO [finetune.py:976] (2/7) Epoch 29, batch 2600, loss[loss=0.1709, simple_loss=0.2343, pruned_loss=0.05378, over 4774.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2414, pruned_loss=0.04644, over 950843.42 frames. ], batch size: 26, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:27:47,802 INFO [optim.py:369] (2/7) 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 04:28:11,114 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3263, 2.5749, 1.7231, 2.0836, 2.7804, 2.1896, 2.0872, 2.2349], + device='cuda:2'), covar=tensor([0.0408, 0.0289, 0.0245, 0.0453, 0.0205, 0.0400, 0.0416, 0.0479], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 04:28:19,510 INFO [finetune.py:976] (2/7) Epoch 29, batch 2650, loss[loss=0.1695, simple_loss=0.2589, pruned_loss=0.04006, over 4916.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2431, pruned_loss=0.04721, over 950150.84 frames. ], batch size: 38, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:28:30,395 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9314, 1.4388, 1.7485, 1.7709, 1.7302, 1.4228, 0.8748, 1.4283], + device='cuda:2'), covar=tensor([0.3140, 0.3145, 0.1704, 0.2006, 0.2469, 0.2595, 0.4041, 0.1972], + device='cuda:2'), in_proj_covar=tensor([0.0291, 0.0245, 0.0226, 0.0312, 0.0221, 0.0234, 0.0226, 0.0184], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 04:28:57,567 INFO [finetune.py:976] (2/7) Epoch 29, batch 2700, loss[loss=0.144, simple_loss=0.2149, pruned_loss=0.03651, over 4820.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2417, pruned_loss=0.04641, over 952105.86 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:29:04,329 INFO [optim.py:369] (2/7) 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 04:29:17,274 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163090.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:29:46,784 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1049, 1.7887, 2.3069, 2.4354, 2.1493, 2.0509, 2.1411, 2.0961], + device='cuda:2'), covar=tensor([0.4104, 0.6657, 0.6131, 0.5045, 0.5384, 0.7691, 0.8053, 0.9921], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0425, 0.0520, 0.0508, 0.0475, 0.0511, 0.0515, 0.0529], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:29:57,695 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-04-28 04:29:58,663 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3126, 1.5723, 1.5546, 1.7571, 1.7095, 1.9012, 1.4759, 3.4930], + device='cuda:2'), covar=tensor([0.0604, 0.0797, 0.0752, 0.1183, 0.0599, 0.0522, 0.0712, 0.0153], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 04:30:02,209 INFO [finetune.py:976] (2/7) Epoch 29, batch 2750, loss[loss=0.1766, simple_loss=0.2416, pruned_loss=0.05584, over 4808.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2397, pruned_loss=0.04635, over 951965.30 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:30:20,493 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-28 04:30:38,994 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163151.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:30:50,657 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163167.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 04:31:01,090 INFO [finetune.py:976] (2/7) Epoch 29, batch 2800, loss[loss=0.1644, simple_loss=0.2323, pruned_loss=0.04823, over 4907.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2371, pruned_loss=0.04601, over 953873.43 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:31:08,097 INFO [optim.py:369] (2/7) 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 04:31:11,257 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163183.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:31:52,458 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163215.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:31:59,014 INFO [finetune.py:976] (2/7) Epoch 29, batch 2850, loss[loss=0.126, simple_loss=0.2076, pruned_loss=0.02222, over 4085.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2354, pruned_loss=0.0455, over 953838.12 frames. ], batch size: 65, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:32:08,309 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163240.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:32:17,196 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.3061, 3.1712, 2.5246, 3.8244, 3.3375, 3.2345, 1.2791, 3.2817], + device='cuda:2'), covar=tensor([0.2239, 0.1715, 0.3933, 0.2490, 0.4296, 0.2215, 0.6794, 0.2968], + device='cuda:2'), in_proj_covar=tensor([0.0250, 0.0222, 0.0255, 0.0308, 0.0305, 0.0255, 0.0279, 0.0278], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 04:32:30,357 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-04-28 04:32:31,404 INFO [finetune.py:976] (2/7) Epoch 29, batch 2900, loss[loss=0.1626, simple_loss=0.2432, pruned_loss=0.041, over 4804.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2374, pruned_loss=0.0462, over 954777.50 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:32:33,719 INFO [optim.py:369] (2/7) 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 04:32:47,875 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163301.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:33:04,180 INFO [finetune.py:976] (2/7) Epoch 29, batch 2950, loss[loss=0.1788, simple_loss=0.2642, pruned_loss=0.04672, over 4807.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2412, pruned_loss=0.04696, over 955342.14 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:33:12,085 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4245, 3.2049, 1.0121, 1.6993, 1.8435, 2.1827, 1.8535, 0.9995], + device='cuda:2'), covar=tensor([0.1423, 0.1114, 0.1873, 0.1260, 0.1075, 0.1100, 0.1542, 0.1945], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0238, 0.0136, 0.0120, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:33:31,501 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-04-28 04:33:37,006 INFO [finetune.py:976] (2/7) Epoch 29, batch 3000, loss[loss=0.167, simple_loss=0.2438, pruned_loss=0.04505, over 4744.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2415, pruned_loss=0.0467, over 954804.73 frames. ], batch size: 27, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:33:37,006 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 04:33:44,591 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7608, 2.1856, 1.8548, 2.1307, 1.6812, 1.8678, 1.8554, 1.4621], + device='cuda:2'), covar=tensor([0.1727, 0.0855, 0.0745, 0.0903, 0.3208, 0.0920, 0.1464, 0.2070], + device='cuda:2'), in_proj_covar=tensor([0.0284, 0.0299, 0.0217, 0.0276, 0.0313, 0.0253, 0.0247, 0.0264], + device='cuda:2'), out_proj_covar=tensor([1.1301e-04, 1.1745e-04, 8.5116e-05, 1.0839e-04, 1.2617e-04, 9.9184e-05, + 9.9584e-05, 1.0383e-04], device='cuda:2') +2023-04-28 04:33:47,841 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6415MB +2023-04-28 04:33:49,650 INFO [optim.py:369] (2/7) 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 04:33:58,143 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4696, 1.3166, 1.8109, 1.7551, 1.3345, 1.2540, 1.4095, 0.9553], + device='cuda:2'), covar=tensor([0.0591, 0.0651, 0.0352, 0.0625, 0.0679, 0.1060, 0.0585, 0.0587], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0076, 0.0095, 0.0073, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:34:19,350 INFO [finetune.py:976] (2/7) Epoch 29, batch 3050, loss[loss=0.144, simple_loss=0.2289, pruned_loss=0.02955, over 4759.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2423, pruned_loss=0.0462, over 955283.36 frames. ], batch size: 28, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:34:23,681 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-28 04:34:34,890 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163446.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:34:50,195 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.04 vs. limit=5.0 +2023-04-28 04:34:53,042 INFO [finetune.py:976] (2/7) Epoch 29, batch 3100, loss[loss=0.1545, simple_loss=0.2346, pruned_loss=0.03719, over 4891.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2388, pruned_loss=0.04505, over 954325.72 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:34:55,824 INFO [optim.py:369] (2/7) 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 04:34:59,437 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163483.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:35:10,682 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1655, 1.9572, 2.3645, 2.6280, 2.1967, 2.1012, 2.2837, 2.2411], + device='cuda:2'), covar=tensor([0.4677, 0.7243, 0.7464, 0.5527, 0.6263, 0.9287, 0.9347, 0.9489], + device='cuda:2'), in_proj_covar=tensor([0.0449, 0.0426, 0.0523, 0.0510, 0.0476, 0.0513, 0.0515, 0.0531], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:35:11,376 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-04-28 04:35:27,054 INFO [finetune.py:976] (2/7) Epoch 29, batch 3150, loss[loss=0.1531, simple_loss=0.2366, pruned_loss=0.03478, over 4846.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2365, pruned_loss=0.04445, over 955264.61 frames. ], batch size: 44, lr: 2.84e-03, grad_scale: 32.0 +2023-04-28 04:35:27,165 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.6217, 1.3754, 1.3984, 1.1497, 1.3044, 1.1043, 1.6115, 1.1635], + device='cuda:2'), covar=tensor([0.3483, 0.1979, 0.4905, 0.2486, 0.1720, 0.2574, 0.1827, 0.5668], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0355, 0.0424, 0.0354, 0.0385, 0.0377, 0.0372, 0.0426], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:35:28,930 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4562, 3.7582, 0.9489, 2.0141, 2.1239, 2.5607, 2.2608, 0.9926], + device='cuda:2'), covar=tensor([0.1365, 0.0916, 0.1854, 0.1225, 0.0958, 0.1020, 0.1358, 0.2183], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0238, 0.0136, 0.0120, 0.0131, 0.0152, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:35:31,734 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163531.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:35:59,149 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-28 04:36:10,392 INFO [finetune.py:976] (2/7) Epoch 29, batch 3200, loss[loss=0.1547, simple_loss=0.2229, pruned_loss=0.04322, over 4752.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2346, pruned_loss=0.04393, over 955876.83 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:36:12,739 INFO [optim.py:369] (2/7) 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 04:36:12,897 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0723, 1.7311, 2.2386, 2.4525, 2.1306, 1.9936, 2.1413, 2.1324], + device='cuda:2'), covar=tensor([0.4834, 0.7394, 0.6753, 0.5676, 0.6201, 0.8784, 0.8601, 0.9581], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0425, 0.0521, 0.0507, 0.0475, 0.0512, 0.0514, 0.0530], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:36:42,511 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163596.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:36:58,177 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1060, 2.5354, 0.8098, 1.4847, 1.5599, 1.7642, 1.6643, 0.8830], + device='cuda:2'), covar=tensor([0.1472, 0.0927, 0.1658, 0.1246, 0.1055, 0.0989, 0.1465, 0.1667], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0238, 0.0136, 0.0121, 0.0132, 0.0153, 0.0117, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:37:05,580 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-04-28 04:37:17,128 INFO [finetune.py:976] (2/7) Epoch 29, batch 3250, loss[loss=0.1818, simple_loss=0.2571, pruned_loss=0.05324, over 4760.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2352, pruned_loss=0.04451, over 955407.01 frames. ], batch size: 54, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:38:20,406 INFO [finetune.py:976] (2/7) Epoch 29, batch 3300, loss[loss=0.148, simple_loss=0.2401, pruned_loss=0.02796, over 4899.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2372, pruned_loss=0.04517, over 952229.36 frames. ], batch size: 32, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:38:22,204 INFO [optim.py:369] (2/7) 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 04:39:05,846 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9050, 1.8847, 2.4612, 2.5708, 1.7356, 1.7377, 1.8835, 1.1547], + device='cuda:2'), covar=tensor([0.0590, 0.0810, 0.0378, 0.0707, 0.0791, 0.1007, 0.0784, 0.0684], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0076, 0.0095, 0.0073, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 04:39:15,522 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-04-28 04:39:22,575 INFO [finetune.py:976] (2/7) Epoch 29, batch 3350, loss[loss=0.1527, simple_loss=0.1972, pruned_loss=0.05405, over 4094.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2399, pruned_loss=0.0463, over 952379.47 frames. ], batch size: 18, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:39:46,990 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163746.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:40:27,591 INFO [finetune.py:976] (2/7) Epoch 29, batch 3400, loss[loss=0.1565, simple_loss=0.2453, pruned_loss=0.03385, over 4853.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.241, pruned_loss=0.04645, over 952484.05 frames. ], batch size: 44, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:40:29,469 INFO [optim.py:369] (2/7) 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 04:40:50,556 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163794.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:41:23,190 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-28 04:41:32,571 INFO [finetune.py:976] (2/7) Epoch 29, batch 3450, loss[loss=0.1535, simple_loss=0.2262, pruned_loss=0.04042, over 4759.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.24, pruned_loss=0.04564, over 952447.20 frames. ], batch size: 27, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:41:44,874 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-04-28 04:42:36,322 INFO [finetune.py:976] (2/7) Epoch 29, batch 3500, loss[loss=0.1546, simple_loss=0.2191, pruned_loss=0.04505, over 4939.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2384, pruned_loss=0.04558, over 952862.32 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:42:38,151 INFO [optim.py:369] (2/7) 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 04:42:59,210 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163896.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:43:28,395 INFO [finetune.py:976] (2/7) Epoch 29, batch 3550, loss[loss=0.1439, simple_loss=0.2243, pruned_loss=0.03171, over 4824.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2357, pruned_loss=0.04471, over 951561.50 frames. ], batch size: 41, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:43:35,165 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3253, 2.2377, 2.0799, 1.9674, 2.4694, 1.9854, 2.9399, 1.8581], + device='cuda:2'), covar=tensor([0.3140, 0.1799, 0.3878, 0.2540, 0.1428, 0.2315, 0.1251, 0.4073], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0354, 0.0423, 0.0352, 0.0385, 0.0376, 0.0371, 0.0425], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:43:39,967 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163944.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:43:53,021 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.05 vs. limit=5.0 +2023-04-28 04:44:01,287 INFO [finetune.py:976] (2/7) Epoch 29, batch 3600, loss[loss=0.1892, simple_loss=0.2589, pruned_loss=0.0598, over 4788.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2343, pruned_loss=0.04459, over 951258.52 frames. ], batch size: 51, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:44:03,546 INFO [optim.py:369] (2/7) 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 04:44:36,729 INFO [finetune.py:976] (2/7) Epoch 29, batch 3650, loss[loss=0.1714, simple_loss=0.2585, pruned_loss=0.04218, over 4829.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2359, pruned_loss=0.04541, over 950419.62 frames. ], batch size: 49, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:45:10,039 INFO [finetune.py:976] (2/7) Epoch 29, batch 3700, loss[loss=0.1376, simple_loss=0.2226, pruned_loss=0.02636, over 4756.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2377, pruned_loss=0.04587, over 947775.51 frames. ], batch size: 28, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:45:11,859 INFO [optim.py:369] (2/7) 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 04:45:15,509 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-04-28 04:45:19,687 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164090.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:45:43,377 INFO [finetune.py:976] (2/7) Epoch 29, batch 3750, loss[loss=0.1644, simple_loss=0.2467, pruned_loss=0.04107, over 4807.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2404, pruned_loss=0.04687, over 947521.79 frames. ], batch size: 45, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:46:11,064 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164151.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:46:18,670 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164155.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:46:22,246 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4166, 2.4514, 1.8943, 2.0667, 2.4812, 2.0717, 3.3685, 1.8549], + device='cuda:2'), covar=tensor([0.3875, 0.2367, 0.4538, 0.3779, 0.2108, 0.2871, 0.1363, 0.4473], + device='cuda:2'), in_proj_covar=tensor([0.0339, 0.0353, 0.0421, 0.0352, 0.0384, 0.0374, 0.0370, 0.0423], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:46:43,103 INFO [finetune.py:976] (2/7) Epoch 29, batch 3800, loss[loss=0.1892, simple_loss=0.2759, pruned_loss=0.05131, over 4921.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2415, pruned_loss=0.04657, over 950605.46 frames. ], batch size: 42, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:46:47,404 INFO [optim.py:369] (2/7) 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 04:46:47,511 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9708, 1.1717, 4.8608, 4.6166, 4.1862, 4.6546, 4.3809, 4.2645], + device='cuda:2'), covar=tensor([0.7258, 0.6391, 0.1022, 0.1713, 0.1072, 0.1337, 0.1759, 0.1635], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0307, 0.0404, 0.0406, 0.0348, 0.0411, 0.0316, 0.0360], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:47:00,266 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0804, 2.5895, 0.9485, 1.4284, 1.8982, 1.3068, 3.5171, 1.9368], + device='cuda:2'), covar=tensor([0.0698, 0.0656, 0.0833, 0.1293, 0.0552, 0.1008, 0.0295, 0.0548], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0063, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 04:47:00,924 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4965, 1.8211, 2.0160, 2.1198, 1.9693, 2.0382, 2.0245, 2.0040], + device='cuda:2'), covar=tensor([0.3737, 0.4977, 0.4289, 0.4015, 0.5055, 0.6392, 0.5016, 0.4745], + device='cuda:2'), in_proj_covar=tensor([0.0347, 0.0375, 0.0333, 0.0343, 0.0354, 0.0395, 0.0364, 0.0336], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 04:47:23,349 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164216.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:47:34,753 INFO [finetune.py:976] (2/7) Epoch 29, batch 3850, loss[loss=0.184, simple_loss=0.2486, pruned_loss=0.05965, over 4794.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2396, pruned_loss=0.04557, over 952098.87 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:47:35,569 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-28 04:47:50,230 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1876, 1.5743, 1.5081, 1.8364, 1.6443, 1.8375, 1.4998, 3.4871], + device='cuda:2'), covar=tensor([0.0659, 0.0813, 0.0777, 0.1231, 0.0663, 0.0466, 0.0710, 0.0127], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0037, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 04:48:06,360 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9835, 2.4619, 1.0313, 1.4023, 1.8600, 1.2456, 3.0379, 1.7037], + device='cuda:2'), covar=tensor([0.0673, 0.0529, 0.0720, 0.1241, 0.0477, 0.0993, 0.0244, 0.0575], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 04:48:38,391 INFO [finetune.py:976] (2/7) Epoch 29, batch 3900, loss[loss=0.1675, simple_loss=0.2377, pruned_loss=0.04867, over 4903.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2374, pruned_loss=0.04513, over 953742.29 frames. ], batch size: 35, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:48:40,199 INFO [optim.py:369] (2/7) 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 04:49:11,022 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8459, 2.2849, 1.9503, 2.2486, 1.6501, 1.9120, 1.9097, 1.4569], + device='cuda:2'), covar=tensor([0.1789, 0.0953, 0.0740, 0.0950, 0.2982, 0.1012, 0.1616, 0.2155], + device='cuda:2'), in_proj_covar=tensor([0.0280, 0.0297, 0.0215, 0.0273, 0.0310, 0.0251, 0.0244, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1169e-04, 1.1671e-04, 8.4236e-05, 1.0735e-04, 1.2477e-04, 9.8598e-05, + 9.8443e-05, 1.0271e-04], device='cuda:2') +2023-04-28 04:49:39,857 INFO [finetune.py:976] (2/7) Epoch 29, batch 3950, loss[loss=0.1398, simple_loss=0.2156, pruned_loss=0.032, over 4832.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2352, pruned_loss=0.04466, over 954653.87 frames. ], batch size: 39, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:50:24,361 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164360.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:50:24,536 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-04-28 04:50:25,010 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8816, 2.2979, 2.1114, 1.8493, 1.4137, 1.4575, 2.1363, 1.4035], + device='cuda:2'), covar=tensor([0.1759, 0.1555, 0.1315, 0.1721, 0.2387, 0.2023, 0.0943, 0.2082], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0210, 0.0171, 0.0204, 0.0201, 0.0187, 0.0157, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 04:50:45,026 INFO [finetune.py:976] (2/7) Epoch 29, batch 4000, loss[loss=0.1994, simple_loss=0.2789, pruned_loss=0.05995, over 4744.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2345, pruned_loss=0.04449, over 953015.38 frames. ], batch size: 59, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:50:47,323 INFO [optim.py:369] (2/7) 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 04:51:41,774 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164421.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:51:48,792 INFO [finetune.py:976] (2/7) Epoch 29, batch 4050, loss[loss=0.1691, simple_loss=0.2232, pruned_loss=0.05752, over 4275.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2378, pruned_loss=0.04562, over 953587.62 frames. ], batch size: 18, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:52:20,687 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164446.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:52:32,240 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1170, 1.2752, 1.2415, 1.5714, 1.4329, 1.4072, 1.3072, 2.4922], + device='cuda:2'), covar=tensor([0.0610, 0.0892, 0.0851, 0.1304, 0.0693, 0.0506, 0.0771, 0.0182], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 04:52:55,386 INFO [finetune.py:976] (2/7) Epoch 29, batch 4100, loss[loss=0.159, simple_loss=0.2473, pruned_loss=0.03537, over 4794.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2401, pruned_loss=0.04604, over 952450.70 frames. ], batch size: 45, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:53:02,651 INFO [optim.py:369] (2/7) 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 04:53:41,169 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164511.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:53:54,830 INFO [finetune.py:976] (2/7) Epoch 29, batch 4150, loss[loss=0.2122, simple_loss=0.2721, pruned_loss=0.07613, over 4907.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2421, pruned_loss=0.04694, over 953259.57 frames. ], batch size: 37, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:54:35,900 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164568.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:54:40,039 INFO [finetune.py:976] (2/7) Epoch 29, batch 4200, loss[loss=0.1352, simple_loss=0.207, pruned_loss=0.03175, over 4849.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2412, pruned_loss=0.04615, over 951493.32 frames. ], batch size: 44, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:54:41,959 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164578.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:54:42,443 INFO [optim.py:369] (2/7) 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 04:54:45,545 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9769, 1.4316, 1.8123, 1.7621, 1.7813, 1.4761, 0.8344, 1.4409], + device='cuda:2'), covar=tensor([0.2790, 0.2989, 0.1538, 0.1965, 0.2250, 0.2587, 0.4076, 0.1919], + device='cuda:2'), in_proj_covar=tensor([0.0295, 0.0248, 0.0231, 0.0317, 0.0224, 0.0239, 0.0231, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 04:55:14,302 INFO [finetune.py:976] (2/7) Epoch 29, batch 4250, loss[loss=0.1862, simple_loss=0.253, pruned_loss=0.05971, over 4902.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2393, pruned_loss=0.0455, over 950750.21 frames. ], batch size: 37, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:55:16,890 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164629.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:55:24,093 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164639.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 04:55:37,623 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164658.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:55:48,428 INFO [finetune.py:976] (2/7) Epoch 29, batch 4300, loss[loss=0.1316, simple_loss=0.2011, pruned_loss=0.03105, over 4789.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2366, pruned_loss=0.04497, over 951910.83 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:55:50,845 INFO [optim.py:369] (2/7) 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 04:56:16,866 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164716.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:56:18,742 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164719.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:56:22,292 INFO [finetune.py:976] (2/7) Epoch 29, batch 4350, loss[loss=0.1892, simple_loss=0.2477, pruned_loss=0.06533, over 4858.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2354, pruned_loss=0.04506, over 955321.86 frames. ], batch size: 31, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:56:33,993 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8529, 1.6708, 1.9414, 2.2672, 1.9792, 1.7984, 1.8577, 1.8255], + device='cuda:2'), covar=tensor([0.4143, 0.6011, 0.5839, 0.4896, 0.5482, 0.7834, 0.7791, 0.9132], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0425, 0.0522, 0.0511, 0.0476, 0.0514, 0.0517, 0.0530], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:56:36,302 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:56:36,314 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:57:17,722 INFO [finetune.py:976] (2/7) Epoch 29, batch 4400, loss[loss=0.165, simple_loss=0.25, pruned_loss=0.03996, over 4820.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2368, pruned_loss=0.04554, over 954671.41 frames. ], batch size: 39, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:57:20,181 INFO [optim.py:369] (2/7) 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 04:57:41,271 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164794.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:57:42,575 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2519, 2.2382, 1.7851, 1.8950, 2.2904, 1.9504, 2.7847, 1.6824], + device='cuda:2'), covar=tensor([0.3717, 0.2070, 0.4376, 0.3272, 0.1822, 0.2500, 0.1411, 0.4299], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0356, 0.0424, 0.0354, 0.0385, 0.0375, 0.0373, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 04:58:01,576 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164807.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:58:04,465 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164811.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:58:24,164 INFO [finetune.py:976] (2/7) Epoch 29, batch 4450, loss[loss=0.1991, simple_loss=0.283, pruned_loss=0.0576, over 4756.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2392, pruned_loss=0.04594, over 955942.92 frames. ], batch size: 28, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:58:47,674 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-04-28 04:58:59,954 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164859.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 04:59:10,502 INFO [finetune.py:976] (2/7) Epoch 29, batch 4500, loss[loss=0.1869, simple_loss=0.2584, pruned_loss=0.05773, over 4821.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2409, pruned_loss=0.04655, over 955237.68 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 04:59:18,300 INFO [optim.py:369] (2/7) 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 04:59:58,683 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3299, 3.0746, 0.8543, 1.6290, 1.7035, 2.2268, 1.7975, 0.9892], + device='cuda:2'), covar=tensor([0.1318, 0.1007, 0.1816, 0.1168, 0.1003, 0.0855, 0.1372, 0.1739], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0235, 0.0134, 0.0119, 0.0130, 0.0151, 0.0116, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:00:20,418 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164924.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:00:20,955 INFO [finetune.py:976] (2/7) Epoch 29, batch 4550, loss[loss=0.1891, simple_loss=0.259, pruned_loss=0.05964, over 4781.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2428, pruned_loss=0.04739, over 956601.15 frames. ], batch size: 51, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 05:00:31,465 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9539, 1.1494, 1.6926, 1.8099, 1.7144, 1.7353, 1.7055, 1.7050], + device='cuda:2'), covar=tensor([0.3617, 0.4942, 0.3877, 0.4146, 0.5053, 0.6719, 0.4479, 0.4094], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0374, 0.0331, 0.0342, 0.0352, 0.0395, 0.0363, 0.0335], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 05:00:32,586 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164934.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 05:01:05,796 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8723, 2.3363, 1.9255, 1.7757, 1.3688, 1.4485, 1.9384, 1.3369], + device='cuda:2'), covar=tensor([0.1652, 0.1320, 0.1383, 0.1635, 0.2407, 0.1843, 0.0982, 0.2047], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0209, 0.0170, 0.0204, 0.0201, 0.0187, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 05:01:26,332 INFO [finetune.py:976] (2/7) Epoch 29, batch 4600, loss[loss=0.1667, simple_loss=0.2405, pruned_loss=0.04644, over 4702.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2416, pruned_loss=0.04694, over 957436.59 frames. ], batch size: 59, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 05:01:29,217 INFO [optim.py:369] (2/7) 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 05:02:09,414 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-04-28 05:02:19,694 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165014.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:02:20,896 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165016.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:02:32,276 INFO [finetune.py:976] (2/7) Epoch 29, batch 4650, loss[loss=0.1178, simple_loss=0.1918, pruned_loss=0.02194, over 4791.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2399, pruned_loss=0.04712, over 956931.78 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 32.0 +2023-04-28 05:03:18,780 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165064.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:03:36,023 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3473, 2.9561, 2.2724, 2.5002, 1.6864, 1.6687, 2.4652, 1.7040], + device='cuda:2'), covar=tensor([0.1571, 0.1288, 0.1252, 0.1521, 0.2249, 0.1880, 0.0916, 0.1987], + device='cuda:2'), in_proj_covar=tensor([0.0199, 0.0210, 0.0171, 0.0204, 0.0202, 0.0188, 0.0157, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 05:03:37,007 INFO [finetune.py:976] (2/7) Epoch 29, batch 4700, loss[loss=0.1055, simple_loss=0.1835, pruned_loss=0.01375, over 4760.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2359, pruned_loss=0.04539, over 955249.68 frames. ], batch size: 27, lr: 2.83e-03, grad_scale: 16.0 +2023-04-28 05:03:40,048 INFO [optim.py:369] (2/7) 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 05:03:58,090 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2397, 3.3636, 0.7564, 1.6925, 1.7313, 2.4415, 1.9523, 1.0598], + device='cuda:2'), covar=tensor([0.1935, 0.1771, 0.2694, 0.1855, 0.1474, 0.1398, 0.1743, 0.2157], + device='cuda:2'), in_proj_covar=tensor([0.0116, 0.0236, 0.0135, 0.0120, 0.0131, 0.0152, 0.0116, 0.0117], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:04:09,167 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165102.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:04:09,316 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-04-28 05:04:18,304 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-04-28 05:04:41,331 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165123.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:04:42,450 INFO [finetune.py:976] (2/7) Epoch 29, batch 4750, loss[loss=0.1505, simple_loss=0.2275, pruned_loss=0.03672, over 4836.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2337, pruned_loss=0.04437, over 954751.36 frames. ], batch size: 47, lr: 2.83e-03, grad_scale: 16.0 +2023-04-28 05:04:43,241 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3399, 2.1275, 2.5633, 2.4668, 2.2969, 2.2843, 2.3986, 2.4082], + device='cuda:2'), covar=tensor([0.5153, 0.8270, 0.8029, 0.7930, 0.7444, 1.0268, 1.1334, 1.0113], + device='cuda:2'), in_proj_covar=tensor([0.0445, 0.0424, 0.0519, 0.0509, 0.0474, 0.0514, 0.0515, 0.0530], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:05:25,305 INFO [finetune.py:976] (2/7) Epoch 29, batch 4800, loss[loss=0.1949, simple_loss=0.27, pruned_loss=0.05986, over 4832.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2364, pruned_loss=0.04595, over 953750.77 frames. ], batch size: 39, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:05:28,761 INFO [optim.py:369] (2/7) 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 05:05:31,323 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165184.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:05:48,359 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1035, 2.6463, 1.0843, 1.4310, 1.9313, 1.2867, 3.5965, 1.9286], + device='cuda:2'), covar=tensor([0.0673, 0.0657, 0.0786, 0.1267, 0.0512, 0.1008, 0.0295, 0.0572], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 05:05:57,697 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165224.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:05:58,220 INFO [finetune.py:976] (2/7) Epoch 29, batch 4850, loss[loss=0.2024, simple_loss=0.2681, pruned_loss=0.06835, over 4926.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2391, pruned_loss=0.04632, over 954439.12 frames. ], batch size: 38, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:06:05,339 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165234.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 05:06:15,009 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1060, 1.5512, 1.6292, 1.9056, 2.2677, 1.8307, 1.6272, 1.5164], + device='cuda:2'), covar=tensor([0.1325, 0.1549, 0.1867, 0.1167, 0.0808, 0.1471, 0.1908, 0.2164], + device='cuda:2'), in_proj_covar=tensor([0.0312, 0.0307, 0.0348, 0.0286, 0.0324, 0.0303, 0.0299, 0.0375], + device='cuda:2'), out_proj_covar=tensor([6.3530e-05, 6.2858e-05, 7.2719e-05, 5.7166e-05, 6.6027e-05, 6.2844e-05, + 6.1752e-05, 7.9230e-05], device='cuda:2') +2023-04-28 05:06:29,516 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 29, batch 4900, loss[loss=0.1564, simple_loss=0.2313, pruned_loss=0.04072, over 3901.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.241, pruned_loss=0.04686, over 954081.86 frames. ], batch size: 16, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:06:35,317 INFO [optim.py:369] (2/7) 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 05:06:37,063 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165282.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:06:57,502 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165314.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:07:04,967 INFO [finetune.py:976] (2/7) Epoch 29, batch 4950, loss[loss=0.1855, simple_loss=0.2661, pruned_loss=0.05241, over 4909.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2425, pruned_loss=0.04737, over 955417.28 frames. ], batch size: 42, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:07:29,753 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165362.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:07:38,077 INFO [finetune.py:976] (2/7) Epoch 29, batch 5000, loss[loss=0.1761, simple_loss=0.2437, pruned_loss=0.05426, over 4729.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.241, pruned_loss=0.04674, over 953775.16 frames. ], batch size: 23, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:07:41,611 INFO [optim.py:369] (2/7) 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 05:07:48,078 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9233, 1.8455, 1.1249, 1.5668, 2.0783, 1.7769, 1.5809, 1.8044], + device='cuda:2'), covar=tensor([0.0447, 0.0344, 0.0300, 0.0520, 0.0237, 0.0466, 0.0461, 0.0506], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0054, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 05:07:50,355 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4789, 1.3681, 1.7640, 1.7369, 1.3207, 1.2486, 1.3786, 0.8347], + device='cuda:2'), covar=tensor([0.0508, 0.0493, 0.0316, 0.0454, 0.0710, 0.1031, 0.0467, 0.0517], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:07:57,213 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165402.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:08:08,237 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-04-28 05:08:12,134 INFO [finetune.py:976] (2/7) Epoch 29, batch 5050, loss[loss=0.1862, simple_loss=0.2532, pruned_loss=0.05964, over 4827.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2382, pruned_loss=0.04573, over 954404.06 frames. ], batch size: 39, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:08:29,763 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165450.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:08:45,559 INFO [finetune.py:976] (2/7) Epoch 29, batch 5100, loss[loss=0.156, simple_loss=0.2227, pruned_loss=0.04462, over 4836.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2353, pruned_loss=0.04484, over 955260.20 frames. ], batch size: 33, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:08:48,018 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8618, 1.4280, 1.4163, 1.7071, 2.0706, 1.6716, 1.4847, 1.4096], + device='cuda:2'), covar=tensor([0.1343, 0.1339, 0.1860, 0.1120, 0.0691, 0.1306, 0.1502, 0.1844], + device='cuda:2'), in_proj_covar=tensor([0.0313, 0.0308, 0.0349, 0.0287, 0.0326, 0.0304, 0.0300, 0.0376], + device='cuda:2'), out_proj_covar=tensor([6.3790e-05, 6.3095e-05, 7.2986e-05, 5.7256e-05, 6.6385e-05, 6.3060e-05, + 6.1860e-05, 7.9395e-05], device='cuda:2') +2023-04-28 05:08:48,563 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:09:48,070 INFO [finetune.py:976] (2/7) Epoch 29, batch 5150, loss[loss=0.2033, simple_loss=0.2722, pruned_loss=0.06716, over 4850.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2358, pruned_loss=0.04515, over 955546.70 frames. ], batch size: 47, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:09:50,651 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5234, 2.9875, 2.4435, 2.8723, 2.0675, 2.6860, 2.8449, 2.0946], + device='cuda:2'), covar=tensor([0.1867, 0.1339, 0.0910, 0.1221, 0.3467, 0.1204, 0.1732, 0.2556], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0300, 0.0216, 0.0275, 0.0312, 0.0253, 0.0247, 0.0262], + device='cuda:2'), out_proj_covar=tensor([1.1214e-04, 1.1809e-04, 8.4759e-05, 1.0801e-04, 1.2544e-04, 9.9390e-05, + 9.9240e-05, 1.0316e-04], device='cuda:2') +2023-04-28 05:10:21,699 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165551.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:10:52,971 INFO [finetune.py:976] (2/7) Epoch 29, batch 5200, loss[loss=0.1616, simple_loss=0.238, pruned_loss=0.04262, over 4895.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2383, pruned_loss=0.04545, over 952635.54 frames. ], batch size: 35, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:10:55,996 INFO [optim.py:369] (2/7) 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 05:11:45,801 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165612.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:11:59,707 INFO [finetune.py:976] (2/7) Epoch 29, batch 5250, loss[loss=0.2012, simple_loss=0.2677, pruned_loss=0.06736, over 4788.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2404, pruned_loss=0.04582, over 951799.40 frames. ], batch size: 51, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:12:44,426 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 05:12:54,747 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8331, 2.4150, 1.9719, 1.8296, 1.3564, 1.4139, 1.9978, 1.3311], + device='cuda:2'), covar=tensor([0.1646, 0.1288, 0.1229, 0.1503, 0.2143, 0.1832, 0.0883, 0.2016], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0211, 0.0172, 0.0205, 0.0203, 0.0188, 0.0158, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 05:13:07,360 INFO [finetune.py:976] (2/7) Epoch 29, batch 5300, loss[loss=0.1903, simple_loss=0.2626, pruned_loss=0.05899, over 4770.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2413, pruned_loss=0.04613, over 951559.14 frames. ], batch size: 26, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:13:16,015 INFO [optim.py:369] (2/7) 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 05:13:47,162 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7022, 0.6726, 1.4780, 2.0107, 1.7577, 1.5734, 1.5778, 1.6038], + device='cuda:2'), covar=tensor([0.4622, 0.7132, 0.6352, 0.6218, 0.6171, 0.7756, 0.7894, 0.8887], + device='cuda:2'), in_proj_covar=tensor([0.0447, 0.0425, 0.0520, 0.0509, 0.0474, 0.0514, 0.0515, 0.0530], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:14:12,925 INFO [finetune.py:976] (2/7) Epoch 29, batch 5350, loss[loss=0.1602, simple_loss=0.2327, pruned_loss=0.04381, over 4703.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2408, pruned_loss=0.04562, over 951866.97 frames. ], batch size: 23, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:14:43,454 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165750.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:15:21,307 INFO [finetune.py:976] (2/7) Epoch 29, batch 5400, loss[loss=0.123, simple_loss=0.2064, pruned_loss=0.01977, over 4760.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2381, pruned_loss=0.0449, over 952587.21 frames. ], batch size: 26, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:15:24,332 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:15:55,095 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165803.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:07,205 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165811.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:21,414 INFO [finetune.py:976] (2/7) Epoch 29, batch 5450, loss[loss=0.1403, simple_loss=0.213, pruned_loss=0.03378, over 4718.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2351, pruned_loss=0.04405, over 953922.53 frames. ], batch size: 23, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:16:22,149 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7949, 2.2215, 1.9062, 2.1652, 1.7407, 1.9358, 1.8488, 1.4244], + device='cuda:2'), covar=tensor([0.1955, 0.1364, 0.0867, 0.1238, 0.3136, 0.1176, 0.1911, 0.2705], + device='cuda:2'), in_proj_covar=tensor([0.0282, 0.0301, 0.0215, 0.0274, 0.0310, 0.0253, 0.0247, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1226e-04, 1.1822e-04, 8.4511e-05, 1.0764e-04, 1.2478e-04, 9.9459e-05, + 9.9302e-05, 1.0270e-04], device='cuda:2') +2023-04-28 05:16:22,707 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165827.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:33,600 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7318, 0.7321, 1.5127, 1.9342, 1.7362, 1.5546, 1.5407, 1.5960], + device='cuda:2'), covar=tensor([0.5917, 0.8340, 0.7604, 0.8866, 0.7327, 0.9540, 0.9870, 1.0531], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0427, 0.0523, 0.0511, 0.0477, 0.0517, 0.0518, 0.0533], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:16:48,219 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165864.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:16:49,424 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3518, 1.6124, 1.3821, 1.6415, 1.4008, 1.3559, 1.3898, 1.1001], + device='cuda:2'), covar=tensor([0.1722, 0.1302, 0.0970, 0.1141, 0.3810, 0.1321, 0.1817, 0.2341], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0301, 0.0216, 0.0274, 0.0310, 0.0254, 0.0247, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1214e-04, 1.1825e-04, 8.4599e-05, 1.0764e-04, 1.2479e-04, 9.9516e-05, + 9.9369e-05, 1.0281e-04], device='cuda:2') +2023-04-28 05:16:55,275 INFO [finetune.py:976] (2/7) Epoch 29, batch 5500, loss[loss=0.1288, simple_loss=0.1908, pruned_loss=0.03339, over 4365.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2323, pruned_loss=0.04359, over 954990.14 frames. ], batch size: 19, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:16:58,257 INFO [optim.py:369] (2/7) 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 05:16:58,350 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2482, 1.0884, 3.7363, 3.5208, 3.2978, 3.5735, 3.5307, 3.3285], + device='cuda:2'), covar=tensor([0.7405, 0.6185, 0.1239, 0.1737, 0.1335, 0.2308, 0.2266, 0.1540], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0311, 0.0410, 0.0412, 0.0353, 0.0416, 0.0320, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:17:16,279 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165907.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:17:18,805 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-04-28 05:17:21,520 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9803, 2.5754, 1.0720, 1.4492, 1.9268, 1.2598, 3.4906, 1.7066], + device='cuda:2'), covar=tensor([0.0716, 0.0699, 0.0815, 0.1184, 0.0507, 0.1019, 0.0228, 0.0626], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 05:17:23,286 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9760, 2.2617, 2.0268, 2.2347, 1.7290, 2.0689, 2.0939, 1.7229], + device='cuda:2'), covar=tensor([0.1617, 0.0987, 0.0657, 0.0938, 0.2492, 0.0855, 0.1370, 0.1698], + device='cuda:2'), in_proj_covar=tensor([0.0281, 0.0300, 0.0215, 0.0273, 0.0309, 0.0253, 0.0246, 0.0260], + device='cuda:2'), out_proj_covar=tensor([1.1181e-04, 1.1786e-04, 8.4376e-05, 1.0723e-04, 1.2445e-04, 9.9153e-05, + 9.9066e-05, 1.0251e-04], device='cuda:2') +2023-04-28 05:17:29,076 INFO [finetune.py:976] (2/7) Epoch 29, batch 5550, loss[loss=0.2032, simple_loss=0.2809, pruned_loss=0.06274, over 4899.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2359, pruned_loss=0.0456, over 953054.88 frames. ], batch size: 43, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:18:28,196 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7000, 1.1430, 4.0816, 3.6052, 3.6615, 3.8258, 3.7182, 3.4770], + device='cuda:2'), covar=tensor([0.9221, 0.8927, 0.1684, 0.3143, 0.2109, 0.3205, 0.4225, 0.3300], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0310, 0.0407, 0.0411, 0.0352, 0.0415, 0.0317, 0.0364], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:18:28,708 INFO [finetune.py:976] (2/7) Epoch 29, batch 5600, loss[loss=0.1738, simple_loss=0.2519, pruned_loss=0.04781, over 4760.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2391, pruned_loss=0.04648, over 954253.07 frames. ], batch size: 54, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:18:37,366 INFO [optim.py:369] (2/7) 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 05:19:30,412 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9383, 1.2702, 3.3086, 3.0840, 2.9440, 3.2409, 3.2449, 2.8994], + device='cuda:2'), covar=tensor([0.7632, 0.5357, 0.1505, 0.2147, 0.1486, 0.2307, 0.1679, 0.1811], + device='cuda:2'), in_proj_covar=tensor([0.0309, 0.0309, 0.0407, 0.0410, 0.0351, 0.0414, 0.0316, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:19:32,347 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-28 05:19:32,711 INFO [finetune.py:976] (2/7) Epoch 29, batch 5650, loss[loss=0.1452, simple_loss=0.2271, pruned_loss=0.03161, over 4836.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2411, pruned_loss=0.04626, over 954636.45 frames. ], batch size: 49, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:20:11,634 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166054.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 05:20:14,605 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7293, 1.5702, 0.7819, 1.4421, 1.5469, 1.5966, 1.4892, 1.5428], + device='cuda:2'), covar=tensor([0.0408, 0.0336, 0.0333, 0.0482, 0.0252, 0.0402, 0.0411, 0.0486], + device='cuda:2'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 05:20:33,889 INFO [finetune.py:976] (2/7) Epoch 29, batch 5700, loss[loss=0.1052, simple_loss=0.1718, pruned_loss=0.01925, over 4164.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2371, pruned_loss=0.04531, over 938090.00 frames. ], batch size: 18, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:20:42,336 INFO [optim.py:369] (2/7) 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 05:21:19,173 INFO [finetune.py:976] (2/7) Epoch 30, batch 0, loss[loss=0.2022, simple_loss=0.2719, pruned_loss=0.06626, over 4805.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2719, pruned_loss=0.06626, over 4805.00 frames. ], batch size: 39, lr: 2.82e-03, grad_scale: 16.0 +2023-04-28 05:21:19,174 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 05:21:20,958 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8137, 1.0924, 1.7603, 2.2974, 1.8799, 1.7161, 1.7193, 1.7513], + device='cuda:2'), covar=tensor([0.4866, 0.7455, 0.6850, 0.5778, 0.6199, 0.8430, 0.9026, 0.9085], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0427, 0.0523, 0.0511, 0.0477, 0.0517, 0.0519, 0.0533], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:21:26,825 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2935, 1.5687, 1.8268, 1.9629, 1.8760, 1.9296, 1.8095, 1.8690], + device='cuda:2'), covar=tensor([0.3675, 0.5451, 0.4332, 0.4282, 0.5327, 0.6803, 0.5211, 0.4460], + device='cuda:2'), in_proj_covar=tensor([0.0347, 0.0377, 0.0333, 0.0346, 0.0354, 0.0396, 0.0365, 0.0336], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 05:21:34,681 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6415MB +2023-04-28 05:21:37,612 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166106.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:21:48,603 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166115.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 05:22:23,300 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-04-28 05:22:31,250 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166147.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:22:35,223 INFO [finetune.py:976] (2/7) Epoch 30, batch 50, loss[loss=0.1765, simple_loss=0.2484, pruned_loss=0.05225, over 4863.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2392, pruned_loss=0.04658, over 214904.39 frames. ], batch size: 34, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:22:41,878 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-04-28 05:22:42,464 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-28 05:22:51,882 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166159.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:23:14,383 INFO [optim.py:369] (2/7) 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 05:23:44,638 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0755, 1.7639, 1.9256, 2.2997, 2.4248, 1.9110, 1.6285, 2.0378], + device='cuda:2'), covar=tensor([0.0738, 0.1101, 0.0688, 0.0550, 0.0535, 0.0833, 0.0726, 0.0533], + device='cuda:2'), in_proj_covar=tensor([0.0183, 0.0202, 0.0183, 0.0170, 0.0177, 0.0177, 0.0149, 0.0176], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:23:47,362 INFO [finetune.py:976] (2/7) Epoch 30, batch 100, loss[loss=0.1521, simple_loss=0.2227, pruned_loss=0.04081, over 4762.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.233, pruned_loss=0.04373, over 377376.68 frames. ], batch size: 28, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:23:50,916 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166207.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:23:56,828 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166208.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:24:41,723 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-04-28 05:24:53,823 INFO [finetune.py:976] (2/7) Epoch 30, batch 150, loss[loss=0.1693, simple_loss=0.2321, pruned_loss=0.05319, over 4886.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2312, pruned_loss=0.04431, over 507800.37 frames. ], batch size: 35, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:24:55,116 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166255.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:25:28,825 INFO [optim.py:369] (2/7) 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 05:25:59,857 INFO [finetune.py:976] (2/7) Epoch 30, batch 200, loss[loss=0.247, simple_loss=0.3187, pruned_loss=0.08767, over 4817.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2314, pruned_loss=0.04468, over 606141.85 frames. ], batch size: 40, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:26:02,452 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166307.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:26:40,371 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-04-28 05:27:02,967 INFO [finetune.py:976] (2/7) Epoch 30, batch 250, loss[loss=0.1689, simple_loss=0.2631, pruned_loss=0.03739, over 4852.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2363, pruned_loss=0.04627, over 685707.09 frames. ], batch size: 47, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:27:19,832 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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] (2/7) Epoch 30, batch 300, loss[loss=0.1377, simple_loss=0.212, pruned_loss=0.03171, over 4758.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2404, pruned_loss=0.04709, over 747208.18 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:28:07,530 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166406.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:28:09,962 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166410.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 05:28:28,641 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-04-28 05:29:01,977 INFO [finetune.py:976] (2/7) Epoch 30, batch 350, loss[loss=0.1491, simple_loss=0.2241, pruned_loss=0.03704, over 4784.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2413, pruned_loss=0.0471, over 794133.23 frames. ], batch size: 29, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:29:08,237 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166454.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:11,870 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:29:32,385 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166481.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:34,819 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 05:29:46,149 INFO [finetune.py:976] (2/7) Epoch 30, batch 400, loss[loss=0.1293, simple_loss=0.2065, pruned_loss=0.02602, over 4828.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.242, pruned_loss=0.04677, over 828990.92 frames. ], batch size: 49, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:29:46,222 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166507.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:29:50,823 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166509.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:06,141 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166531.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:13,231 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166542.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:19,823 INFO [finetune.py:976] (2/7) Epoch 30, batch 450, loss[loss=0.1814, simple_loss=0.2461, pruned_loss=0.05837, over 4763.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2397, pruned_loss=0.0458, over 855452.16 frames. ], batch size: 28, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:30:31,328 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166570.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:39,261 INFO [optim.py:369] (2/7) 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 05:30:47,075 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166592.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:30:53,785 INFO [finetune.py:976] (2/7) Epoch 30, batch 500, loss[loss=0.1549, simple_loss=0.2169, pruned_loss=0.04644, over 4872.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2363, pruned_loss=0.04452, over 877185.86 frames. ], batch size: 49, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:31:02,346 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=5.53 vs. limit=5.0 +2023-04-28 05:31:16,388 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-04-28 05:31:27,597 INFO [finetune.py:976] (2/7) Epoch 30, batch 550, loss[loss=0.1517, simple_loss=0.2234, pruned_loss=0.04001, over 4756.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2341, pruned_loss=0.04406, over 896669.92 frames. ], batch size: 27, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:31:34,187 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166663.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:31:45,971 INFO [optim.py:369] (2/7) 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 05:31:48,287 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6048, 3.7816, 0.7634, 2.1278, 2.1637, 2.7033, 2.1928, 1.0883], + device='cuda:2'), covar=tensor([0.1262, 0.0893, 0.2005, 0.1133, 0.1008, 0.0890, 0.1364, 0.1985], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0239, 0.0136, 0.0121, 0.0132, 0.0153, 0.0118, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:32:01,358 INFO [finetune.py:976] (2/7) Epoch 30, batch 600, loss[loss=0.2171, simple_loss=0.2812, pruned_loss=0.07644, over 4816.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2357, pruned_loss=0.04464, over 908845.57 frames. ], batch size: 45, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:32:05,746 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166710.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 05:32:07,097 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-04-28 05:32:23,052 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 650, loss[loss=0.1886, simple_loss=0.2688, pruned_loss=0.05416, over 4828.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2371, pruned_loss=0.04488, over 917182.99 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:32:37,945 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166758.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 05:32:47,101 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166764.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:32:55,018 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9324, 1.4375, 1.9539, 2.3721, 1.9951, 1.8666, 1.8748, 1.8822], + device='cuda:2'), covar=tensor([0.4278, 0.6546, 0.6242, 0.5452, 0.5630, 0.7921, 0.8243, 0.9246], + device='cuda:2'), in_proj_covar=tensor([0.0449, 0.0426, 0.0523, 0.0510, 0.0476, 0.0516, 0.0519, 0.0532], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:33:07,158 INFO [optim.py:369] (2/7) 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 05:33:29,696 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 700, loss[loss=0.1698, simple_loss=0.2558, pruned_loss=0.04191, over 4904.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2395, pruned_loss=0.04553, over 926031.62 frames. ], batch size: 35, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:33:39,563 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166803.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:33:40,204 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8852, 4.0566, 0.7609, 2.4968, 2.6581, 2.8689, 2.4453, 1.0441], + device='cuda:2'), covar=tensor([0.1233, 0.0815, 0.2158, 0.0997, 0.0873, 0.0953, 0.1491, 0.2066], + device='cuda:2'), in_proj_covar=tensor([0.0117, 0.0238, 0.0136, 0.0121, 0.0132, 0.0153, 0.0118, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:34:03,940 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166825.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:34:22,693 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166837.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:34:37,355 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166851.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:34:43,121 INFO [finetune.py:976] (2/7) Epoch 30, batch 750, loss[loss=0.1511, simple_loss=0.2342, pruned_loss=0.03401, over 4722.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2408, pruned_loss=0.04581, over 933122.49 frames. ], batch size: 59, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:34:51,920 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166865.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:34:52,609 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1160, 2.1217, 1.7611, 1.7422, 2.0437, 1.6788, 2.6559, 1.4658], + device='cuda:2'), covar=tensor([0.3536, 0.1943, 0.4614, 0.3073, 0.1829, 0.2651, 0.1340, 0.4816], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0356, 0.0428, 0.0354, 0.0388, 0.0377, 0.0373, 0.0428], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:35:11,850 INFO [optim.py:369] (2/7) 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 05:35:21,896 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 800, loss[loss=0.1504, simple_loss=0.2273, pruned_loss=0.0367, over 4931.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2421, pruned_loss=0.04588, over 939656.35 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:36:17,513 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166929.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:36:49,757 INFO [finetune.py:976] (2/7) Epoch 30, batch 850, loss[loss=0.1713, simple_loss=0.2417, pruned_loss=0.05045, over 4777.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2401, pruned_loss=0.04577, over 943081.15 frames. ], batch size: 29, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:36:59,928 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166963.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:37:20,967 INFO [optim.py:369] (2/7) 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 05:37:29,603 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-04-28 05:37:37,945 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166990.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:37:42,142 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-04-28 05:37:47,449 INFO [finetune.py:976] (2/7) Epoch 30, batch 900, loss[loss=0.1273, simple_loss=0.2093, pruned_loss=0.02267, over 4804.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2377, pruned_loss=0.04532, over 947536.73 frames. ], batch size: 25, lr: 2.81e-03, grad_scale: 16.0 +2023-04-28 05:37:52,367 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167011.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:38:12,963 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-04-28 05:38:21,329 INFO [finetune.py:976] (2/7) Epoch 30, batch 950, loss[loss=0.1676, simple_loss=0.2355, pruned_loss=0.04988, over 4831.00 frames. ], tot_loss[loss=0.163, simple_loss=0.236, pruned_loss=0.045, over 946692.01 frames. ], batch size: 30, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:38:38,159 INFO [optim.py:369] (2/7) 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 05:38:40,059 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.78 vs. limit=5.0 +2023-04-28 05:38:45,983 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167091.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:38:52,915 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167100.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:38:55,141 INFO [finetune.py:976] (2/7) Epoch 30, batch 1000, loss[loss=0.1787, simple_loss=0.2607, pruned_loss=0.04835, over 4912.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2381, pruned_loss=0.04591, over 948452.33 frames. ], batch size: 37, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:38:55,873 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167104.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:06,123 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167120.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:14,263 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.82 vs. limit=5.0 +2023-04-28 05:39:17,578 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167137.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:23,083 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-04-28 05:39:28,662 INFO [finetune.py:976] (2/7) Epoch 30, batch 1050, loss[loss=0.1738, simple_loss=0.2523, pruned_loss=0.04766, over 4918.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2404, pruned_loss=0.04594, over 949330.26 frames. ], batch size: 36, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:39:31,675 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5912, 3.2495, 1.2169, 1.9547, 1.9772, 2.5081, 2.0177, 1.3048], + device='cuda:2'), covar=tensor([0.1250, 0.0759, 0.1729, 0.1143, 0.1037, 0.0908, 0.1332, 0.2089], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0239, 0.0137, 0.0121, 0.0132, 0.0153, 0.0118, 0.0119], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:39:34,163 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167161.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:36,567 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:36,595 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:46,132 INFO [optim.py:369] (2/7) 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 05:39:49,157 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167185.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:39:50,860 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167187.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:40:01,670 INFO [finetune.py:976] (2/7) Epoch 30, batch 1100, loss[loss=0.1652, simple_loss=0.2393, pruned_loss=0.04551, over 4862.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2406, pruned_loss=0.04571, over 949610.12 frames. ], batch size: 34, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:40:08,808 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167213.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:40:22,661 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167235.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:40:34,490 INFO [finetune.py:976] (2/7) Epoch 30, batch 1150, loss[loss=0.1797, simple_loss=0.2578, pruned_loss=0.05084, over 4913.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2414, pruned_loss=0.04612, over 949959.52 frames. ], batch size: 46, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:41:03,287 INFO [optim.py:369] (2/7) 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 05:41:11,484 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167285.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:41:33,517 INFO [finetune.py:976] (2/7) Epoch 30, batch 1200, loss[loss=0.1397, simple_loss=0.2139, pruned_loss=0.03276, over 4786.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2404, pruned_loss=0.04648, over 949372.19 frames. ], batch size: 29, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:42:37,940 INFO [finetune.py:976] (2/7) Epoch 30, batch 1250, loss[loss=0.1337, simple_loss=0.2117, pruned_loss=0.02782, over 4764.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2372, pruned_loss=0.04558, over 952231.69 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:43:19,233 INFO [optim.py:369] (2/7) 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 05:43:22,506 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.18 vs. limit=5.0 +2023-04-28 05:43:32,135 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167391.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:43:42,859 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4436, 2.9502, 1.0329, 1.5903, 2.1536, 1.2857, 3.9547, 1.8606], + device='cuda:2'), covar=tensor([0.0676, 0.0805, 0.0862, 0.1241, 0.0501, 0.1034, 0.0246, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 05:43:45,349 INFO [finetune.py:976] (2/7) Epoch 30, batch 1300, loss[loss=0.1823, simple_loss=0.2477, pruned_loss=0.05847, over 4877.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2346, pruned_loss=0.04484, over 954088.59 frames. ], batch size: 34, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:44:12,877 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167420.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:44:25,757 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167432.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:44:35,812 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167439.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:44:45,739 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4674, 1.4203, 1.7002, 1.7112, 1.3218, 1.1733, 1.4902, 1.0364], + device='cuda:2'), covar=tensor([0.0468, 0.0414, 0.0356, 0.0396, 0.0568, 0.0837, 0.0470, 0.0488], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:44:56,435 INFO [finetune.py:976] (2/7) Epoch 30, batch 1350, loss[loss=0.184, simple_loss=0.2567, pruned_loss=0.05571, over 4918.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2348, pruned_loss=0.04481, over 955339.99 frames. ], batch size: 37, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:44:58,349 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167456.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:45:00,777 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167460.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:45:17,749 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167468.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:45:25,587 INFO [optim.py:369] (2/7) 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 05:45:34,228 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 1400, loss[loss=0.1999, simple_loss=0.28, pruned_loss=0.05985, over 4748.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2382, pruned_loss=0.04598, over 951473.29 frames. ], batch size: 54, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:46:14,262 INFO [finetune.py:976] (2/7) Epoch 30, batch 1450, loss[loss=0.1857, simple_loss=0.255, pruned_loss=0.05826, over 4131.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.24, pruned_loss=0.04644, over 952649.84 frames. ], batch size: 65, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:46:34,040 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5841, 2.7079, 2.1493, 2.3780, 2.7322, 2.3060, 3.5190, 1.9355], + device='cuda:2'), covar=tensor([0.3874, 0.2610, 0.4309, 0.3614, 0.1822, 0.2768, 0.1702, 0.4734], + device='cuda:2'), in_proj_covar=tensor([0.0337, 0.0352, 0.0420, 0.0349, 0.0382, 0.0371, 0.0369, 0.0421], + device='cuda:2'), out_proj_covar=tensor([9.9381e-05, 1.0460e-04, 1.2700e-04, 1.0392e-04, 1.1299e-04, 1.1024e-04, + 1.0730e-04, 1.2642e-04], device='cuda:2') +2023-04-28 05:46:52,053 INFO [optim.py:369] (2/7) 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 05:47:00,605 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167585.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:47:21,140 INFO [finetune.py:976] (2/7) Epoch 30, batch 1500, loss[loss=0.163, simple_loss=0.2382, pruned_loss=0.04387, over 4781.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2423, pruned_loss=0.04726, over 954439.05 frames. ], batch size: 51, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:47:52,026 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.65 vs. limit=2.0 +2023-04-28 05:47:56,720 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167633.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:48:20,792 INFO [finetune.py:976] (2/7) Epoch 30, batch 1550, loss[loss=0.1669, simple_loss=0.2365, pruned_loss=0.04868, over 4920.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2415, pruned_loss=0.04651, over 955044.57 frames. ], batch size: 38, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:48:40,247 INFO [optim.py:369] (2/7) 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 05:48:54,269 INFO [finetune.py:976] (2/7) Epoch 30, batch 1600, loss[loss=0.1673, simple_loss=0.2285, pruned_loss=0.05309, over 4837.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2391, pruned_loss=0.0455, over 956356.74 frames. ], batch size: 49, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:49:05,586 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167719.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:28,180 INFO [finetune.py:976] (2/7) Epoch 30, batch 1650, loss[loss=0.1351, simple_loss=0.2078, pruned_loss=0.0312, over 4702.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2365, pruned_loss=0.0448, over 957941.21 frames. ], batch size: 23, lr: 2.81e-03, grad_scale: 32.0 +2023-04-28 05:49:30,083 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167756.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:32,990 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167760.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:37,857 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 05:49:47,182 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167780.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:49:48,861 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1566, 2.4575, 2.3560, 2.7574, 2.7033, 2.6007, 2.4157, 4.9254], + device='cuda:2'), covar=tensor([0.0476, 0.0640, 0.0669, 0.0965, 0.0493, 0.0411, 0.0616, 0.0150], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 05:49:52,950 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167788.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:02,179 INFO [finetune.py:976] (2/7) Epoch 30, batch 1700, loss[loss=0.2193, simple_loss=0.2836, pruned_loss=0.07747, over 4909.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2342, pruned_loss=0.04389, over 956816.21 frames. ], batch size: 36, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:50:02,850 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167804.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:05,283 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167808.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:18,837 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167828.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:50:35,263 INFO [finetune.py:976] (2/7) Epoch 30, batch 1750, loss[loss=0.175, simple_loss=0.255, pruned_loss=0.04749, over 4814.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.237, pruned_loss=0.04515, over 955986.77 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:50:53,215 INFO [optim.py:369] (2/7) 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 05:51:08,163 INFO [finetune.py:976] (2/7) Epoch 30, batch 1800, loss[loss=0.154, simple_loss=0.2367, pruned_loss=0.03563, over 4927.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2403, pruned_loss=0.04586, over 956627.28 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:51:41,402 INFO [finetune.py:976] (2/7) Epoch 30, batch 1850, loss[loss=0.1862, simple_loss=0.2564, pruned_loss=0.05807, over 4833.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2421, pruned_loss=0.04684, over 955204.96 frames. ], batch size: 49, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:52:04,617 INFO [optim.py:369] (2/7) 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 05:52:26,790 INFO [finetune.py:976] (2/7) Epoch 30, batch 1900, loss[loss=0.1579, simple_loss=0.2367, pruned_loss=0.03954, over 4883.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2427, pruned_loss=0.04669, over 956098.50 frames. ], batch size: 43, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:52:48,421 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.5259, 4.4881, 3.0409, 5.2061, 4.5223, 4.5249, 2.1118, 4.4766], + device='cuda:2'), covar=tensor([0.1741, 0.0921, 0.3518, 0.0976, 0.3091, 0.1471, 0.5252, 0.2018], + device='cuda:2'), in_proj_covar=tensor([0.0248, 0.0218, 0.0253, 0.0303, 0.0300, 0.0249, 0.0274, 0.0274], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 05:53:09,283 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168035.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:53:09,895 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0475, 2.5384, 1.0501, 1.3602, 1.8505, 1.2135, 3.5026, 1.6861], + device='cuda:2'), covar=tensor([0.0732, 0.0820, 0.0842, 0.1263, 0.0546, 0.1087, 0.0188, 0.0613], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 05:53:11,143 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0813, 2.3365, 1.0201, 1.3045, 1.7213, 1.2086, 3.0855, 1.5389], + device='cuda:2'), covar=tensor([0.0703, 0.0575, 0.0738, 0.1352, 0.0530, 0.1077, 0.0291, 0.0660], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0050, 0.0052, 0.0073, 0.0051], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 05:53:31,845 INFO [finetune.py:976] (2/7) Epoch 30, batch 1950, loss[loss=0.1594, simple_loss=0.2409, pruned_loss=0.03894, over 4844.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2412, pruned_loss=0.04614, over 954406.64 frames. ], batch size: 44, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:53:56,263 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168075.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:54:04,671 INFO [optim.py:369] (2/7) 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 05:54:13,526 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168088.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:54:24,479 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=168096.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 05:54:32,405 INFO [finetune.py:976] (2/7) Epoch 30, batch 2000, loss[loss=0.1573, simple_loss=0.2276, pruned_loss=0.04348, over 4933.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2372, pruned_loss=0.04495, over 954506.45 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:54:55,111 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168136.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:55:30,826 INFO [finetune.py:976] (2/7) Epoch 30, batch 2050, loss[loss=0.1394, simple_loss=0.2095, pruned_loss=0.0347, over 4053.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2338, pruned_loss=0.04381, over 954798.62 frames. ], batch size: 17, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:55:40,689 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7984, 2.3207, 1.8030, 2.1737, 1.6072, 1.9535, 1.8535, 1.3699], + device='cuda:2'), covar=tensor([0.1867, 0.1083, 0.0852, 0.1088, 0.3347, 0.1028, 0.1872, 0.2474], + device='cuda:2'), in_proj_covar=tensor([0.0279, 0.0298, 0.0216, 0.0271, 0.0306, 0.0250, 0.0245, 0.0258], + device='cuda:2'), 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:2') +2023-04-28 05:55:43,772 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9447, 2.4396, 1.9848, 1.9929, 1.4521, 1.5021, 2.0032, 1.3637], + device='cuda:2'), covar=tensor([0.1646, 0.1302, 0.1349, 0.1468, 0.2063, 0.1807, 0.0894, 0.1914], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0210, 0.0170, 0.0205, 0.0201, 0.0188, 0.0157, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 05:55:44,982 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8340, 2.5529, 2.7333, 3.3814, 3.1957, 2.6658, 2.4489, 3.0116], + device='cuda:2'), covar=tensor([0.0801, 0.0976, 0.0627, 0.0532, 0.0550, 0.0773, 0.0623, 0.0492], + device='cuda:2'), in_proj_covar=tensor([0.0181, 0.0201, 0.0181, 0.0169, 0.0176, 0.0176, 0.0148, 0.0174], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 05:55:47,311 INFO [optim.py:369] (2/7) 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 05:56:04,271 INFO [finetune.py:976] (2/7) Epoch 30, batch 2100, loss[loss=0.1575, simple_loss=0.2349, pruned_loss=0.04007, over 4816.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2346, pruned_loss=0.04447, over 954915.45 frames. ], batch size: 41, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:56:23,095 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168233.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:56:31,969 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4700, 1.4368, 1.8611, 1.8307, 1.3563, 1.2422, 1.4755, 0.8911], + device='cuda:2'), covar=tensor([0.0577, 0.0586, 0.0348, 0.0579, 0.0749, 0.1061, 0.0593, 0.0604], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:56:37,574 INFO [finetune.py:976] (2/7) Epoch 30, batch 2150, loss[loss=0.1818, simple_loss=0.241, pruned_loss=0.06128, over 4102.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.238, pruned_loss=0.04545, over 953798.78 frames. ], batch size: 18, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:56:54,473 INFO [optim.py:369] (2/7) 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 05:57:04,546 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 2200, loss[loss=0.1937, simple_loss=0.2684, pruned_loss=0.05954, over 4743.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.24, pruned_loss=0.04563, over 955143.70 frames. ], batch size: 59, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:57:18,061 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4406, 1.3820, 1.8065, 1.7229, 1.3247, 1.2445, 1.3843, 0.8754], + device='cuda:2'), covar=tensor([0.0556, 0.0509, 0.0307, 0.0466, 0.0682, 0.0967, 0.0517, 0.0512], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0076, 0.0095, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 05:58:09,636 INFO [finetune.py:976] (2/7) Epoch 30, batch 2250, loss[loss=0.1687, simple_loss=0.2485, pruned_loss=0.04444, over 4898.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2417, pruned_loss=0.04632, over 952418.77 frames. ], batch size: 36, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:58:40,749 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168375.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:58:43,710 INFO [optim.py:369] (2/7) 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 05:58:55,053 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168391.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 05:59:13,990 INFO [finetune.py:976] (2/7) Epoch 30, batch 2300, loss[loss=0.1372, simple_loss=0.2171, pruned_loss=0.02861, over 4787.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.242, pruned_loss=0.04558, over 954300.18 frames. ], batch size: 51, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 05:59:38,059 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:59:38,102 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 05:59:39,305 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.8845, 2.8372, 2.2060, 3.3134, 2.8872, 2.9081, 1.1670, 2.8396], + device='cuda:2'), covar=tensor([0.2210, 0.1712, 0.3512, 0.3022, 0.3386, 0.2188, 0.5774, 0.2869], + device='cuda:2'), in_proj_covar=tensor([0.0247, 0.0216, 0.0252, 0.0302, 0.0298, 0.0248, 0.0273, 0.0272], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:00:16,725 INFO [finetune.py:976] (2/7) Epoch 30, batch 2350, loss[loss=0.1401, simple_loss=0.225, pruned_loss=0.02764, over 4816.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2392, pruned_loss=0.04483, over 955582.00 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:00:37,860 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168471.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:00:47,296 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-04-28 06:00:48,854 INFO [optim.py:369] (2/7) 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 06:01:21,061 INFO [finetune.py:976] (2/7) Epoch 30, batch 2400, loss[loss=0.1175, simple_loss=0.1855, pruned_loss=0.02471, over 4747.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2362, pruned_loss=0.04421, over 955120.46 frames. ], batch size: 28, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:01:33,554 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3585, 1.7601, 2.1725, 2.5003, 2.1133, 1.7695, 1.2858, 1.9004], + device='cuda:2'), covar=tensor([0.2835, 0.2916, 0.1457, 0.2035, 0.2468, 0.2534, 0.4188, 0.1798], + device='cuda:2'), in_proj_covar=tensor([0.0293, 0.0247, 0.0229, 0.0316, 0.0223, 0.0237, 0.0230, 0.0186], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 06:02:28,242 INFO [finetune.py:976] (2/7) Epoch 30, batch 2450, loss[loss=0.1841, simple_loss=0.2525, pruned_loss=0.05786, over 4864.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.234, pruned_loss=0.04369, over 952616.23 frames. ], batch size: 34, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:02:40,502 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1920, 1.4735, 1.3009, 1.7175, 1.5413, 1.6820, 1.3653, 3.0668], + device='cuda:2'), covar=tensor([0.0640, 0.0798, 0.0800, 0.1226, 0.0642, 0.0510, 0.0756, 0.0161], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 06:03:10,378 INFO [optim.py:369] (2/7) 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 06:03:21,581 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168589.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:03:36,318 INFO [finetune.py:976] (2/7) Epoch 30, batch 2500, loss[loss=0.1832, simple_loss=0.2655, pruned_loss=0.05048, over 4919.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2353, pruned_loss=0.0442, over 953977.83 frames. ], batch size: 36, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:04:06,210 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5411, 1.5165, 1.9266, 1.8973, 1.4155, 1.2879, 1.5037, 0.9308], + device='cuda:2'), covar=tensor([0.0456, 0.0604, 0.0349, 0.0635, 0.0765, 0.1030, 0.0656, 0.0597], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:04:31,462 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-04-28 06:04:49,840 INFO [finetune.py:976] (2/7) Epoch 30, batch 2550, loss[loss=0.1935, simple_loss=0.2678, pruned_loss=0.05966, over 4828.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2396, pruned_loss=0.04503, over 953570.22 frames. ], batch size: 47, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:05:22,980 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1847, 1.4136, 1.2721, 1.7074, 1.6075, 1.5533, 1.3995, 2.4383], + device='cuda:2'), covar=tensor([0.0577, 0.0810, 0.0804, 0.1213, 0.0613, 0.0467, 0.0721, 0.0234], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0042, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0013, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 06:05:23,460 INFO [optim.py:369] (2/7) 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 06:05:31,392 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 2600, loss[loss=0.1961, simple_loss=0.2749, pruned_loss=0.05868, over 4816.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2412, pruned_loss=0.04581, over 953809.06 frames. ], batch size: 40, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:05:38,797 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6495, 2.1579, 1.7351, 2.0850, 1.6052, 1.8181, 1.7305, 1.3129], + device='cuda:2'), covar=tensor([0.1773, 0.0986, 0.0821, 0.1032, 0.3292, 0.1045, 0.1901, 0.2384], + device='cuda:2'), in_proj_covar=tensor([0.0279, 0.0296, 0.0215, 0.0269, 0.0305, 0.0250, 0.0244, 0.0256], + device='cuda:2'), out_proj_covar=tensor([1.1094e-04, 1.1637e-04, 8.4232e-05, 1.0573e-04, 1.2264e-04, 9.8118e-05, + 9.8084e-05, 1.0077e-04], device='cuda:2') +2023-04-28 06:06:14,256 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.3040, 1.4562, 1.8545, 1.8915, 1.8617, 1.8709, 1.8330, 1.8438], + device='cuda:2'), covar=tensor([0.3858, 0.5394, 0.4026, 0.4309, 0.5055, 0.6463, 0.4882, 0.4295], + device='cuda:2'), in_proj_covar=tensor([0.0346, 0.0376, 0.0332, 0.0344, 0.0353, 0.0396, 0.0363, 0.0336], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:06:26,244 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168739.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:06:36,633 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4688, 1.3865, 1.8157, 1.7607, 1.3314, 1.2320, 1.4513, 0.9309], + device='cuda:2'), covar=tensor([0.0534, 0.0596, 0.0339, 0.0672, 0.0703, 0.1007, 0.0585, 0.0587], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0069, 0.0075, 0.0095, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:06:40,211 INFO [finetune.py:976] (2/7) Epoch 30, batch 2650, loss[loss=0.1464, simple_loss=0.2206, pruned_loss=0.0361, over 4895.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2414, pruned_loss=0.0456, over 951368.92 frames. ], batch size: 37, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:06:46,597 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9465, 2.4481, 0.9748, 1.2206, 1.6599, 1.2629, 2.9995, 1.5510], + device='cuda:2'), covar=tensor([0.0860, 0.0660, 0.0915, 0.1776, 0.0634, 0.1318, 0.0514, 0.0926], + device='cuda:2'), in_proj_covar=tensor([0.0051, 0.0064, 0.0047, 0.0046, 0.0049, 0.0051, 0.0073, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 06:07:17,594 INFO [optim.py:369] (2/7) 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 06:07:44,072 INFO [finetune.py:976] (2/7) Epoch 30, batch 2700, loss[loss=0.1412, simple_loss=0.2195, pruned_loss=0.03141, over 4763.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2393, pruned_loss=0.04461, over 953406.65 frames. ], batch size: 51, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:08:02,871 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-04-28 06:08:08,612 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5588, 1.8331, 2.0228, 2.0581, 1.9373, 1.9193, 2.0086, 1.9953], + device='cuda:2'), covar=tensor([0.3748, 0.5573, 0.4189, 0.4381, 0.5452, 0.6878, 0.5265, 0.4724], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0374, 0.0331, 0.0343, 0.0351, 0.0395, 0.0362, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:08:17,670 INFO [finetune.py:976] (2/7) Epoch 30, batch 2750, loss[loss=0.1469, simple_loss=0.2166, pruned_loss=0.03857, over 4815.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2365, pruned_loss=0.04353, over 952593.19 frames. ], batch size: 25, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:08:28,574 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-04-28 06:08:35,306 INFO [optim.py:369] (2/7) 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 06:08:41,776 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168889.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:08:50,765 INFO [finetune.py:976] (2/7) Epoch 30, batch 2800, loss[loss=0.1769, simple_loss=0.2544, pruned_loss=0.04967, over 4912.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2345, pruned_loss=0.04328, over 954246.98 frames. ], batch size: 37, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:09:02,328 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([4.0962, 3.9530, 2.9880, 4.7919, 4.0335, 4.0996, 1.6301, 4.1029], + device='cuda:2'), covar=tensor([0.1834, 0.1285, 0.4189, 0.1084, 0.4281, 0.1778, 0.6161, 0.2312], + device='cuda:2'), in_proj_covar=tensor([0.0249, 0.0218, 0.0253, 0.0303, 0.0300, 0.0250, 0.0275, 0.0275], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:09:12,960 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168937.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:09:24,482 INFO [finetune.py:976] (2/7) Epoch 30, batch 2850, loss[loss=0.141, simple_loss=0.2177, pruned_loss=0.03213, over 4780.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2336, pruned_loss=0.0433, over 956674.11 frames. ], batch size: 27, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:09:31,353 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1447, 1.6018, 1.9928, 2.3725, 2.0367, 1.5828, 1.2862, 1.8485], + device='cuda:2'), covar=tensor([0.2830, 0.2947, 0.1591, 0.1977, 0.2288, 0.2524, 0.4136, 0.1789], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0247, 0.0231, 0.0317, 0.0225, 0.0237, 0.0231, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 06:09:33,162 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168967.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:09:34,349 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168969.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 06:09:41,977 INFO [optim.py:369] (2/7) 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 06:09:58,543 INFO [finetune.py:976] (2/7) Epoch 30, batch 2900, loss[loss=0.2018, simple_loss=0.2829, pruned_loss=0.06029, over 4813.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2372, pruned_loss=0.04448, over 955204.34 frames. ], batch size: 39, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:10:10,334 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-04-28 06:10:14,812 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169028.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:10:16,014 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169030.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 06:10:31,849 INFO [finetune.py:976] (2/7) Epoch 30, batch 2950, loss[loss=0.1529, simple_loss=0.2234, pruned_loss=0.04123, over 4775.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2394, pruned_loss=0.04471, over 954858.47 frames. ], batch size: 26, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:10:44,249 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7701, 1.2192, 1.8028, 2.2501, 1.8476, 1.6902, 1.7397, 1.7212], + device='cuda:2'), covar=tensor([0.4522, 0.7176, 0.6260, 0.5219, 0.5578, 0.7588, 0.7973, 0.9368], + device='cuda:2'), in_proj_covar=tensor([0.0447, 0.0425, 0.0519, 0.0506, 0.0473, 0.0513, 0.0513, 0.0529], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:10:49,917 INFO [optim.py:369] (2/7) 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 06:10:50,643 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169082.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:11:05,765 INFO [finetune.py:976] (2/7) Epoch 30, batch 3000, loss[loss=0.1291, simple_loss=0.1947, pruned_loss=0.03173, over 4703.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2413, pruned_loss=0.04525, over 954757.24 frames. ], batch size: 23, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:11:05,765 INFO [finetune.py:1001] (2/7) Computing validation loss +2023-04-28 06:11:10,910 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9362, 2.2324, 1.8865, 1.6271, 1.4555, 1.5009, 1.8889, 1.4384], + device='cuda:2'), covar=tensor([0.1654, 0.1336, 0.1352, 0.1592, 0.2254, 0.1806, 0.0975, 0.2048], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0210, 0.0171, 0.0205, 0.0202, 0.0189, 0.0158, 0.0188], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 06:11:16,542 INFO [finetune.py:1010] (2/7) 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] (2/7) Maximum memory allocated so far is 6415MB +2023-04-28 06:11:24,701 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169116.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:11:52,085 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169143.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:12:04,394 INFO [finetune.py:976] (2/7) Epoch 30, batch 3050, loss[loss=0.1722, simple_loss=0.2385, pruned_loss=0.05301, over 4693.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2409, pruned_loss=0.04445, over 955382.58 frames. ], batch size: 59, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:12:13,891 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169159.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:12:36,608 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169177.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:12:42,082 INFO [optim.py:369] (2/7) 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 06:13:06,207 INFO [finetune.py:976] (2/7) Epoch 30, batch 3100, loss[loss=0.2053, simple_loss=0.2684, pruned_loss=0.07112, over 4926.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.238, pruned_loss=0.0432, over 956550.93 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:13:29,640 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169220.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:13:50,606 INFO [finetune.py:976] (2/7) Epoch 30, batch 3150, loss[loss=0.1783, simple_loss=0.247, pruned_loss=0.05481, over 4850.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2365, pruned_loss=0.04337, over 959053.86 frames. ], batch size: 47, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:14:10,085 INFO [optim.py:369] (2/7) 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 06:14:16,889 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.1126, 2.5838, 1.1627, 1.4807, 2.0098, 1.2644, 3.4954, 1.8430], + device='cuda:2'), covar=tensor([0.0689, 0.0634, 0.0794, 0.1168, 0.0482, 0.1024, 0.0231, 0.0584], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0046, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 06:14:22,311 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5165, 1.9219, 1.9562, 2.0632, 1.9339, 1.9021, 1.9467, 1.9508], + device='cuda:2'), covar=tensor([0.3862, 0.5470, 0.4553, 0.4259, 0.5625, 0.6913, 0.5523, 0.4991], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0373, 0.0330, 0.0341, 0.0350, 0.0393, 0.0361, 0.0334], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:14:23,996 INFO [finetune.py:976] (2/7) Epoch 30, batch 3200, loss[loss=0.1508, simple_loss=0.2215, pruned_loss=0.04002, over 4772.00 frames. ], tot_loss[loss=0.1588, simple_loss=0.2332, pruned_loss=0.04223, over 958795.25 frames. ], batch size: 26, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:14:38,825 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169323.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:14:40,578 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169325.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 06:14:57,967 INFO [finetune.py:976] (2/7) Epoch 30, batch 3250, loss[loss=0.1785, simple_loss=0.2615, pruned_loss=0.04772, over 4853.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2348, pruned_loss=0.04381, over 955100.70 frames. ], batch size: 47, lr: 2.80e-03, grad_scale: 32.0 +2023-04-28 06:15:11,016 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2965, 1.5480, 1.4230, 1.7724, 1.6420, 1.8057, 1.4246, 3.5436], + device='cuda:2'), covar=tensor([0.0622, 0.0838, 0.0816, 0.1263, 0.0665, 0.0539, 0.0762, 0.0147], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0039, 0.0039, 0.0043, 0.0040, 0.0038, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 06:15:17,082 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-04-28 06:15:18,034 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 30, batch 3300, loss[loss=0.1665, simple_loss=0.234, pruned_loss=0.04956, over 4743.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2382, pruned_loss=0.04499, over 955375.64 frames. ], batch size: 23, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:15:56,490 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169438.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:15:58,399 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4325, 1.8345, 1.9170, 1.9739, 1.8622, 1.8648, 1.9995, 1.8673], + device='cuda:2'), covar=tensor([0.3669, 0.5182, 0.4152, 0.4326, 0.5162, 0.6579, 0.5013, 0.4585], + device='cuda:2'), in_proj_covar=tensor([0.0344, 0.0375, 0.0331, 0.0343, 0.0352, 0.0396, 0.0363, 0.0336], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:16:05,662 INFO [finetune.py:976] (2/7) Epoch 30, batch 3350, loss[loss=0.1596, simple_loss=0.2384, pruned_loss=0.04038, over 4881.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2405, pruned_loss=0.04569, over 955917.87 frames. ], batch size: 43, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:16:18,727 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 06:16:33,500 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6473, 2.7115, 2.2006, 2.4257, 2.6796, 2.4655, 3.6116, 2.1215], + device='cuda:2'), covar=tensor([0.3672, 0.2341, 0.4309, 0.3028, 0.1710, 0.2273, 0.1639, 0.4102], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0357, 0.0427, 0.0353, 0.0386, 0.0376, 0.0372, 0.0427], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:16:39,414 INFO [finetune.py:976] (2/7) Epoch 30, batch 3400, loss[loss=0.2286, simple_loss=0.2897, pruned_loss=0.08379, over 4110.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2409, pruned_loss=0.04579, over 955821.09 frames. ], batch size: 65, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:16:41,450 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.71 vs. limit=5.0 +2023-04-28 06:16:47,268 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=169515.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:17:07,998 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.19 vs. limit=5.0 +2023-04-28 06:17:12,721 INFO [finetune.py:976] (2/7) Epoch 30, batch 3450, loss[loss=0.1524, simple_loss=0.2177, pruned_loss=0.04357, over 4867.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.24, pruned_loss=0.0452, over 955417.30 frames. ], batch size: 31, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:17:15,232 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9659, 1.8161, 2.3518, 2.5156, 1.7158, 1.6135, 1.8189, 0.9781], + device='cuda:2'), covar=tensor([0.0531, 0.0701, 0.0370, 0.0550, 0.0812, 0.1039, 0.0641, 0.0725], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:17:21,146 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5529, 1.4230, 1.9023, 1.8663, 1.3859, 1.3125, 1.4867, 0.9356], + device='cuda:2'), covar=tensor([0.0470, 0.0612, 0.0313, 0.0533, 0.0762, 0.1073, 0.0503, 0.0559], + device='cuda:2'), in_proj_covar=tensor([0.0071, 0.0067, 0.0065, 0.0068, 0.0075, 0.0094, 0.0072, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:17:36,960 INFO [optim.py:369] (2/7) 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 06:18:07,732 INFO [finetune.py:976] (2/7) Epoch 30, batch 3500, loss[loss=0.1799, simple_loss=0.2412, pruned_loss=0.05937, over 4867.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2375, pruned_loss=0.0445, over 955662.02 frames. ], batch size: 34, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:18:29,321 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169621.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:18:31,006 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169623.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:18:32,264 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169625.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 06:19:13,502 INFO [finetune.py:976] (2/7) Epoch 30, batch 3550, loss[loss=0.1479, simple_loss=0.217, pruned_loss=0.03935, over 4851.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.235, pruned_loss=0.04386, over 956746.54 frames. ], batch size: 49, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:19:25,261 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.2442, 2.5634, 0.8397, 1.5103, 1.5428, 1.8619, 1.6415, 0.8811], + device='cuda:2'), covar=tensor([0.1485, 0.1120, 0.1779, 0.1364, 0.1158, 0.0997, 0.1650, 0.1624], + device='cuda:2'), in_proj_covar=tensor([0.0118, 0.0238, 0.0136, 0.0121, 0.0131, 0.0154, 0.0118, 0.0118], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:19:30,804 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169673.0, num_to_drop=1, layers_to_drop={1} +2023-04-28 06:19:42,308 INFO [optim.py:369] (2/7) 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 06:19:43,559 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169682.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:20:12,670 INFO [finetune.py:976] (2/7) Epoch 30, batch 3600, loss[loss=0.1779, simple_loss=0.2607, pruned_loss=0.04757, over 4822.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2337, pruned_loss=0.04387, over 957016.24 frames. ], batch size: 39, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:20:45,831 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169729.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:20:57,512 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169738.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:21:10,890 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169747.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:21:20,083 INFO [finetune.py:976] (2/7) Epoch 30, batch 3650, loss[loss=0.1562, simple_loss=0.2456, pruned_loss=0.03345, over 4898.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2358, pruned_loss=0.04435, over 955409.62 frames. ], batch size: 43, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:21:43,295 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=169772.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:21:53,879 INFO [optim.py:369] (2/7) 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 06:22:03,118 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169786.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:03,159 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169786.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:11,321 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169790.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:16,004 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.8333, 1.7702, 1.8517, 1.4365, 1.9109, 1.6359, 2.4005, 1.5969], + device='cuda:2'), covar=tensor([0.3276, 0.1839, 0.4358, 0.2699, 0.1425, 0.2243, 0.1429, 0.4385], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0360, 0.0430, 0.0356, 0.0388, 0.0379, 0.0373, 0.0428], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:22:25,953 INFO [finetune.py:976] (2/7) Epoch 30, batch 3700, loss[loss=0.2314, simple_loss=0.2899, pruned_loss=0.08645, over 4811.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2401, pruned_loss=0.04501, over 956390.30 frames. ], batch size: 40, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:22:35,270 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169808.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:22:44,958 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169820.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:23:22,784 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169847.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:23:31,474 INFO [finetune.py:976] (2/7) Epoch 30, batch 3750, loss[loss=0.1786, simple_loss=0.2655, pruned_loss=0.04589, over 4798.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2411, pruned_loss=0.04531, over 956978.18 frames. ], batch size: 51, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:23:42,221 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=169863.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:24:04,805 INFO [optim.py:369] (2/7) 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 06:24:14,009 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-04-28 06:24:14,494 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2685, 2.9044, 2.2423, 2.3487, 1.6648, 1.6271, 2.5551, 1.6376], + device='cuda:2'), covar=tensor([0.1676, 0.1478, 0.1285, 0.1523, 0.2123, 0.1854, 0.0876, 0.1895], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0208, 0.0169, 0.0203, 0.0200, 0.0187, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 06:24:16,925 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4094, 1.2664, 1.6425, 1.5770, 1.3113, 1.1787, 1.2531, 0.6946], + device='cuda:2'), covar=tensor([0.0518, 0.0657, 0.0368, 0.0579, 0.0673, 0.1120, 0.0557, 0.0608], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0064, 0.0068, 0.0075, 0.0094, 0.0072, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:24:36,633 INFO [finetune.py:976] (2/7) Epoch 30, batch 3800, loss[loss=0.1393, simple_loss=0.1959, pruned_loss=0.04133, over 3970.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.241, pruned_loss=0.04457, over 955320.54 frames. ], batch size: 17, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:25:05,034 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8320, 1.4830, 1.9495, 2.3055, 1.9112, 1.7875, 1.8968, 1.8125], + device='cuda:2'), covar=tensor([0.4605, 0.6965, 0.6573, 0.5483, 0.6015, 0.7985, 0.8386, 1.0050], + device='cuda:2'), in_proj_covar=tensor([0.0448, 0.0429, 0.0524, 0.0509, 0.0477, 0.0517, 0.0517, 0.0533], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:25:18,399 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.3311, 1.3856, 1.5091, 1.5822, 1.7089, 1.3219, 0.9900, 1.4790], + device='cuda:2'), covar=tensor([0.0803, 0.1350, 0.0861, 0.0612, 0.0640, 0.0843, 0.0834, 0.0583], + device='cuda:2'), in_proj_covar=tensor([0.0184, 0.0205, 0.0185, 0.0173, 0.0179, 0.0180, 0.0151, 0.0178], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:25:40,862 INFO [finetune.py:976] (2/7) Epoch 30, batch 3850, loss[loss=0.1268, simple_loss=0.1974, pruned_loss=0.02807, over 3968.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2385, pruned_loss=0.04343, over 955268.87 frames. ], batch size: 17, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:26:12,313 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 06:26:45,747 INFO [finetune.py:976] (2/7) Epoch 30, batch 3900, loss[loss=0.1875, simple_loss=0.2563, pruned_loss=0.05932, over 4828.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2369, pruned_loss=0.04351, over 955075.44 frames. ], batch size: 33, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:27:36,463 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.6457, 1.4851, 4.4463, 4.2249, 3.9464, 4.3212, 4.1581, 3.9591], + device='cuda:2'), covar=tensor([0.7029, 0.5425, 0.1029, 0.1490, 0.0957, 0.1835, 0.1546, 0.1599], + device='cuda:2'), in_proj_covar=tensor([0.0308, 0.0308, 0.0403, 0.0406, 0.0346, 0.0412, 0.0314, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:27:43,927 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=170045.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:27:48,653 INFO [finetune.py:976] (2/7) Epoch 30, batch 3950, loss[loss=0.1769, simple_loss=0.2473, pruned_loss=0.05323, over 4914.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2347, pruned_loss=0.04309, over 954139.01 frames. ], batch size: 46, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:28:24,771 INFO [optim.py:369] (2/7) 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 06:28:27,244 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170085.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:28:37,438 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-04-28 06:28:46,489 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.91 vs. limit=5.0 +2023-04-28 06:28:46,920 INFO [finetune.py:976] (2/7) Epoch 30, batch 4000, loss[loss=0.1581, simple_loss=0.2266, pruned_loss=0.04486, over 4736.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2342, pruned_loss=0.04347, over 954495.98 frames. ], batch size: 59, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:28:46,985 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170103.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:28:53,809 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=170106.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:29:39,699 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170142.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:29:50,083 INFO [finetune.py:976] (2/7) Epoch 30, batch 4050, loss[loss=0.1861, simple_loss=0.2645, pruned_loss=0.05382, over 4850.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2368, pruned_loss=0.04419, over 953706.54 frames. ], batch size: 47, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:30:07,133 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-04-28 06:30:22,184 INFO [optim.py:369] (2/7) 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 06:30:36,789 INFO [finetune.py:976] (2/7) Epoch 30, batch 4100, loss[loss=0.1777, simple_loss=0.2568, pruned_loss=0.04925, over 4903.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2381, pruned_loss=0.04448, over 952348.03 frames. ], batch size: 43, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:30:47,305 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8179, 1.4528, 1.9246, 2.2995, 1.8988, 1.7948, 1.8687, 1.7716], + device='cuda:2'), covar=tensor([0.4280, 0.7033, 0.6186, 0.5374, 0.5850, 0.8280, 0.7881, 0.9650], + device='cuda:2'), in_proj_covar=tensor([0.0447, 0.0427, 0.0521, 0.0507, 0.0475, 0.0515, 0.0516, 0.0531], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:31:08,389 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0048, 2.0853, 1.2146, 1.6275, 2.4078, 1.8056, 1.6912, 1.8106], + device='cuda:2'), covar=tensor([0.0447, 0.0343, 0.0261, 0.0512, 0.0214, 0.0448, 0.0467, 0.0520], + device='cuda:2'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0027], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 06:31:10,085 INFO [finetune.py:976] (2/7) Epoch 30, batch 4150, loss[loss=0.1661, simple_loss=0.246, pruned_loss=0.04313, over 4706.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2385, pruned_loss=0.04477, over 949940.53 frames. ], batch size: 54, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:31:26,667 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 06:31:40,696 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([3.1326, 3.3805, 2.6085, 2.9634, 3.4908, 2.9258, 2.8626, 3.1081], + device='cuda:2'), covar=tensor([0.0293, 0.0201, 0.0181, 0.0332, 0.0165, 0.0310, 0.0314, 0.0309], + device='cuda:2'), in_proj_covar=tensor([0.0027, 0.0023, 0.0021, 0.0028, 0.0019, 0.0027, 0.0027, 0.0027], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0046, 0.0039, 0.0053, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 06:31:42,966 INFO [finetune.py:976] (2/7) Epoch 30, batch 4200, loss[loss=0.1513, simple_loss=0.2238, pruned_loss=0.03938, over 4755.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2381, pruned_loss=0.04402, over 950494.36 frames. ], batch size: 27, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:31:49,336 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-04-28 06:31:58,818 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170325.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:32:16,768 INFO [finetune.py:976] (2/7) Epoch 30, batch 4250, loss[loss=0.1586, simple_loss=0.2307, pruned_loss=0.04324, over 4872.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2358, pruned_loss=0.04362, over 952525.05 frames. ], batch size: 34, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:32:36,243 INFO [optim.py:369] (2/7) 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 06:32:38,778 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170385.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:32:39,583 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-04-28 06:32:48,584 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170401.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:32:50,165 INFO [finetune.py:976] (2/7) Epoch 30, batch 4300, loss[loss=0.1663, simple_loss=0.2463, pruned_loss=0.04313, over 4793.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2334, pruned_loss=0.04288, over 952316.32 frames. ], batch size: 29, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:32:50,272 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170403.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:10,859 INFO [zipformer.py:1188] (2/7) 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] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=170442.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:22,271 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170451.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:23,431 INFO [finetune.py:976] (2/7) Epoch 30, batch 4350, loss[loss=0.1287, simple_loss=0.2074, pruned_loss=0.025, over 4765.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2307, pruned_loss=0.0423, over 951247.76 frames. ], batch size: 28, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:33:28,866 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=170461.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:33:30,134 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-04-28 06:33:52,314 INFO [optim.py:369] (2/7) 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 06:34:04,604 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170490.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:34:23,443 INFO [finetune.py:976] (2/7) Epoch 30, batch 4400, loss[loss=0.1449, simple_loss=0.2223, pruned_loss=0.03378, over 4932.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.233, pruned_loss=0.04337, over 952028.43 frames. ], batch size: 33, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:34:34,471 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-04-28 06:34:47,082 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=170522.0, num_to_drop=1, layers_to_drop={2} +2023-04-28 06:35:08,309 INFO [finetune.py:976] (2/7) Epoch 30, batch 4450, loss[loss=0.2076, simple_loss=0.2892, pruned_loss=0.06296, over 4824.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2369, pruned_loss=0.04425, over 953570.34 frames. ], batch size: 40, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:35:18,637 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 06:35:41,956 INFO [finetune.py:976] (2/7) Epoch 30, batch 4500, loss[loss=0.2063, simple_loss=0.267, pruned_loss=0.07278, over 4797.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2382, pruned_loss=0.04445, over 954386.65 frames. ], batch size: 45, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:35:45,265 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-04-28 06:35:46,294 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9485, 1.2269, 3.2641, 3.0212, 2.9353, 3.2517, 3.2216, 2.8751], + device='cuda:2'), covar=tensor([0.7700, 0.5484, 0.1613, 0.2496, 0.1556, 0.1990, 0.1535, 0.1890], + device='cuda:2'), in_proj_covar=tensor([0.0307, 0.0306, 0.0402, 0.0406, 0.0345, 0.0411, 0.0314, 0.0361], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:35:59,496 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 4550, loss[loss=0.1887, simple_loss=0.2466, pruned_loss=0.06543, over 4885.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2385, pruned_loss=0.04443, over 955121.89 frames. ], batch size: 32, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:36:33,303 INFO [optim.py:369] (2/7) 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 06:36:47,828 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 4600, loss[loss=0.1912, simple_loss=0.243, pruned_loss=0.06972, over 4008.00 frames. ], tot_loss[loss=0.163, simple_loss=0.238, pruned_loss=0.04396, over 955263.57 frames. ], batch size: 17, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:37:10,211 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.16 vs. limit=5.0 +2023-04-28 06:37:10,673 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.5962, 1.2969, 4.2494, 4.0038, 3.6940, 4.0854, 3.9656, 3.7477], + device='cuda:2'), covar=tensor([0.7056, 0.5586, 0.1029, 0.1739, 0.1223, 0.1837, 0.1613, 0.1578], + device='cuda:2'), in_proj_covar=tensor([0.0310, 0.0308, 0.0405, 0.0409, 0.0348, 0.0414, 0.0316, 0.0363], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:37:19,720 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=170749.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:37:22,624 INFO [finetune.py:976] (2/7) Epoch 30, batch 4650, loss[loss=0.1422, simple_loss=0.2155, pruned_loss=0.03445, over 4822.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2372, pruned_loss=0.04445, over 953397.03 frames. ], batch size: 39, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:37:33,014 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7523, 2.6813, 2.1432, 2.4189, 2.7399, 2.3364, 3.4699, 2.0846], + device='cuda:2'), covar=tensor([0.3232, 0.2134, 0.4299, 0.3180, 0.1557, 0.2467, 0.1486, 0.4107], + device='cuda:2'), in_proj_covar=tensor([0.0340, 0.0357, 0.0428, 0.0353, 0.0386, 0.0376, 0.0371, 0.0424], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:37:39,986 INFO [optim.py:369] (2/7) 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 06:37:40,869 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-04-28 06:37:49,193 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7607, 1.2672, 1.7693, 2.2540, 1.8199, 1.6917, 1.7326, 1.6908], + device='cuda:2'), covar=tensor([0.4360, 0.6911, 0.5926, 0.5416, 0.5734, 0.7272, 0.8018, 0.8820], + device='cuda:2'), in_proj_covar=tensor([0.0450, 0.0429, 0.0523, 0.0510, 0.0476, 0.0518, 0.0518, 0.0534], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:37:55,525 INFO [finetune.py:976] (2/7) Epoch 30, batch 4700, loss[loss=0.1211, simple_loss=0.1931, pruned_loss=0.02453, over 4816.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2343, pruned_loss=0.04356, over 953859.66 frames. ], batch size: 25, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:38:05,045 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170817.0, num_to_drop=1, layers_to_drop={3} +2023-04-28 06:38:07,596 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-04-28 06:38:08,127 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7633, 1.3213, 1.8042, 2.2646, 1.8309, 1.7080, 1.7694, 1.7073], + device='cuda:2'), covar=tensor([0.4149, 0.6567, 0.5732, 0.4915, 0.5296, 0.7437, 0.7497, 0.8491], + device='cuda:2'), in_proj_covar=tensor([0.0449, 0.0428, 0.0522, 0.0510, 0.0475, 0.0517, 0.0517, 0.0533], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:38:23,810 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-04-28 06:38:29,291 INFO [finetune.py:976] (2/7) Epoch 30, batch 4750, loss[loss=0.1448, simple_loss=0.2099, pruned_loss=0.03987, over 4753.00 frames. ], tot_loss[loss=0.1597, simple_loss=0.2326, pruned_loss=0.04341, over 950244.62 frames. ], batch size: 23, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:38:31,257 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4637, 1.0721, 0.3516, 1.2214, 1.1314, 1.3527, 1.2677, 1.2999], + device='cuda:2'), covar=tensor([0.0505, 0.0400, 0.0404, 0.0548, 0.0296, 0.0503, 0.0468, 0.0582], + device='cuda:2'), in_proj_covar=tensor([0.0028, 0.0023, 0.0022, 0.0028, 0.0019, 0.0027, 0.0027, 0.0028], + device='cuda:2'), out_proj_covar=tensor([0.0052, 0.0047, 0.0039, 0.0054, 0.0039, 0.0051, 0.0051, 0.0053], + device='cuda:2') +2023-04-28 06:38:47,636 INFO [optim.py:369] (2/7) 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 06:38:47,795 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7094, 1.2318, 1.8774, 2.1747, 1.7764, 1.7109, 1.8090, 1.7023], + device='cuda:2'), covar=tensor([0.4397, 0.6561, 0.5442, 0.5558, 0.5579, 0.7312, 0.6877, 0.9225], + device='cuda:2'), in_proj_covar=tensor([0.0450, 0.0428, 0.0523, 0.0511, 0.0476, 0.0519, 0.0518, 0.0534], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:38:49,593 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7300, 1.6213, 1.9693, 2.0829, 1.5581, 1.3543, 1.5790, 0.8452], + device='cuda:2'), covar=tensor([0.0621, 0.0558, 0.0457, 0.0757, 0.0757, 0.1053, 0.0702, 0.0663], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0065, 0.0069, 0.0075, 0.0094, 0.0071, 0.0062], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:39:16,799 INFO [finetune.py:976] (2/7) Epoch 30, batch 4800, loss[loss=0.2018, simple_loss=0.2682, pruned_loss=0.06776, over 4901.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2366, pruned_loss=0.04491, over 951967.82 frames. ], batch size: 35, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:39:45,897 INFO [zipformer.py:1188] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=170925.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:39:47,794 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=170928.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:39:59,623 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-04-28 06:40:21,165 INFO [finetune.py:976] (2/7) Epoch 30, batch 4850, loss[loss=0.1732, simple_loss=0.2446, pruned_loss=0.05094, over 4823.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2382, pruned_loss=0.04451, over 952407.83 frames. ], batch size: 38, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:40:52,178 INFO [optim.py:369] (2/7) 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 06:40:59,532 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=170989.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:41:19,848 INFO [finetune.py:976] (2/7) Epoch 30, batch 4900, loss[loss=0.1445, simple_loss=0.2331, pruned_loss=0.02791, over 4923.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2386, pruned_loss=0.04457, over 951008.64 frames. ], batch size: 33, lr: 2.79e-03, grad_scale: 32.0 +2023-04-28 06:42:00,282 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7322, 2.3247, 1.7518, 1.8981, 1.2999, 1.3316, 1.8553, 1.2675], + device='cuda:2'), covar=tensor([0.1675, 0.1259, 0.1315, 0.1456, 0.2264, 0.1847, 0.0940, 0.2031], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0207, 0.0169, 0.0202, 0.0199, 0.0186, 0.0155, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 06:42:25,534 INFO [finetune.py:976] (2/7) Epoch 30, batch 4950, loss[loss=0.2286, simple_loss=0.2922, pruned_loss=0.08252, over 4788.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2408, pruned_loss=0.04557, over 950918.89 frames. ], batch size: 51, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:42:31,533 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.0400, 1.7869, 1.9954, 2.3746, 2.4632, 1.9814, 1.7412, 2.1907], + device='cuda:2'), covar=tensor([0.0758, 0.1081, 0.0727, 0.0529, 0.0527, 0.0830, 0.0640, 0.0478], + device='cuda:2'), in_proj_covar=tensor([0.0180, 0.0201, 0.0181, 0.0171, 0.0176, 0.0177, 0.0148, 0.0175], + device='cuda:2'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:2') +2023-04-28 06:42:45,067 INFO [optim.py:369] (2/7) 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 06:42:46,475 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.2689, 1.7070, 2.1072, 2.5256, 2.1309, 1.7024, 1.4147, 1.9457], + device='cuda:2'), covar=tensor([0.2997, 0.3058, 0.1518, 0.2086, 0.2333, 0.2490, 0.4237, 0.1889], + device='cuda:2'), in_proj_covar=tensor([0.0294, 0.0247, 0.0230, 0.0316, 0.0225, 0.0236, 0.0230, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:2') +2023-04-28 06:42:52,370 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171092.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:42:59,424 INFO [finetune.py:976] (2/7) Epoch 30, batch 5000, loss[loss=0.1338, simple_loss=0.2074, pruned_loss=0.03008, over 4758.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2393, pruned_loss=0.04501, over 950295.50 frames. ], batch size: 28, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:43:09,468 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171117.0, num_to_drop=1, layers_to_drop={0} +2023-04-28 06:43:18,355 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7715, 2.1666, 1.6971, 1.5663, 1.3495, 1.3859, 1.7403, 1.2632], + device='cuda:2'), covar=tensor([0.1631, 0.1268, 0.1433, 0.1704, 0.2272, 0.1882, 0.0962, 0.2081], + device='cuda:2'), in_proj_covar=tensor([0.0198, 0.0208, 0.0169, 0.0203, 0.0200, 0.0186, 0.0156, 0.0187], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 06:43:33,012 INFO [finetune.py:976] (2/7) Epoch 30, batch 5050, loss[loss=0.1346, simple_loss=0.2088, pruned_loss=0.03024, over 4245.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2366, pruned_loss=0.04436, over 951517.27 frames. ], batch size: 18, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:43:33,150 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171153.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:43:39,793 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.7898, 2.4721, 2.0560, 2.3852, 1.7952, 2.1272, 1.9965, 1.5174], + device='cuda:2'), covar=tensor([0.2017, 0.1127, 0.0812, 0.1012, 0.3028, 0.1004, 0.1741, 0.2346], + device='cuda:2'), in_proj_covar=tensor([0.0283, 0.0300, 0.0218, 0.0274, 0.0309, 0.0256, 0.0250, 0.0261], + device='cuda:2'), out_proj_covar=tensor([1.1265e-04, 1.1776e-04, 8.5391e-05, 1.0726e-04, 1.2444e-04, 1.0055e-04, + 1.0029e-04, 1.0261e-04], device='cuda:2') +2023-04-28 06:43:40,802 INFO [zipformer.py:1188] (2/7) 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] (2/7) 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 06:44:02,202 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171196.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:44:06,504 INFO [finetune.py:976] (2/7) Epoch 30, batch 5100, loss[loss=0.1497, simple_loss=0.228, pruned_loss=0.03575, over 4897.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2328, pruned_loss=0.04256, over 953873.36 frames. ], batch size: 35, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:44:21,429 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171225.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:44:22,723 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-04-28 06:44:25,625 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.71 vs. limit=2.0 +2023-04-28 06:44:40,210 INFO [finetune.py:976] (2/7) Epoch 30, batch 5150, loss[loss=0.2245, simple_loss=0.2955, pruned_loss=0.07673, over 4907.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2357, pruned_loss=0.0446, over 951232.88 frames. ], batch size: 43, lr: 2.78e-03, grad_scale: 64.0 +2023-04-28 06:44:42,787 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171257.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:44:53,817 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=171273.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:45:05,608 INFO [optim.py:369] (2/7) 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] (2/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=171284.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:45:35,542 INFO [finetune.py:976] (2/7) Epoch 30, batch 5200, loss[loss=0.2205, simple_loss=0.29, pruned_loss=0.07549, over 4841.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2388, pruned_loss=0.04547, over 953198.38 frames. ], batch size: 47, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:45:39,800 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.1640, 1.4250, 1.3162, 1.6805, 1.5605, 1.6704, 1.3618, 3.0353], + device='cuda:2'), covar=tensor([0.0639, 0.0840, 0.0825, 0.1226, 0.0649, 0.0540, 0.0757, 0.0150], + device='cuda:2'), in_proj_covar=tensor([0.0037, 0.0038, 0.0039, 0.0043, 0.0040, 0.0037, 0.0038, 0.0054], + device='cuda:2'), out_proj_covar=tensor([0.0012, 0.0013, 0.0014, 0.0015, 0.0014, 0.0013, 0.0013, 0.0016], + device='cuda:2') +2023-04-28 06:46:31,213 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171345.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:46:36,593 INFO [finetune.py:976] (2/7) Epoch 30, batch 5250, loss[loss=0.1493, simple_loss=0.2295, pruned_loss=0.03449, over 4782.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2408, pruned_loss=0.04635, over 952141.35 frames. ], batch size: 29, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:47:18,737 INFO [optim.py:369] (2/7) 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] (2/7) Epoch 30, batch 5300, loss[loss=0.1427, simple_loss=0.2204, pruned_loss=0.03244, over 4771.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2424, pruned_loss=0.04725, over 950575.48 frames. ], batch size: 28, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:47:43,406 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.9794, 2.5101, 2.1021, 2.0979, 1.4382, 1.5178, 2.2664, 1.4001], + device='cuda:2'), covar=tensor([0.1837, 0.1418, 0.1405, 0.1592, 0.2429, 0.2033, 0.0961, 0.2177], + device='cuda:2'), in_proj_covar=tensor([0.0200, 0.0210, 0.0170, 0.0205, 0.0202, 0.0188, 0.0157, 0.0189], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:2') +2023-04-28 06:47:50,516 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171406.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:48:10,605 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-04-28 06:48:41,620 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 5350, loss[loss=0.1529, simple_loss=0.2292, pruned_loss=0.03829, over 4777.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2425, pruned_loss=0.04716, over 950575.20 frames. ], batch size: 23, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:49:01,601 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.4812, 1.8486, 1.9498, 2.0486, 1.9023, 1.9270, 1.9710, 1.9641], + device='cuda:2'), covar=tensor([0.4195, 0.5595, 0.4477, 0.4259, 0.5781, 0.7162, 0.5338, 0.5017], + device='cuda:2'), in_proj_covar=tensor([0.0342, 0.0375, 0.0331, 0.0344, 0.0352, 0.0394, 0.0362, 0.0335], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:49:03,280 INFO [optim.py:369] (2/7) 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 06:49:10,279 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 06:49:15,609 INFO [zipformer.py:1188] (2/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=171498.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:49:18,572 INFO [finetune.py:976] (2/7) Epoch 30, batch 5400, loss[loss=0.1505, simple_loss=0.223, pruned_loss=0.03897, over 4764.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2403, pruned_loss=0.04646, over 951843.30 frames. ], batch size: 27, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:49:34,909 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.4250, 1.3130, 1.6553, 1.6141, 1.3412, 1.2072, 1.3768, 0.7959], + device='cuda:2'), covar=tensor([0.0510, 0.0584, 0.0347, 0.0449, 0.0692, 0.1090, 0.0479, 0.0510], + device='cuda:2'), in_proj_covar=tensor([0.0070, 0.0067, 0.0064, 0.0069, 0.0075, 0.0094, 0.0072, 0.0061], + device='cuda:2'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:2') +2023-04-28 06:49:51,781 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 5450, loss[loss=0.1478, simple_loss=0.2167, pruned_loss=0.03948, over 4819.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.237, pruned_loss=0.04532, over 950751.04 frames. ], batch size: 40, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:49:54,378 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.5359, 1.7868, 1.8676, 1.9766, 1.7827, 1.9406, 1.9943, 1.8966], + device='cuda:2'), covar=tensor([0.3609, 0.5393, 0.4651, 0.4595, 0.5802, 0.7295, 0.4916, 0.4875], + device='cuda:2'), in_proj_covar=tensor([0.0343, 0.0375, 0.0331, 0.0345, 0.0352, 0.0394, 0.0362, 0.0335], + device='cuda:2'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:2') +2023-04-28 06:49:56,141 INFO [zipformer.py:1188] (2/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=171559.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:50:10,996 INFO [optim.py:369] (2/7) 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 06:50:12,293 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171584.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:50:25,852 INFO [finetune.py:976] (2/7) Epoch 30, batch 5500, loss[loss=0.1401, simple_loss=0.2119, pruned_loss=0.03414, over 4759.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2335, pruned_loss=0.04392, over 952940.27 frames. ], batch size: 28, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:50:25,980 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([1.8976, 1.2637, 1.4119, 1.6642, 2.0030, 1.6195, 1.4769, 1.2936], + device='cuda:2'), covar=tensor([0.1636, 0.1899, 0.2029, 0.1548, 0.1093, 0.1812, 0.2241, 0.2480], + device='cuda:2'), in_proj_covar=tensor([0.0314, 0.0308, 0.0349, 0.0285, 0.0324, 0.0306, 0.0300, 0.0376], + device='cuda:2'), out_proj_covar=tensor([6.3650e-05, 6.2773e-05, 7.2912e-05, 5.6819e-05, 6.5944e-05, 6.3343e-05, + 6.1795e-05, 7.9403e-05], device='cuda:2') +2023-04-28 06:50:45,015 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 5550, loss[loss=0.2182, simple_loss=0.2954, pruned_loss=0.0705, over 4821.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2347, pruned_loss=0.04417, over 947299.78 frames. ], batch size: 40, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:51:38,772 INFO [optim.py:369] (2/7) 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 06:52:00,697 INFO [zipformer.py:1188] (2/7) 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] (2/7) Epoch 30, batch 5600, loss[loss=0.1286, simple_loss=0.1928, pruned_loss=0.03219, over 4471.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.237, pruned_loss=0.04426, over 948683.48 frames. ], batch size: 19, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:52:24,133 INFO [scaling.py:679] (2/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-04-28 06:52:36,164 INFO [scaling.py:679] (2/7) Whitening: num_groups=1, num_channels=384, metric=3.37 vs. limit=5.0 +2023-04-28 06:52:55,570 INFO [zipformer.py:1188] (2/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=171748.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:52:58,356 INFO [finetune.py:976] (2/7) Epoch 30, batch 5650, loss[loss=0.1946, simple_loss=0.2465, pruned_loss=0.07141, over 4698.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2386, pruned_loss=0.04436, over 947124.18 frames. ], batch size: 23, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:53:37,490 INFO [optim.py:369] (2/7) 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 06:53:37,609 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([0.9951, 0.6807, 0.7501, 0.7783, 1.0796, 0.9061, 0.8181, 0.7977], + device='cuda:2'), covar=tensor([0.1635, 0.1631, 0.1904, 0.1482, 0.1026, 0.1409, 0.1709, 0.2229], + device='cuda:2'), in_proj_covar=tensor([0.0315, 0.0308, 0.0351, 0.0286, 0.0325, 0.0307, 0.0302, 0.0378], + device='cuda:2'), out_proj_covar=tensor([6.3956e-05, 6.2899e-05, 7.3270e-05, 5.7083e-05, 6.6151e-05, 6.3443e-05, + 6.2251e-05, 7.9734e-05], device='cuda:2') +2023-04-28 06:53:50,970 INFO [zipformer.py:1188] (2/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=171796.0, num_to_drop=0, layers_to_drop=set() +2023-04-28 06:53:59,019 INFO [zipformer.py:2441] (2/7) attn_weights_entropy = tensor([2.9514, 3.5558, 1.5533, 2.2586, 3.1485, 2.1263, 4.8509, 2.7679], + device='cuda:2'), covar=tensor([0.0523, 0.0567, 0.0697, 0.1032, 0.0388, 0.0826, 0.0157, 0.0496], + device='cuda:2'), in_proj_covar=tensor([0.0050, 0.0064, 0.0046, 0.0046, 0.0049, 0.0051, 0.0072, 0.0050], + device='cuda:2'), out_proj_covar=tensor([0.0008, 0.0010, 0.0007, 0.0008, 0.0008, 0.0008, 0.0010, 0.0008], + device='cuda:2') +2023-04-28 06:54:00,720 INFO [finetune.py:976] (2/7) Epoch 30, batch 5700, loss[loss=0.1151, simple_loss=0.1747, pruned_loss=0.02773, over 4261.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2339, pruned_loss=0.04341, over 929873.62 frames. ], batch size: 18, lr: 2.78e-03, grad_scale: 32.0 +2023-04-28 06:54:33,364 INFO [finetune.py:1241] (2/7) Done!